Design of a model and controller for a lawn mowing robot

HO CHI MINH CITY UNIVERSITY OF TECHNOLOGY AND EDUCATIONGRADUATION THESIS MAJOR: AUTOMATION AND CONTROL ENGINEERING TECHNOLOGY INSTRUCTOR: DANG XUAN BA ADVISOR STUDENT: TRAN THANH DUNG NG

INTRODUCTION

Problem Statement

The rise of technology and the growing need for automated solutions have made robotic lawn mowers an effective answer to various practical challenges Particularly in expansive spaces, these advanced machines provide high precision and efficiency, ensuring that lawns remain clean and hygienic.

Robotic lawn mowers offer significant advantages by saving users time and effort in lawn care, allowing them to focus on other tasks These innovative machines enhance safety and convenience, especially in challenging terrains where manual mowing can be risky.

The integration of advanced remote-control technology, such as LoRa, significantly enhances the functionality of robotic lawn mowers, making them more flexible and user-friendly This innovation enables users to effortlessly manage and control their mowers remotely through mobile devices and the Internet, thereby extending their application to remote and challenging areas By leveraging cutting-edge technologies like LoRa in the development of robotic lawn mowers, we address practical needs while advancing towards a smarter, more modern, and sustainable digital gardening environment.

Figure 1.1: Husqvarna Automower 550 Robotic Lawn Mower

DEPARTMENT OF AUTOMATIC CONTROL 2 urban green spaces This will undoubtedly bring superior benefits to both users and the living environment around them

Robotic lawn mowers offer significant strengths and potential, paving the way for continued exploration and development of this technology By advancing these innovations, we can create smarter and more convenient solutions that enhance modern living.

Robotic lawn mowers have become increasingly popular worldwide, particularly in developed markets like Europe, the United States, and China These regions lead the way in the research, development, and commercialization of advanced robotic lawn mowers, offering a wide range of models and features to meet various consumer needs.

Europe is home to many globally renowned robotic lawn mower brands, such as

Husqvarna, Gardena, Bosch, and Robomow are leading brands in the European lawn care market, known for their innovative designs and advanced grass-cutting technologies These products excel in navigating complex terrains and feature cutting-edge technologies such as GPS navigation, mobile app control, and automatic battery charging for enhanced user convenience.

Figure 1.2: Robotic Lawn Mower Market – Growth Rate By Region, 2022 - 2027

China has become a leading center for robotic lawn mower production, featuring brands such as Ecovacs Robotics, Mamibot, and Novabot These Chinese models typically offer more competitive pricing than their European or American equivalents while still providing essential functions, including automatic grass cutting, self-charging capabilities, and app control.

Global research and development in robotic lawn mowers focus on the following directions:

Enhancing artificial intelligence in robots is crucial for improving their autonomy in complex situations This includes their ability to recognize and navigate obstacles, identify areas that require mowing, and automatically return to their charging stations when battery levels are low.

• Increasing Connectivity: Allowing users to control the robot remotely via mobile apps, monitor operating status, and schedule automatic mowing

• Developing Specialized Robots: Meeting the needs for mowing on special terrains such as golf courses, parks, and industrial areas

Figure 1.3: Robotic Lawn Mower Market – Growth Rate By Region, 2022 - 2027

Table 1.1: Classification of types of lawn cutting robots produced in different countries

• Suitable for many types of garden areas and terrains

• GPS positioning, control via mobile application, and automatic battery charging

• Ability to cut grass accurately, quietly and save energy

• Compact design, suitable for small, medium sized gardens

• Easy to install and use

• Ability to cut grass evenly and beautifully, operating smoothly

• Modern design, many smart features

• Ability to operate durably over large areas

• Easily set up and program automatic lawn mowing schedules

• Remote control via mobile application

• GPS positioning system helps the robot operate effectively and accurately

• Ability to automatically avoid obstacles and automatically return to the charging station

• Large capacity, suitable for large grass

The global research and development of robotic lawn mowers is rapidly evolving, drawing interest from technology firms, research institutions, and universities alike Current studies aim to improve the automation, efficiency, and intelligence of these mowers to satisfy the growing user demands and diverse applications across multiple sectors.

Some prominent research trends include:

Artificial Intelligence (AI) and Machine Learning enhance a robot's capacity to recognize and analyze its surroundings, facilitating automatic navigation and obstacle avoidance These technologies also enable the detection of areas requiring cutting and optimize the cutting path for improved efficiency.

Advanced Sensor Technology: Using modern sensors such as LiDAR, cameras, ultrasound, GPS, etc., to enhance environmental awareness, ensuring the safety and operational efficiency of the robot

Enhance your remote connectivity and control capabilities with innovative mobile applications and cloud platforms designed for seamless robot management These tools empower users to effortlessly monitor and control robots from anywhere, while also enabling the collection and analysis of operational data to optimize product performance.

Energy and Performance: Researching new energy solutions such as high-capacity lithium-ion batteries, wireless charging, solar energy, etc., to extend operating time and minimize environmental impact

Specialized Lawn Mowing Robots: Developing specialized robot models for different environments and applications such as golf courses, parks, industrial areas, and farms

Recent studies have led to substantial enhancements in lawn mowing robot technology, paving the way for future applications and significantly advancing the robotics and automation industry.

Although there are no official statistics, indirect sources indicate that the demand for lawn mowers in Vietnam is on the rise This trend is driven by several factors:

• Increasing aesthetic and landscaping needs of the population

• Development of urban areas, industrial zones, golf courses, etc

• Trend towards automation in agriculture and gardening

• More affordable lawn mower prices

Vietnam's economy is rapidly growing, driving a significant increase in the demand for robots This trend highlights the country's commitment to advancing robotics and high technology, which aims to lower labor and operational costs while enhancing business competitiveness Robots are gaining traction across various sectors, including mobile device production, beverages, food, consumer goods, and manufacturing industries.

Given these potentials, robots have become a topic of interest for researchers in Vietnam in general and educators in particular Several studies related to lawn mowing robots have been conducted:

Table 1.2: Research Topics on Lawn Mowing Robots in Vietnam

NO Research Topic Author Year Content

Design, manufacture, and control of lawn mowing robots

TS Cái Việt Anh Dũng 2012

Design and manufacture of remote-controlled multi- terrain lawn mowers

Vietnam possesses significant potential for developing lawn mowing robots, driven by its skilled technical workforce, low production costs, and an increasing demand for automation in agriculture and landscaping Nonetheless, the industry encounters various challenges that need to be addressed.

• Technological Limitations: Domestic research mainly focuses on simple robot models using basic technology Accessing and mastering advanced technologies such as

AI, machine learning, modern sensors, etc., remains challenging

• Lack of Investment: Research and development projects for lawn mowing robots often struggle to attract investment, especially from domestic resources

• Underdeveloped Market: The market for lawn mowing robots in Vietnam is still small and not effectively exploited Consumers lack awareness and experience in using these products

To advance the research and development of lawn mowing robots in Vietnam, it is essential to foster strong collaboration among universities, research institutes, businesses, and the government Key initiatives should focus on enhancing partnerships, sharing knowledge, and leveraging resources to drive innovation in this emerging technology sector.

• Increase investment in research and technological development

• Train high-quality human resources

• Develop supportive policies and encourage businesses to produce and market lawn mowing robots

These steps will help overcome current challenges and unlock the potential for significant advancements in the field of robotics in Vietnam, particularly in the application of lawn mowing robots.

Purpose and Objective of the Project

The "Design and Remote Control of Lawn Mowing Robot" project focuses on reducing labor accidents associated with manual lawn mowing in Vietnam, where traditional methods pose significant risks By developing an automated and remotely controlled robot, the initiative aims to enhance worker safety, minimize hazards, and improve productivity in lawn maintenance.

To achieve the above purpose, the project sets out the following specific objectives:

Conducting thorough research on cutting-edge technologies and solutions in lawn mowing robots is essential for designing an optimal system tailored to the specific conditions and requirements of Vietnam.

Manufacturing and Testing: Manufacture a prototype of the lawn mowing robot based on the researched design, and conduct tests to evaluate the robot's performance in various environmental conditions

Remote Control: Develop a reliable and user-friendly remote-control system, allowing users to monitor and control the robot conveniently and safely

Performance Optimization: Optimize control algorithms and the robot's operating mechanism to enhance mowing efficiency, save energy, and ensure the system's durability

Practical Application: Implement field trials of the lawn mowing robot in real-world settings, evaluate its effectiveness, and assess its potential for widespread application in the future.

Research Methods

Data analysis involves using tools such as Matlab and QT Creator to collect and observe data, generate response graphs, and assess system outcomes By referencing scholarly articles from Google Scholar, the accuracy of the results can be validated effectively.

Simulation Analysis: Use Matlab Simulink to simulate the inverse kinematics solution algorithm using the Jacobian method and the remote-control system on a two-wheeled moving robot model

Experimental Analysis: Based on simulation results, make observations and comments in practical experiments, comparing experimental results with theoretical knowledge learned from articles

System Effectiveness Criteria: Includes checking the tracking error of the robot compared to the desired trajectory and assessing control quality based on data collected from the model.

