Ho Chi Minh City, July 2024 MINISTRY OF EDUCATION AND TRAINING HO CHI MINH CITY UNIVERSITY OF TECHNOLOGY AND EDUCATION FACULTY OF MECHANICAL ENGINEERING GRADUATION THESIS RESEARCH, D
INTRODUCTION
Motivation
The hand is essential for daily activities, including grasping and holding, but many individuals face challenges due to limited motor capabilities Factors such as accidents, chronic diseases, strokes, age-related muscle deterioration, or paralysis can lead to the loss of limbs or impaired hand function This loss significantly impacts daily living and labor, affecting various aspects of patients' lives and their mental well-being.
Supporting patient rehabilitation through direct assistance fosters confidence and proactivity in daily activities While a physical therapy regimen can facilitate quick recovery, it requires substantial economic and human resources However, advancements in technology have made patient support and rehabilitation more accessible and less resource-intensive Numerous studies and available devices offer solutions that provide support in both active and passive forms.
Passive support methods utilize prosthetic limbs that lack control capabilities, making them affordable and visually appealing but functionally limited In contrast, active support methods feature self-operating limbs powered by actuators like motors and artificial muscles, offering enhanced support by responding more flexibly to the user's intentions Key characteristics of these advanced devices include their lightweight design, wearability on the hand, and moderate gripping force, with options for both rigid and soft joint types.
Traditional rigid structures made from metals and hard plastics provided durability and precise movement but were often stiff and cumbersome due to their external frames These limitations hindered the adjustment of finger joints and made controllers heavy and inconvenient for users In contrast, recent advancements in soft materials like silicone and flexible plastics offer lightweight, bendable options that enhance user-friendliness and ease of maintenance This project focuses on developing devices with soft structures, which not only improve flexibility and durability but also eliminate alignment issues by creating jointless designs, making them ideal for healthcare applications.
Figure 1.1: The rigid exoskeleton has many disadvantages
Soft mechanisms have been extensively studied, with pneumatic, hydraulic, and tendon-driven methods being the most prevalent Pneumatic and hydraulic systems provide significant flexibility and ease of use, yet they face challenges such as potential leakage and the need for specialized chamber designs to effectively manage air or fluid pressure.
Tendon-driven mechanisms offer a simpler design for collaboration with soft gloves, effectively managing the pulling forces from mechanical components while minimizing operational malfunctions This transmission method is also more cost-effective in both design and fabrication compared to pneumatic or hydraulic systems Furthermore, gloves utilizing tendon-driven mechanisms can be designed to be more compact and straightforward than those powered by hydraulic or pneumatic systems.
In this thesis, the authors present an innovative assistive glove featuring a tendon drive mechanism and friction wheels to manage tendon slack during operation This project emerged from a Mechatronics initiative focused on resolving tendon slack and eliminating initial tension Users can customize their grip strength using an encoder knob, while the system maintains a safe contraction level by utilizing data from flex sensors on the fingers.
The authors developed an assistive glove designed specifically for the index and middle fingers, allowing the thumb to remain free for enhanced comfort The glove features a unique transmission mechanism that isolates tendon coils and wires, enabling comfortable movement when the system is inactive To enhance safety, two flex sensors are integrated to monitor the index and middle fingers, preventing them from entering dangerous flexion or extension zones The system operates with a single DC motor that mimics natural hand movement, ensuring that deep contractions of the index or middle finger do not limit the motion of the other fingers, thus avoiding discomfort for the user.
The authors conducted experiments to assess the device's ability to effectively contract and extend fingers It demonstrated impressive stability while gripping cylindrical objects, enabling users to hold items in diverse ways that align with their individual hand shape preferences.
Objectives
- Design, fabricate, and test a soft glove suitable for the target users, with a focus on safety, comfort, and convenience
- Research, design, fabricate, and test a transmission system that uses only a single motor to control the position of two primary fingers, the index and middle fingers
- Evaluate the hand's ability to flex/extend and grip test objects to collect data and evaluate the system's effectiveness.
Research subjects and scope
This thesis focuses on an innovative assistive glove designed for patients with myasthenia, utilizing flex sensors as feedback signals for its control system This technology enables users to adjust the level of finger flexion and extension according to their individual preferences, enhancing their mobility and independence.
This research focuses on the development of a tendon-driven mechanism that operates without the need for pre-tensioning, enhancing user safety and comfort by isolating the glove from the mechanism during design, manufacturing, and testing processes.
The integration of flex sensors and filters facilitates the collection of feedback signals essential for motion control systems This research further assesses the force applied to test objects during gripping and holding processes.
Research contents
- Study the anatomy of the human hand Study characteristics and symptoms of patients with