Project Limitations

The project "Design and Remote Control of Lawn Mowing Robot" is conducted within the scope of a student project, hence it has certain limitations

The use of budget-friendly GPS modules in the project may lead to increased positioning errors, impacting the robot's accuracy in movement and mowing tasks compared to systems equipped with high-end GPS technology.

The accuracy of sensors, including ultrasonic and infrared types, is limited due to their low cost, which may hinder the robot's capability to effectively detect and avoid obstacles.

The quality of components and materials used in robot manufacturing can significantly impact its durability and operational performance, especially when budget constraints limit the selection of high-grade resources.

• Control Algorithms: The control algorithms developed in the project may not be optimal, leading to less smooth and efficient robot operations

• Remote Control Software: The remote-control software interface may be simple and not integrated with many advanced features

• Terrain: The project focuses on testing the robot on relatively flat and simple terrain

Testing on more complex terrains (such as slopes, many obstacles, etc.) might be beyond the robot's capabilities

• Grass Types: The project mainly tests the robot on common and easy-to-cut grass types Cutting harder, thicker, or unevenly grown grass might pose challenges for the robot

The project successfully designs and manufactures a lawn mowing robot that operates autonomously and can be controlled remotely, enhancing mowing efficiency and reducing labor-related accidents Additionally, the identified limitations will inform future research and improvement strategies.

Content introduction

This article provides a comprehensive overview of robotics, covering key concepts such as robot kinematics, trajectory planning, and the principles of LoRa communication networks It also explores the theories behind remote-control systems, force feedback methods, and distributed control techniques, highlighting their significance in enhancing robotic functionality and performance.

Presentation on calculating and verifying the kinematics for the joystick device

(console) used to control the remote lawn mowing robot and the kinematics of the lawn mowing robot using algebraic methods and the Jacobian method

This article provides a comprehensive overview of a simulation system that effectively simulates and verifies the Jacobian method for solving the inverse kinematics problem in robotics It outlines the necessary steps to construct a remote control system and demonstrates the simulation of this system both with and without force feedback, highlighting the implications for robotic control and interaction.

Chapter 5: System design and construction:

This article provides a comprehensive overview of the development of a lawn mowing robot, detailing the design and construction processes involved in creating the robot model, establishing a LoRa communication network, and developing control circuits Additionally, it covers the construction of monitoring and data collection interfaces, highlighting the integration of these components into a cohesive system.

Chapter 6: System experimentation and evaluation:

Presents an overview of the experimental and evaluation process, experiments and evaluates the drive method of the lawn mowing robot, experiments and evaluates the remote robot control method

Chapter 7: Conclusion and development directions:

Presents the conclusions of the conducted topic and possible directions for developing the topic in the future.

THEORETICAL FOUNDATION

Robot and Mobile robot overview

Robots are automated machines designed to replicate human activities, primarily to take on complex and hazardous tasks These programmable devices execute a variety of intricate operations through computer control and electronic circuits, utilizing electric motors for their rotary movements, making them ideal for light load applications Their key advantages include the ability to perform repetitive tasks with high speed, precision, and durability As a result, robots are commonly employed in assembly lines, production processes, repairs, and operations in challenging, toxic, or dangerous environments.

Mobile robots are robots capable of moving freely in their surroundings They are equipped with sensor systems, actuators, and controllers to autonomously perform tasks

DEPARTMENT OF AUTOMATIC CONTROL 12 such as navigation, obstacle avoidance, positioning, and environmental interaction Mobile robots have diverse applications in industries, agriculture, services, and scientific research

Wheeled robots, including two-wheeled, three-wheeled, and four-wheeled variants, are the most prevalent type of mobile robot, known for their simplicity and efficiency in movement.

• Tracked Robots: Use tracks to move over rough and uneven terrains

• Legged Robots: Use legs to move, mimicking the movement of animals or humans

• Flying Robots: Use rotors or jet engines to fly in the air

• Swimming Robots: Use fins or thrusters to move underwater

Main Components of Mobile Robots:

• Sensors: Provide information about the surrounding environment, including distance sensors, collision sensors, light sensors, cameras, GPS, etc

• Actuators: Perform the movements of the robot, including motors, wheels, tracks, legs, rotors, etc

• Controller: Processes information from the sensors and makes control decisions for the actuators

• Industrial: Transporting goods in factories and warehouses; inspecting and maintaining equipment in hazardous environments

• Agricultural: Irrigation, pesticide spraying, harvesting crops

• Services: Vacuuming, mopping, serving in restaurants and hotels

• Scientific Research: Exploring hazardous environments, space exploration

• Navigation and Obstacle Avoidance: Developing intelligent navigation algorithms to help robots move safely and efficiently in complex environments

• Human-Machine Interaction: Creating intuitive and easy-to-use human-machine interfaces that allow people to interact with and control robots easily

• Artificial Intelligence (AI): Applying AI to enhance robots' self-learning, adaptability, and decision-making capabilities

• Energy: Developing efficient and durable energy sources to enable robots to operate for extended periods without recharging

Mobile robots are essential in today's world, offering significant advantages to humanity As technology evolves, these robots are expected to grow in intelligence and adaptability, effectively addressing a wide range of human requirements.

Robot kinematics

The kinematics of a two-wheeled mobile robot is essential in robotics, emphasizing the analysis and control of its movement This field relies on mathematical equations that illustrate the relationship between control parameters, like wheel angular velocity, and the robot's actual movement in space.

A two-wheeled mobile robot features two fixed wheels on an axle and a balancing free or caster wheel The independent control of the main wheels enables versatile movement and the ability to turn in place The kinematics of the robot is described using a differential geometric model, with key state variables including its position (x, y) and orientation (θ) within a plane coordinate system.

The basic kinematic equations of a two-wheeled robot can be expressed as:

• v is the linear velocity of the robot

• ω is the angular velocity of the robot

• x˙ and y˙ are the rates of change of position along the x and y axes

• θ˙ is the rate of change of the orientation angle

The robot's linear velocity (v) and angular velocity (ω) are regulated by the rotational speeds of its left (ωL) and right (ωR) wheels, establishing a direct relationship between these velocities.

• r is the radius of the wheels

• L is the distance between the two wheels

The inverse kinematics of a two-wheeled mobile robot involves calculating the required wheel angular velocities to achieve a specific movement in space This process contrasts with forward kinematics, where control parameters are utilized to determine the robot's actual movement.

In a two-wheeled mobile robot, inverse kinematics aims to calculate the rotational velocities of the left (ωL) and right (ωr) wheels, which are essential for achieving the desired linear velocity (v) and angular velocity (ω) of the robot.

To achieve this, we use the basic kinematic equations mentioned earlier, but in their inverse form:

• v is the desired linear velocity of the robot

• ω is the desired angular velocity of the robot

• r is the radius of the wheels

• L is the distance between the two wheels

The equations enable precise calculation of the necessary rotational speeds for each wheel, ensuring the robot achieves targeted linear and angular velocities This capability is essential for guiding the robot along designated paths, executing tasks with high accuracy, and maneuvering through intricate environments.

To achieve straight movement at a constant velocity in a robot, set the angular velocity (ω) to zero, resulting in equal rotational velocities for both wheels (ωL=ωR=v/r) In contrast, to enable the robot to turn in place, set the linear velocity (v) to zero, causing the wheels to rotate in opposite directions (ωL/2r).

Understanding inverse kinematics is fundamental for developing feedback control algorithms, enabling the robot to automatically adjust wheel velocities to follow a desired trajectory, improving accuracy and stability during movement.

Theory of PID Controller

PID (Proportional-Integral-Derivative) is a popular feedback control algorithm utilized in automatic control systems It functions by assessing the difference between the current control variable and the desired target value, making necessary adjustments to enhance stability and accuracy.

- The proportional component calculates the control signal based on the difference between the current value of the control variable and the target value

- If the difference value is large, the control signal increases to a significant level

- The integral component calculates the control signal based on the total amount of integration error during the control process

- This component helps overcome temporary errors and helps achieve long-term accuracy

- If the total amount of integration error is large, the control signal will increase or decrease in the appropriate direction

- The derivative component calculates the control signal based on the rate of change of the control variable value

- This component helps predict the changing trend of the control variable and stabilize the system

- If the speed changes are large, the control signal will be adjusted to limit spikes

The PID algorithm calculates the sum of the above three components to create the final control signal The overall formula of the PID algorithm can be expressed as follows:

- u(t) is the last control signal injected into the system

Kp, Ki, and Kd are essential tuning coefficients in a PID system, crucial for optimizing its response These coefficients are identified through rigorous testing and fine-tuning processes to ensure optimal performance tailored to the specific system's needs.

- e(t) is the difference between the current value of the control variable and the target value Global positioning system GNSS – GPS

Overview of GNSS - GPS

Global Navigation Satellite System (GNSS)

The Global Navigation Satellite System (GNSS) is a network of various satellite systems from different countries worldwide Current satellite systems include:

• GPS (USA): The first and most widely used global navigation system

• GLONASS (Russia): The Russian navigation system

• GALILEO (Europe): The European Union's navigation system

• IRNSS (India): India's navigation system

• BEIDOU (China): China's navigation system

• QZSS (Japan): Japan's navigation system

GNSS satellites orbit the Earth, continuously sending signals to ground-based receivers, which allows for precise location determination globally Despite their independent development and varying technical specifications, these systems share the common goal of accurately determining the position, velocity, and time of connected antennas The applications of GNSS are vast and impactful.

GNSS applications are diverse and include:

• Aviation: Supports accurate navigation and landing for aircraft

Figure 2.3: GNSS in the world

• Maritime: Provides positioning and navigation for ships and boats

• Agriculture: Assists in activities such as precise planting, irrigation, and harvesting

• Emergency Services: Helps locate and coordinate rescue and emergency operations

GNSS significantly enhances safety, efficiency, and convenience in daily life

Concepts of Positioning and Navigation

Positioning refers to the capability of identifying the location of an object within a defined space using a specific coordinate system This process relies on signals from GNSS satellites to accurately calculate the position of ground-based receiving devices.

• Navigation: The ability to guide an object to move through space from point A to point B To navigate an object, its coordinates must first be determined

The Process of Positioning and Navigation

• Receiving Satellite Signals: The device receives signals from at least four GNSS satellites to determine its position

• Calculating Position: The device uses signal transmission time and satellite positions to calculate its coordinates on the Earth's surface

• Determining the Route: Using the current position and the destination point to calculate the optimal travel route

• Navigating: Provides detailed guidance to move from point A to point B, including real-time adjustments to avoid obstacles and optimize the route

Thanks to GNSS, positioning and navigation have become more accurate and convenient, effectively supporting various fields of life and production

The Global Positioning System (GPS) is a satellite-based navigation system managed by the United States government and operated by the U.S Space Force It offers geolocation and time information to GPS receivers located anywhere on or near the Earth, provided there is a clear line of sight to at least four GPS satellites Notably, GPS does not require users to send any data and functions independently of telephonic or internet services, although these technologies can improve the effectiveness of GPS positioning.

Figure 2.4: Overview of the GNSS

Figure 2.5: Illustration of GPS satellite

1 Space Segment: Consists of a constellation of at least 24 operational satellites that orbit the Earth and transmit one-way signals giving the current GPS satellite position and time

2 Control Segment: A global network of ground facilities that track the GPS satellites, monitor their transmissions, perform analyses, and send commands and data to the constellation

3 User Segment: The GPS receivers and the user community Receivers are available in a wide variety of formats, from hand-held devices to units integrated into cars, phones, and other devices

Triangulation is a key process in GPS technology, where the receiver calculates its location in three dimensions—latitude, longitude, and altitude—by measuring the time it takes for signals to travel from at least four satellites By knowing the precise positions of these satellites and determining the distance to each one, the GPS receiver can accurately pinpoint its location.

• Time Synchronization: GPS also provides highly accurate time synchronization, which is crucial for many applications such as telecommunications and network synchronization

• Navigation: Used in aviation, maritime, and land navigation to provide accurate positioning and directions

• Surveying and Mapping: Used in geodesy, land surveying, and mapping to create accurate maps and models

• Timing: Provides precise time reference needed for various applications including financial transactions, mobile communications, and power grid management

• Emergency Services: Assists in locating individuals in distress and dispatching emergency services efficiently

• Global Coverage: Provides accurate position information anywhere on Earth

• High Accuracy: Capable of providing position information with an accuracy of a few meters or even centimeters with advanced techniques

• Reliability: Functions continuously, 24/7, under all weather conditions

2.4.3 Determine the object’s coordinates from the GPS system

Latitude is the angle between a point on the Earth's surface and the plane of latitude that passes through the origin, with parallels of latitude represented as concentric circles The equator serves as the base parallel at 0° latitude, while latitude values range from 0° at the equator to 90° N at the North Pole and 90° S at the South Pole Points sharing the same latitude lie on the same latitude line, which is essential for determining global positions in geographic coordinate systems and technologies like GPS.

Longitude is the angular measurement of a location on Earth in relation to the meridian plane that passes through it and the prime meridian A meridian connects the North and South poles, intersecting the equator at a right angle Locations sharing the same longitude create lines that run parallel to each other.

The prime meridian, located at 0° longitude and passing through the Royal Greenwich Observatory, serves as a global reference point Conversely, the antipodal meridian is marked by longitudes of 180° west and -180° east, illustrating the Earth's movement from east to west and vice versa.

5.3.2.4 Method for determining the coordinates of a point

An ellipsoid, also known as a rotated sphere, is a key concept in spatial geometry and geography This three-dimensional shape resembles a sphere but is elongated or flattened along various axes The surface of an ellipsoid is characterized by three primary semi-axes of varying lengths, which determine its unique size and shape.

The Ellipsoid coordinate reference system has the following components:

• Origin O coincides with the geometric center of gravity of the reference ellipsoid

(usually close with the center of the earth)

• The Z axis coincides with the major axis of the ellipsoid along the minor semi-axis

(or the earth's axis of rotation)

• The X axis is the secant of the prime meridian and the equatorial plane

• The Y axis is perpendicular to the OXZ plane according to the principles of the

In geography and earth science, ellipsoids are utilized to accurately represent the Earth's shape Unlike a perfect sphere, the Earth is an ellipsoid, reflecting its true form.

Figure 2.6: Geodetic coordinate system and geocentric coordinate system

The reference ellipsoid serves as a crucial model for approximating the Earth's surface shape, defining the equator and meridians based on its parameters Coordinates on Earth are typically represented using latitude and longitude derived from this ellipsoidal framework.

The parameters of the reference ellipsoid are essential for navigation and cartographic systems, including the WGS84 geographic coordinate system (World Geodetic System 1984), enabling accurate and consistent location and measurement of positions on the Earth's surface.

Research indicates that the coordinates of a point can be determined by calculating the distance covered by one degree when altering objects, resulting in changes to both longitude and latitude.

Determine the coordinates of a point using equations (2.2), (2.3)

11132.92 559.82cos(2 ) 1.175cos(4 ) 0.0023cos(6 )−  +  −  (2.2) 111412.87 cos( ) 93.5cos(3 ) 0.118cos(5 ) −  +  (2.3)

• (2.2) is the length in meters when the latitude changes along the North - South direction, the unit is m/Deg

• (2.3) is the length in meters when the longitude changes along the North - South direction, the unit is m/Deg

For the current Vietnamese coordinate system, WGS-84 is positioned in accordance with the determined parameters: major semi-axis a= 6,378,137 m; flatness

298.257223563 a f = ; rotation speed around the axis wr92115 10 − 11 rad s/ ; Earth's gravitational constant

The National Coordinate Origin (N00) is situated at the Institute of Cadastral Research on Hoang Quoc Viet Street in Hanoi Each province utilizes a specific sutra, with individual axes and projection zones set at 3 degrees, applying a coefficient of k=0.9999.

Overview LORA

LoRa (Long Range) is a wireless communication technology designed for long- range, low-power, and low-data-rate applications It is a popular choice for Internet of

The Internet of Things (IoT) is increasingly leveraging Chirp Spread Spectrum (CSS) technology for deployments, as it allows devices to connect over long distances while consuming minimal energy Operating in unlicensed frequency bands, CSS ensures reliable communication even in challenging, noisy environments.

• LoRa Devices: These include sensors, actuators, and other IoT devices equipped with LoRa transceivers These devices can transmit and receive data over long distances

• LoRa Gateway: Gateways receive LoRa signals from multiple devices and relay this information to the network server via higher bandwidth networks like Ethernet or cellular

• LoRa Network Server: This component manages the network, processes data, and handles authentication and device management It ensures that data from devices is routed correctly and efficiently

• Application Server: The final destination for data, where it is processed, analyzed, and used to trigger actions or provide insights

• Long-Range Communication: LoRa can communicate over distances up to 15 km in rural areas and 2-5 km in urban environments, making it ideal for wide-area applications

• Low Power Consumption: Devices using LoRa can operate for years on a single battery due to the technology’s low power requirements

LoRa technology is optimized for low data rate applications, with speeds typically between 0.3 kbps and 50 kbps This makes it ideal for transmitting sensor data and other low-bandwidth information, ensuring efficient communication in various IoT scenarios.

• Scalability: LoRa networks can support millions of devices, making it suitable for large-scale IoT deployments.

Connection protocol

UART (Universal Asynchronous Receiver-Transmitter) is an asynchronous sequential data transmission protocol widely utilized in electronic applications It enables the sequential transmission of bit data from a transmitter to a receiver using two key pins: TX (Transmit) and RX (Receive) Unlike other communication methods, UART does not require a shared clock signal; instead, it transmits data at a fixed rate, known as the baud rate, allowing for synchronization and comprehension of the transmitted data between devices.

To perform UART communication between two devices, you need to ensure that the devices have compatible communication configurations, including:

• Baud rate: The data transmission speed between the two devices must match Baud rate defines the number of data bits transmitted per second For example: 9600, 115200 baud

• Number of data bits (Data bits): Determines the number of data bits transmitted in each data frame Typically, 8 data bits are used

• Parity (Error checking): Error checking to detect data transmission errors There can be no error checking (None), even error checking (Even) or odd error checking (Odd)

• Stop bits: Determines the number of stop bits used after data transmission Typically,

After configuring the UART for both devices, data transmission occurs through the TX and RX pins Each byte is sent sequentially, starting with the least significant bit (LSB) and followed by the most significant bit (MSB).

UART communication is often used for relatively simple data transfer between devices, e.g., sensor information transmission, device control, communication between microcontroller and computer, and many other applications

SPI (Serial Peripheral Interface) is a widely-used serial communication protocol that facilitates communication between microcontrollers and peripheral devices like sensors, displays, and flash memory It is specifically designed for fast and efficient data transfer in embedded and electronic systems.

PI utilizes a master-slave model where a master device oversees multiple slave devices This protocol relies on four primary wires for data transmission and control.

• MOSI (Master Out Slave In): Wire transmits data from master to slave

• MISO (Master In Slave Out): Wire transmits data from slave to master

• SCLK (Serial Clock): Clock signal generated by the master to synchronize data transmission

• SS (Slave Select) or CS (Chip Select): Signal selects a specific slave device for communication

In an SPI (Serial Peripheral Interface) system, data is sent as serial bits over the MOSI (Master Out Slave In) and MISO (Master In Slave Out) lines The master device generates clock pulses on the SCLK (Serial Clock) line, allowing for the transmission of data bits from the master to the slave via MOSI, and from the slave back to the master through MISO, if feedback is necessary The SS (Slave Select) or CS (Chip Select) signal is utilized to select and activate the appropriate slave device.

• SPI is capable of transmitting data quickly, with transmission speeds up to several tens of MHz

• Simple structure with few wires makes it easy to connect and communicate with many peripheral devices

• SPI has low latency in transmitting and receiving data, suitable for applications that require real time

• Each slave device requires a separate SS wire, which may result in the need to use more wires in a multi-slave system

• SPI does not have a mechanism to acknowledge data transfer, so error checking is required at the application level if necessary

I2C (Inter-Integrated Circuit) is a widely used two-wire serial communication protocol that facilitates communication between microcontrollers and various peripheral devices, including sensors, LCD displays, and EEPROMs Developed by Philips Semiconductor in the 1980s, I2C has established itself as a standard for connecting electronic devices in embedded systems.

I2C uses two main wires to transfer data and synchronize between devices:

• SDA (Serial Data Line): Serial data transmission line

• SCL (Serial Clock Line): Clock signal transmission wire

I2C operates on a master-slave model, in which the master device controls data transmission and one or more slave devices respond

• (Start Condition): The Master will pull the SDA wire from high to low while SCL remains high to signal the start of communication

In data transmission, the master device generates clock pulses on the SCL line and transfers data bit by bit on the SDA line Each byte, consisting of 8 bits, is sent with an acknowledgment bit from the slave device following its transmission.

• (Stop Condition): The master pulls the SDA wire from low to high while SCL remains high to signal the end of communication

In an I2C system, each slave device is assigned a unique address, enabling the master to identify and communicate with specific devices Typically, these addresses are either 7 bits or 10 bits in length, facilitating the connection of multiple slave devices on a single I2C bus.

• Only two wires SDA and SCL are needed to transmit data between multiple devices, helping to reduce the number of wires and circuit complexity

• Supports multiple devices on the same bus through addressing, easily expanding the system

• Allows multiple host devices to share the I2C bus, creating flexibility in system design

• Compared to protocols such as SPI, I2C has lower data transfer rates, typically only reaching a few hundred kHz

• The system may experience problems if two devices have the same I2C address, leading to conflicts

• Because only two wires are used, the signal on I2C is susceptible to noise, especially in complex electromagnetic environments.

Software used

SolidWorks, developed by Dassault Systèmes, is a leading 3D design software widely utilized in mechanical engineering It provides a comprehensive set of reliable tools and features that enable users to create precise and intricate 3D models.

SolidWorks is a powerful software that enables users to design a wide range of 3D parts and models, utilizing advanced tools for shaping, assembly, trimming, and rendering Additionally, it facilitates the creation and management of detailed technical drawings, offering features like automatic drawing generation and accurate representation of dimensions and specifications.

SolidWorks boasts a user-friendly interface that enables intuitive creation and editing of 3D models, allowing users to easily adjust parameters and conduct plausibility checks within the software Additionally, it offers simulation capabilities to test product features prior to actual production, ultimately saving time and resources.

Matlab, developed by MathWorks, is a robust programming and numerical computing environment extensively utilized in science, engineering, and technology It offers a comprehensive set of tools and functions for numerical calculations, data analysis, and simulations, making it an essential resource for professionals in these fields.

Matlab enables users to develop and run programs for numerical calculations, simulations, and solutions to complex challenges across various disciplines, including mathematics, physics, electrical engineering, and signal processing The software offers extensive support for diverse data types and graphical representations, facilitating seamless numerical operations, graph creation, and statistical analysis.

Matlab is a versatile programming language that enables users to develop and tailor their own programs, functions, and algorithms Additionally, it facilitates integration with external tools and libraries, enhancing the software's functionality and performance.

Arduino IDE (Integrated Development Environment) is a software with an open-source code, used mainly to write and compile code into Arduino hardware boards

The Arduino IDE is a versatile and user-friendly platform designed for electronic application development and embedded programming Its popularity among beginners makes it an ideal starting point for those new to programming, while its robust features also cater to the needs of professional developers working on Arduino-based projects.

Figure 2.10: Software Arduino IDE 2.7.4 Visual code

Visual Studio Code (VS Code) is a widely-used open-source integrated development environment (IDE) created by Microsoft, offering a variety of features designed to enhance the coding experience for developers.

Visual Studio Code (VS Code) is a versatile, cross-platform integrated development environment (IDE) compatible with Windows, macOS, and Linux It offers essential features including syntax checking, code hinting, error detection, source code editing, and version management Supporting a wide range of programming languages and frameworks, VS Code is ideal for developers working with popular languages such as JavaScript, Python, and C++, as well as frameworks like Node.js and React.

VS Code excels in web application development with JavaScript and TypeScript, offering features such as code hints, error detection, and autocompletion Additionally, it seamlessly integrates with popular tools like Node.js and Angular, enhancing the development experience.

- Python: VS Code provides powerful features for Python development, including syntax checking, code hinting, debugging, test visualization, and integration with environment management tools like pipenv and conda

- C++ and C#: VS Code supports C++ and C# application development It provides features such as code hinting, debugging, debugging, and integration with compilers and toolkits such as gcc, clang, and NET Core

Several VS Code extensions enhance Java development, enabling users to efficiently write and build Java applications while integrating with tools such as Maven and Gradle.

- HTML/CSS and PHP: VS Code supports website development with HTML, CSS and PHP It provides syntax checking, code hints, live browser viewing, and integration with frameworks like Laravel

SYSTEM KINEMATIC CALCULATION

Inverse Kinematics of the Robot

Chain wheels, or sprocket wheels, feature a unique design that sets them apart from traditional round wheels, providing distinct advantages in specific applications Their specialized structure enhances performance, making them ideal for various mechanical systems The benefits of using chain wheels include improved torque transfer, increased durability, and greater efficiency in power transmission, making them a preferred choice in numerous industries.

1 Ability to Move on Difficult Terrain

Chain wheels offer a larger contact area with the ground compared to traditional round wheels, allowing for better weight distribution and reduced pressure per unit area This design minimizes the risk of sinking into soft surfaces such as mud, sand, or snow, making chain wheels an ideal choice for challenging terrains.

- Improved Traction: With their design of multiple links, chain wheels provide more contact points and better grip, allowing vehicles to move easily on rough, steep, or slippery terrain

- Even Weight Distribution: Chain wheels distribute the vehicle's weight more evenly over the ground, enabling the vehicle to carry heavier loads without damaging the ground or the vehicle itself

Chain wheels provide enhanced stability on uneven terrain, making them crucial for military, construction, and agricultural vehicles that must navigate challenging landscapes while ensuring consistent performance.

Chain wheels offer a significant advantage over round wheels when it comes to getting unstuck, particularly in muddy conditions Their larger contact area and superior grip enable them to self-extricate more effectively, reducing the struggle often faced by round wheels.

- High Wear Resistance: Chain wheels are made from high wear-resistant materials, increasing their longevity and durability under harsh conditions

Chain wheels offer advantages such as high durability, making them ideal for harsh conditions, despite their slower speeds and higher maintenance needs compared to traditional round wheels When analyzing the movement of tracked vehicles, their kinematics closely resemble that of differential drive robots, with both utilizing two independently controlled motors per side The direction of movement is determined by the speed difference between the two tracks or wheels, allowing for the application of similar kinematic equations To calculate a tracked robot's average and angular velocity, it is essential to know the speed of each track and the distance between them, enabling effective control of the robot's coordinates in space.

Figure 3.1: a) Coordinate system of the differential drive robot in space; b) Perspective view of the coordinate system from the robot body

- Xi, Yi: Coordinates of the robot relative to coordinate system {I}

- Xr, Yr: x and y axes of coordinate system {R}

- V R ,V L : Velocity of the right and left wheels

- ω: Angular velocity of the robot

- L: Distance from the center of the robot to the center of the steering wheel

- R: Radius of the steering wheel

-  : Angle formed by the orientation of the robot in coordinate systems {R} and {I}

Based on Figure 3.1, we can describe the vectors as follows:

In which the components on the right-hand side are:

Substituting formula (3.2) into (3.1), we obtain:

The components of linear velocities x and y are derived from the coordinate system that includes the positions x and y os sin x vc y v

Substituting the linear velocity (3.3) into (3.4), we obtain:

The angular velocity (ω) of a mobile robot during a turn is calculated by taking the difference between the velocities of the right and left wheels and dividing it by the distance between the axes of the two driving wheels.

From (3.5) and (3.6), the continuous-time kinematic model of the robot in state space can be represented as:

Representation in matrix form yields the matrix of the robot's kinematic model: cos 0 sin 0

To accurately calculate the robot's inverse kinematics, it is essential to take into account both the robot's coordinate system and the wheels' coordinate system in relation to the robot.

- V sli e d : Sliding velocity of the steering wheel

-  B : Angular velocity of the passive roller (if applicable)

- p B : Radius of the roller (if applicable)

-  B : The angle between V sli e d and the direction of y B

From the figure above, we observe that the vector y B is parallel to  B p B cos B Therefore, we conclude that.: cos cos

Following that, considering the vector x B along the same direction as V d ive r and opposite direction to  B p B sin B , we have the following:

Figure 3.2: Coordinate system {B} of the steering wheel

From (3.9) and (3.10), we obtain the general equation as follows: cos sin cos sin tan tan

Consider the coordinate system of the steering wheel relative to the coordinate system of the robot:

Figure 3.3: Coordinate system of the steering wheel {B} projected onto the coordinate system of the robot {I}

From the figure above, we can determine the linear velocity along the vector x B :

By multiplying the rotation matrix of coordinate system {I} by B V B , we obtain the linear velocity of the steering wheel projected onto the {I} axis:

( ) cos -sin cos sin sin cos sin cos cos sin sin cos

Let's assume we have u as the velocity vector along the x I , v as the velocity vector along the y I and the angular velocity r

Figure 3.4: Representation of velocity vectors in coordinate system {I} and {B}

Similarly to Figure 3.3, we can observe the linear velocity of the steering wheel in the robot's coordinate system as another equation: cos -sin sin cos

0 1 d cos sin d sin cos cos -sin

I yi B xi yi B BI B BI xi B BI B BI yi BI BI

From (3.11) and (3.14), we derive the equation as follows: cos sin

From Figure 3.4, we can deduce the relationship between u, v, r and the velocities in the x, y directions, and angular velocity  by using the Jacobian matrix: cos sin 0 sin cos 0

From equations (3.15) and (3.16), we can derive the general equation for the steering wheel velocity: cos sin

1 1 tan 0 1 -sin cos cos sin 0 sin cos 0

-  B : Angle deviation of the passive wheel axis relative to the steering wheel axis

- R: Radius of the steering wheel

- − d yi ,d xi : Distance from the steering wheel axis to the robot axis

-  BI : Angle deviation of the steering wheel axis relative to the robot axis

- u: Linear velocity along the I-axis

- v: Lateral velocity along the I-axis

- r: Angular velocity about the I-axis

Based on the previous Figure 3.1, we can identify the following basic parameters:

0 0 0 yiR xiR BIR BR yiL xiL BIL BL d L d d L d

= = = = − = = So, we plug in the parameters we have found into the formula (3.17):

If we combine the robot's coordinate system with the global coordinate system, we obtain the matrix of the inverse kinematics model: cos sin 0

To address the inverse kinematics challenge for a lawn mowing robot, it is essential to ascertain the velocity components along the x and y axes, along with the rotational angle in the global coordinate system However, the available data typically only includes the robot's position and orientation, leaving the velocity matrix components unknown Consequently, a more sophisticated approach is necessary to effectively solve the inverse kinematics problem for this robotic system.

Solving the inverse kinematics of a lawn mowing robot using the Odometry method

Odometry is essential for determining a robot's position and orientation using wheel encoder data In differential drive robots, this technique tracks wheel rotations to accurately compute the robot's trajectory.

To solve the inverse kinematics for the robot using the Odometry method, we need to determine the following specific conditions:

+ Desired velocity:  d ( )t , for setpoint control,  d ( )t =0

+ Asymptotically (exponentially) stable: t−  , ( )t −  d ( )t or in the other words, t −  , ( )t − 0 (Error model:( )t = d ( )t −( )t )

Steps to calculate inverse kinematics using the Odometry method:

Step 1: By using closed-loop control, assuming that

Step 2: Differentiating with respect to time

Step 3: Choosing the control input as per the relation:

It’s called a computed velocity control, based on this control input vector, we can find the individual wheel velocities of the robot:

The forward kinematics of the robot

Using the formulas from the inverse kinematics section, we can derive the forward kinematics matrix model for the lawn-mowing robot This derivation is based on formula (3.19), leading to the results of cos, sin, and 0.

Utilizing the Odometry method, we calculate the robot's velocity based on its position and orientation, while applying the inverse kinematics model for the steering wheels This approach allows us to determine the linear and angular velocity required for the robot to achieve its target position and orientation effectively.

SYSTEM SIMULATION

Verification of inverse kinematics

Verify the inverse kinematics of the robot by simulating the Odometry method on Matlab and Simulink

Figure 4.1: Inverse kinematics using Odometry method in Simulink

Robot specifications: R=0.112m, L= 0.195m where R is the radius of the wheel and

L is the distance from the center of the vehicle to the center of the wheel

- Initial position of the robot: x=0; y=0;  =0

Figure 4.2: Inverse kinematics response of the robot in Matlab Case 1 a) Robot’s position; b) Robot’s velocity

Figure 4.3: Inverse kinematics response of the robot in Simulink Case 1 a) Angular velocity of the robot’s two wheels on the scope; b) Visualization of the robot’s movement in space

When all positions are set to 0, the robot will remain stationary

- Initial position of the robot: x=0; y=0;  =0

Figure 4.4: Inverse kinematics response of the robot in Matlab Case 2 a) robot's position; b) Robot’s velocity

Figure 4.5: Inverse kinematics response of the robot in Simulink Case 2 a) Angular velocity of the robot’s two wheels on the scope; b) Visualization of the robot’s movement in space

When providing a position x = 5m providing a position where x = 5m and with all other positions equal to 0, robot will move along the x-axis and gradually decrease its speed

- Initial position of the robot: x=0; y=0;  =0

Figure 4.6: Inverse kinematics response of the robot in Matlab Case 3 a) Robot’s position; b) Robot’s velocity

Figure 4.7: Inverse kinematics response of the robot in Simulink Case 3 a) Angular velocity of the robot’s two wheels on the scope; b) Visualization of the robot’s movement

When we set y = 5m and all other positions are 0, the robot will move along a curved trajectory and gradually decrease its speed as it approaches the chosen position

- Initial position of the robot: x=-2; y=3;  =pi/6=30

Figure 4.8: Inverse kinematics response of the robot in Matlab Case 4 a) Robot’s position; b) Robot’s velocity

Figure 4.9: Inverse kinematics response of the robot in Simulink Case 4 a) Angular velocity of the robot’s two wheels on the scope; b) Visualization of the robot’s movement in space

➔ When given a position x d =3;y d = −3; d = −pi/ 3= − 60 When given a position Therefore, we can see that using the Odometry method, inverse kinematics has effectively matched the desired position values that we sent.

Verification of forward kinematics

Verify the forward kinematics of the lawn mowing robot by simulating the inverse kinematics on Matlab and Simulink

Figure 4.10: Forward kinematics in Simulink

In a manner akin to inverse kinematics, the robot's parameters include a wheel radius (R) of 0.112 meters and a distance (L) of 0.195 meters from the vehicle's center to the wheel's center.

- Initial position of the robot: x=1; y=3;  =pi/4

Figure 4.11: Forward kinematics response of the robot simulation in Matlab Case 1 a) position of the robot; b) Velocity of the robot Simulation in Simulink: a) b)

Figure 4.12: Forward kinematics response of the robot simulation in Simulink Case 1 a) Angular velocity of the robot’s two wheels on the scope; b) Visualization of the robot’s movement in space

➔ The robot moves along a circular trajectory

- Initial position of the robot: x=1; y=3;  =pi/4

Figure 4.13: Forward kinematics response of the robot simulation in Matlab Case 2 a) position of the robot; b) velocity of the robot

Figure 4.14: Forward kinematics response of the robot simulated in Simulink Case 1 a) Angular velocity of the robot’s two wheels on the scope; b) Visualization of the robot’s movement in space

The robot navigates in a counterclockwise circular path around its starting point, demonstrating that the forward kinematics accurately aligns with the angular velocity values of the two wheels derived from the inverse kinematics.

SYSTEM DESIGN AND CONSTRUCTION

Overview of the Construction System

Figure 5.1: Block digram of robot

The diagram illustrates a remote-control system with two main components: MASTER and SLAVE

• Control Interface: This is where the user interacts to send control commands The interface can be a computer or a dedicated remote-control device

• Wireless Transmitter Module: This module is responsible for transmitting control commands from the interface to the SLAVE device via radio waves

• Wireless Receiver Module: This module receives control commands from the

• UART (Universal Asynchronous Receiver/Transmitter): This is a serial communication interface between the receiver module and the microcontroller (MCU)

• MCU (Microcontroller Unit): This is the brain of the SLAVE, responsible for processing control commands received from the MASTER

• Executive Block: This block contains components that execute control commands, such as: o Motor: Controls movement o Alarm: Emits warning sounds o Other peripheral devices: Depending on the specific application

1 The MASTER receives control commands from the user through the interface

2 Control commands are encoded and sent to the SLAVE via the wireless transmitter module

3 The SLAVE receives the signal, decodes it, and transfers it to the MCU through the UART

4 The MCU processes the command and controls the components in the executive block to perform the corresponding action.

Design and Construction of the Robot Model

Figure 5.2: Overview of the 3D model of the lawn mowing robot

The 3D model of the lawn mowing robot, created using SolidWorks, features three primary components: the cutting mechanism, the movement mechanism, and the robot body The cutting mechanism incorporates both in-and-out translation and up-and-down movement, while the movement mechanism consists of two driving wheels, two passive wheels, tracks, and a shock absorption system The robot's body is constructed from an aluminum profile frame, which contains the electrical cabinet and two DC motors.

5.2.1.1 Design of the Robot Cutting Mechanism

The in-and-out linear motion mechanism in the cutting mechanism of a lawn mowing robot has several important functions:

Retractable cutting blades offer enhanced protection by minimizing wear and safeguarding against impacts and contact with obstacles when not in use, ultimately extending the blade's lifespan.

The lawn mowing robot prioritizes safety by automatically retracting its cutting blade when not in use or when it detects obstacles, significantly lowering the risk of injury to both people and pets in the vicinity.

The cutting mechanism features a base constructed from a mica sheet and shaped aluminum profiles, housing the systems responsible for the motorized in-and-out translation and vertical movement of the cutting motor.

The mica sheet is designed with dimensions of 230x302 mm (length, width) and a thickness of 10 mm In the front part of the mica sheet (Figure 5.4), a section is cut out to

The Department of Automatic Control has designed a system to facilitate the movement of the cutting motor, allowing it to move in and out seamlessly The mica surface is equipped with screw holes and three additional holes specifically for attaching slide rails and lead screws.

Figure 5.4: 3D model of the mica sheet a) b)

Figure 5.5: 3D Model of the Cutting Motor Box a) 3D perspective view; b) top-down view

The cutting motor section is centrally located within the box, featuring a shaft that connects to the front-positioned cutting blade This motor utilizes a GT20 belt with a 1:3 transmission ratio to enhance the blade's torque Supporting the box's movement along aluminum profiles are V-slot wheels and belt tensioners on either side Additionally, a stepper motor box equipped with a gearbox facilitates the linear movement of the cutting motor through a pulley drive system with a 1:1 transmission ratio.

Figure 5.6: 3D Model of the Stepper Motor Box a) 3D perspective view; b) top-down view

5.2.1.2 Design of the motion mechanism

To facilitate movement over various challenging terrains such as swamps or surfaces unsuitable for conventional wheels, the movement mechanism of the robot is replaced with caterpillar tracks

This motion mechanism consists of a total of 5 components, namely the planetary

A DC motor powers the driving wheels, while passive wheels and tracks enhance movement and stability The shock absorber system is supported by both main and auxiliary support wheels, ensuring optimal performance and comfort during operation.

Figure 5.7: Overview of the 3D Model of the Motion Mechanism a) 3D perspective view; b) top-down view

The motion mechanism features wheels and assist joints that are designed for easy assembly and disassembly Its innovative hinge design enables the simple addition of auxiliary assist joints to the shock absorber in load-bearing areas As the assist joints share similar structures, the construction process for a single load-bearing joint will be detailed below.

Figure 5.8: 3D Model of the Assist Joint a) model; b) detail 1; c) detail 2; d) detail 3

The construction of the assist joint for the shock absorber includes 3 parts, with parts

The connection between parts 1 and 3 is secured by bolts and nuts, while the auxiliary load-bearing joint features a wheel that utilizes the same connection method Additionally, the wheels mounted at the top play a crucial role in supporting the tracks, effectively preventing sagging.

Figure 5.9: 3D Model of the Auxiliary Assist Joint a) auxiliary assist joint; b) upper wheel

The shock absorber and assist joint are secured with two hinge plates, which also attach the aluminum profile to the vehicle body These hinges are fastened together using brass screws, sized according to the thickness of the wheel and assist joint, typically measuring 25 mm based on the designed thickness.

Figure 5.10: 3D Model of the Motion Mechanism Hinge a) 3D perspective view; b) top-down view

The driving mechanism of the tracks relies on a large gear known as the driving wheel Due to the initial wheel not matching the vehicle's size and weight, a larger wheel was engineered The specifications indicate that the pitch of the driving wheel is 8.8 mm, leading to the design of a wheel with 23 teeth, which has a radius of 56 mm measured from the contact surface of the teeth.

Figure 5.11: 3D Model of the Driving Wheel a) 3D perspective view; b) top-down view

The passive wheel plays a crucial role in track systems by maintaining track tension, guiding and supporting the tracks, and minimizing wear and tear Additionally, it supports the suspension system, ensuring the overall efficiency and durability of the track system This functionality is vital for enabling vehicles to move stably and safely across various terrains.

The front wheel features two deep grooves specifically designed to accommodate the track links As the vehicle moves, the rows of teeth engage with these grooves, effectively preventing the tracks from slipping off during left or right turns.

Figure 5.12: 3D Model of the Passive Wheel a) 3D perspective view; b) top-down view 5.2.1.3 Design of the robot body

The fixed frame is constructed from extruded aluminum bars, featuring two primary bars measuring 200x400 mm and extending 550 mm in length, along with an additional bar positioned at the rear.

The Department of Automatic Control has designed a planetary DC motor with a length of 320 mm, accompanied by an electrical box frame measuring 290x240x80 mm The front of the frame features two 100 mm bars and eleven horizontal bars, each 240 cm long, which are utilized for mounting the shaft that facilitates vertical movement.

Figure 5.13: 3D model of the robot frame a) 3D perspective view; b) top-down view

Design and construction of electrical and control systems

The lawn mowing robot control system consists of 3 main parts (3 MCUs):

The Robot Actuator component features a microcontroller unit (MCU) that directly sends pulse outputs to control a variety of motors, including DC motors, stepper motors, servos, and brushless motors It also integrates warning devices and receives signals from limit switches, ensuring efficient data communication from the MCU center.

• Robot Signal Processing part: includes 1 MCU connected to receive signals from the GNSS Chip and digital compass, Lora signal processing and data transmission functions

Figure 5.20: Result in the constructed of body robot structure

• Monitoring and control center: includes 1 MCU connected to the Lora module and computer to receive monitoring data and transmit control commands to the robot

Figure 5.21: Block diagram of electrical system

Figure 5.22: Wiring diagram of the robot

5.3.2 Select equipment for the system

5.3.2.1 Board Arduino MEGA 2560 and MEGA pro 2560

• Size: Larger, suitable for projects that need a lot of space for components and connections

• Connection: Full range of connection pins, including power pins, digital I/O pins, analog pins, serial communication pins, PWM pins, SPI pins, I2C pins

• Application: Suitable for large, complex projects that require a lot of resources and scalability

Figure 5.23: Structure of the Arduino mega 2560

• Size: More compact, suitable for projects requiring mobility and space saving

• Connection: Maintain the important connection pins, but some rarely used pins can be removed to reduce size

• Application: Ideal for compact, embedded projects that require high performance but do not need too many connection pins

Figure 5.24: Pinout diagram of Mega 2560 Pro

Table 5.1 Arduino Mega 2560 and Mega 2560 Pro parameters

Number of I/O pins 54 pins (including 15 PWM pins)

5.3.2.2 Dynamic expansion motor for grass cutting mechanism

The engine functions as the main engine to rotate the lawn mowing mechanism The engine will have a rotating capacity of 1400rpm/V

Here the motor voltage will be supplied by the 20A ES

5.3.2.3 Step motor with speed reduction gear boxes SK4248

Two Step 42 motors with gear reduction boxes enhance the functionality of the grass-cutting mechanism One motor is responsible for adjusting the cutting height, while the other motor controls the extension and retraction of the grass-cutting unit during operation.

Figure 5.27: Step 42 motor with speed reduction

Figure 5.28: Mechanical drawing of the motor

The Planet 30W 600RPM motor with a 12PPR encoder and 8mm shaft is a versatile and efficient motor designed for precise control applications With a power rating of

Operating at a steady 600 revolutions per minute, this 30-watt motor is perfect for tasks that demand consistent and reliable performance Its integrated encoder delivers 12 pulses per revolution, ensuring precise feedback on motor position and speed, making it essential for robotics, automation, and CNC machinery applications.

Figure 5.30: Mechanical drawing of the motor

Table 5.3 DC Planet Motor Parameters

Two TB6600 stepper motor drivers are utilized to manage the pulse supply for two stepper motors, each configured with a microstepping setting of 8 (OFF-ON-OFF) to meet the pulse supply needs of both motors.

Figure 5.32: Stepper Driver TB6600 Wiring diagram

Table 5.4 Stepper Driver TB6600 Parameters

The HI216 H-Bridge circuit is crucial for controlling the direction and speed of DC motors, offering full bidirectional control for forward, backward, and stop functions Engineered to manage high current and voltage demands, the HI216 is ideal for diverse applications such as robotics, automotive systems, and industrial automation Its efficient design enhances performance, making it a reliable choice for motor control tasks.

DEPARTMENT OF AUTOMATIC CONTROL 78 minimal power loss and provides robust protection against overheating and short circuits, ensuring reliable performance in demanding environments

Figure 5.34: Wiring diagram of HI216 H-Bridge

Table 5.5 HI216 H-Bridge Circuit Parameters

5.3.2.7 LM2596 Step Down Buck Converter

The LM2596 step down buck converter is used to step down the voltage from the vehicle's power supply to 5V, providing power for the MCUs, encoders, and other components

Figure 5.35: LM2596 DC-DC Step Down Buck Converter

Table 5.6 DC-DC Step Down Buck Parameters

5.3.2.8 SX1278 433MHz RF LoRa AS32 UART Transceiver Module

Our team utilizes the SX1278 433MHz RF LoRa AS32 transceiver module for signal transmission to the remote controller, thanks to its compact design and user-friendly wireless connectivity This cost-effective solution offers stable performance and is ideal for long-distance communication, achieving a maximum range of up to 3000 meters in optimal conditions with minimal obstructions.

Figure 5.36: LoRa AS32 UART Transceiver Module

Figure 5.37: LoRa AS32 UART Transceiver Module Wiring diagram

Table 5.7 LoRa AS32 UART Transceiver Module Parameters

UART Communication Data bits 8, Stop bits 1

Transmission Rate 0.3 – 19.2 Kbps (Default: 2.4 Kbps)

Figure 5.38: Anten WiFi 4G RF Lora

Table 5.8 Anten WiFi 4G RF Lora Parameters

UART Communication Data bits 8, Stop bits 1 Frequency Range 700 – 2700 Mhz / 433Mhz / 315Mhz

Transmission Rate 0.3 – 19.2 Kbps (Default: 2.4 Kbps)

5.3.2.10 Camera integrated with VTX module

The VTX-integrated camera allows users to remotely control the Lawn Mowing Robot by transmitting real-time video, enhancing user experience and convenience.

Figure 5.39: Camera integrated with VTX module and RTX Module

A reliable robot necessitates a stable power source with high discharge current, making the use of a 4S 14.8V LiPo battery an excellent choice This battery offers advantages such as compact size, lightweight design, optimal performance, and enhanced safety Additionally, it can efficiently power the servo, DC motor control circuit, and microcontroller through a voltage reduction circuit.

Sensors are essential for maintaining balance in lawn mowing robots, as they provide critical data for effective operation Key information collected includes angular acceleration, angular velocity, device orientation angles, ground clearance, and magnetic field data Feedback accelerometers and inertial sensors are among the primary components that facilitate precise control of the robot's movements.

DEPARTMENT OF AUTOMATIC CONTROL 84 accelerometers, which provide essential signals for the robot to maintain stability and ensure efficient control during lawn mowing operations

The HMC5883L sensor, a three-axis magnetometer manufactured by Honeywell, is essential for lawn mowing robots as it measures the Yaw angle and determines the Earth's geomagnetic orientation This digital compass is widely used in positioning and navigation systems, including drones and autonomous robots, thanks to its ability to measure magnetic field intensity and direction accurately It integrates easily with microcontrollers via I2C or SPI interfaces and features noise filters and internal amplifiers to ensure measurement accuracy and stability With its high resolution and low power consumption, the HMC5883L is an excellent choice for projects requiring precise magnetic field measurements and positioning.

Table 5.10 HMC5883L Compass Sensor Parameters

Factors to consider when choosing the right GPS chip for the application:

When selecting a GPS chip, accuracy is paramount, as it significantly enhances the effectiveness of your positioning and navigation system For applications demanding precise location tracking, such as lawn mowing robots, opting for a highly accurate GPS chip is essential to ensure the robot operates as intended.

The GPS chip's update rate refers to how frequently location data is refreshed each second A higher update rate enables devices to swiftly adapt to positional changes, making it crucial for fast-moving applications like drones, autonomous vehicles, and robotic lawn mowers that demand immediate responsiveness.

The sensitivity of a GPS chip is crucial for its performance in various environments, as it affects signal capture A highly sensitive GPS chip excels in challenging conditions, such as dense forests, urban landscapes with tall buildings, or beneath tree cover, ensuring reliable navigation even in areas with weak signals.

Time to first acquisition (TTFF) refers to the duration required for a GPS chip to establish a position upon startup A quicker first signal acquisition time enhances device readiness, allowing for faster operation and significantly improving overall system performance.

GPS chips can enhance positioning and navigation by integrating additional sensors like magnetic field sensors, accelerometers, and gyroscopes This capability is particularly beneficial for applications that demand precise and comprehensive positioning data.

• Power consumption: For battery-powered applications such as drones or autonomous robots, choosing a GPS chip with low power consumption is important to extend device operating time

Build algorithms and control programs

Figure 5.49: Robot signal processing flowchart

5.4.3 Monitor and control center flowchart

Figure 5.50: Monitor and control center flowchart 5.4.4 Image Processing Application

The image processing application aims to halt the robot upon detecting a human figure through its camera To accomplish this, the team employs OpenCV and MediaPipe for effective pose recognition.

OpenCV, or Open Source Computer Vision Library, is a widely recognized open-source library that offers a comprehensive suite of tools for computer vision tasks It enables efficient image and video processing, object detection, and supports a variety of applications in the field of computer vision.

MediaPipe, a framework created by Google, enables real-time multimedia processing and features advanced deep learning models for various applications, including pose, hand, and face recognition It effectively detects and tracks human poses, making it a powerful tool for developers in the field of computer vision.

The application utilizes OpenCV to access and read data from the camera, initiating the default camera with the command cv2.VideoCapture(0) If the camera fails to open, the program will display an error message and exit.

Each frame read from the camera is converted from BGR format (OpenCV's default format) to RGB, then sent to MediaPipe for pose recognition processing

After analyzing the frame, the system checks if a human pose is detected by verifying that results.pose_landmarks is not None If a pose is detected, the robot's operation pauses, and the detection time is recorded Conversely, if no pose is detected within a specified timeframe of 3 seconds, the system resumes normal operation.

Figure 5.51: Flowchart for person detection

Figure 5.52: Draw the line when detected

5.5 Build control interface, monitor and collect data

Design interface and algorithm requirements:

– The interface needs to fully display all robot parameters: current position coordinates, destination coordinates on a map interface

– The interface should be able to calculate the distance between the current position and the destination

– The interface should be able to display information about the position

The software for programming the control interface and data collection utilizes a laptop due to hardware constraints Our team used QT Designer for this purpose

Some features of QT Designer: it is efficient in designing graphical interfaces for applications using the Qt framework This software can design interfaces for

DEPARTMENT OF AUTOMATIC CONTROL 102 applications in various fields It supports diverse programming languages such as C/C++, Python, Java, and has a wide user community

The control interface comprises three main tabs, each with distinct functions described in Table 5.10 below

1 Tab_Intro Introduces the project title, supervising teacher, and team members

2 Tab_Setup Sets up communication ports and transmission speeds

3 Tab_Control Selects control modes for the robot, displays positions, and selects features

Tab_Intro is a page that introduces the project title, presents the supervising teacher, and lists the team members involved in executing the project

Tab_Setup includes functions to set up necessary UART connections for the robot's operation and configure its position

– Area 1: Configure Serial port connection and baud rate

– Area 2: Functions such as opening the camera and map

Tab_Control displays the number of satellites, current GPS coordinates, and destination coordinates of the robot on the map, along with several functions

Figure 5.56: Interface of the map

All the functions in Tab_Control:

– Area 1: Control the robot in two modes, Manual and Auto

– Area 2: Select the cutting blade height relative to the ground

In Area 3, display the robot's current coordinates alongside the destination coordinates Include two essential buttons: the "Clear Coordinate" button to reset the selected positions for the robot, and the "Send Coordinate" button to transmit the selected coordinates to the robot for effective processing and navigation.

SYSTEM EXPERIMENTATION AND EVALUATION

Establishing UART Protocol Using RF UART Lora Module

To operate the robot model, the remote-control system built by the team needs to have high accuracy, ensuring reliability and stability during operation

After configuring and programming the communication protocol, the team will test data transmission and reception using a GPS module to select a real-world destination The overview diagram illustrates the experimental verification of LORA technology.

Figure 6.1: Overview diagram of the experimental verification of the LORA protocol

Following the established data structure for the LORA protocol, the data sent from the computer consists of coordinates selected on the map The team conducted several experiments for verification

By communicating and sending data as strings from the host computer (laptop) to the processing unit, the team checked whether the received string was correctly separated into individual values

Figure 6.2: Connecting the RF UART Lora Module box

The LORA BOX features a Mega 2560 for processing signals between two LORAs, along with an RF UART LORA module After setting up UART communication between the laptop and LORA, a formatted string with semicolon-separated values was transmitted The system was optimized for data packet transmission and reception between the remote control and the robot, leading to tests of the LORA modules' communication range and stability in diverse environments, including open areas and locations with possible signal interference.

Experimental Results

In open areas, transmission speeds were notably quick, exhibiting no interference or delays Under optimal conditions, the transmission range could extend up to 5 kilometers even in the absence of nearby transmission towers.

Figure 6.3: a) Signal transmission and reception experiment in an open space; b) Successful declaration and connection to LORA BOX; c) GPS coordinates sent from the

In densely populated areas, the LORA's operating range is reduced to 30m-50m, resulting in noticeable signal delays This limitation has caused issues with signal transmission, particularly when pressing buttons on the interface or switching control modes.

Figure 6.4: a) Space with many trees and bushes; b) Interference occurring during experiments in a space with many obstacles

Conclusion: Based on the data, the transmitted values exhibited high accuracy, demonstrating the feasibility of LORA technology for the remote-control system

6 2 GPS and Compass Sensor Experiments

The GPS module was integrated with the robot, and the compass sensor was calibrated to provide accurate heading data The system was programmed to continuously

DEPARTMENT OF AUTOMATIC CONTROL 110 track the robot's position and orientation The experimental setup included pre-determined markers and obstacles to test navigation algorithms a) b)

Figure 6.5: a) Setup for real-world observation; b) Selecting a new coordinate point on the interface map

The team conducted LORA experiments to identify the optimal distance and spacing for effective signal transmission between the robot and the laptop As illustrated in Figure 6.5, the coordinate values transmitted from the robot demonstrated high accuracy, attributed to the open space facilitating clearer communication.

GPS receiving signals from multiple satellites The compass sensor accurately read the Earth's magnetic field and converted it to the yaw angle

The accuracy of GPS readings is influenced by the number of connected satellites, with more satellites resulting in improved precision However, obstacles in the sky can hinder the GPS's ability to receive signals, and adverse weather conditions may also impact the accuracy of the readings.

In conclusion, the robot's GPS readings were inconsistent due to its operating environment and weather conditions, whereas the compass sensor functioned properly The team analyzed the discrepancies from the anticipated values and made necessary adjustments to both the algorithms and hardware to enhance overall performance.

The MANUAL mode was tested by controlling the joystick to check its functionalities a)

Figure 6.6: a) Remote control joystick for the robot; b) Switching to MANUAL mode

Figure 6.7: Remotely controlling the robot using the joystick

Conclusion: The joystick control had a relatively long operating range, and functions like adjusting the cutting blade's height and rotation speed could be operated without issues

The AUTO mode was tested by selecting a point on the map and sending the coordinates to the robot for processing and navigation to the chosen location

Figure 6.8: Selecting a location on the control interface

After pinpointing the location on the map, the team utilized the "SEND COORDINATE" button to transmit the coordinates to the robot The robot then processed the received data, assessed the current angle against the target angle, and navigated to the designated point.

Figure 6.9: a) Initial position; b) Adjusting the heading; c) Robot moving towards the designated point; d) Robot reaching the chosen location

In conclusion, the AUTO control program effectively guided the robot to its designated location on the map However, issues arose with overshoots during heading adjustments due to the PID controller, leading to inconsistencies in the trajectory and deviations from the intended path The team assessed these deviations and refined the control algorithms to enhance performance Overall, the successful autonomous operation highlights the viability of the automatic control system.

CONCLUSION AND FUTURE DIRECTIONS

* After conducting research and implementing the project "Design of a model and controller for a lawn mowing robot," the team has achieved the following result:

- Designed a complete model of the lawn mowing robot using Solidworks 2021

- Calculated both forward and inverse kinematics for the lawn mowing robot using the Odometry method

- Integrated a GPS positioning system and compass sensor to determine and adjust the robot's position during operation

- Simulated the movement and validated the kinematics of the lawn mowing robot using MATLAB Simulink 2023b

- Successfully built and validated a remote-control system using LORA technology to transmit data between the robot and the control device

- Successfully constructed the autonomous lawn mowing robot model and control cabinet Developed a user interaction interface using Visual Studio Code for control and data collection

- Experimented with the remote-control system, tested, and verified the accuracy of the robot in mowing based on the compass sensor and GPS positioning

* The limitations encountered during the project implementation are as follows:

- Determining the robot's rotation angle is heavily dependent on hardware, leading to operational errors

- The suspension system for the movement mechanism is not optimized for the vehicle's tracks

- The GPS positioning system is affected by weather and surrounding environment, resulting in low accuracy

Future development directions for the project aim to enhance operational accuracy by utilizing an inertial sensor to determine the robot's initial state Integrating an IMU sensor with the positioning system will support GPS, further improving accuracy and stability Additionally, research will focus on enhancing the energy system by incorporating backup batteries or multiple power sources to ensure continuous robot operation Optimizing the control algorithm will minimize mowing errors, while studying the material and structure of the cutting blades will increase the robot's efficiency and lifespan.

[1] M Họgele, K Nilsson, J N Pires, and R Bischoff, "Industrial robotics," Springer handbook of robotics, pp 1385-1422, 2016

[2] M Yoshida, K Nonami, and M Kumagai, "Development of autonomous lawn mower robot with multiple sensors," in IEEE/RSJ International Conference on Intelligent Robots and Systems, 2005, pp 4351-4356

[3] C C Cho, "Design and implementation of an autonomous robotic lawn mower," in 2011 IEEE International Conference on Mechatronics and Automation, 2011, pp 767-772

[4] N Hirose, "Autonomous lawn mowing robot using vision system," in SICE Annual Conference, 2010, pp 1762-1767

[5] L Merino, J M P G Sánchez, and A Ollero, "A review of recent contributions in multi-robot systems: Applications and future challenges," Sensors, vol 21, no 5, p 1519,

[6] G Gutiérrez, J Olaverri, and L E Díez, "Path planning for autonomous mobile robot using multi-objective optimization," in 2008 IEEE International Conference on Emerging Technologies and Factory Automation, 2008, pp 1047-1050

[7] D M Anderson and J N Martin, "Robot programming: a practical guide to behavior- based robotics," McGraw-Hill Professional, 2007

[8] R Siegwart, I R Nourbakhsh, and D Scaramuzza, "Introduction to Autonomous Mobile Robots," MIT Press, 2011

[9] J K Petersen and J P Thomsen, "Robotic lawn mower with GPS-based navigation," in

2010 International Conference on Consumer Electronics, Communications and Networks (CECNet), 2010, pp 4312-4315

[10] H Takahashi and T Ohtsubo, "Autonomous navigation of a robotic lawn mower using RTK-GPS," in 2013 IEEE International Conference on Robotics and Automation (ICRA),

[11] R Paranjape, A Arkin, and S Chatterjee, "Precision agriculture using robotic lawn mowers," in 2014 IEEE Conference on Technologies for Sustainability (SusTech), 2014, pp 167-172

[12] M Reiser, M Kuderer, and W Burgard, "Using learned behavior models to track human motion in indoor environments," in 2013 IEEE International Conference on Robotics and Automation (ICRA), 2013, pp 3450-3456

[13] E Ackerman and E Guizzo, "Why we don’t have autonomous lawn mowers yet (and how we will)," IEEE Spectrum, vol 53, no 9, pp 34-39, 2016

[14] S Thrun, W Burgard, and D Fox, "Probabilistic Robotics," MIT Press, 2005

[15] J V Miró, A J F P Lorenzo, and A Ollero, "Autonomous robots for precision agriculture: State of the art and future perspectives," Precision Agriculture, vol 13, no 1, pp 1-27, 2012.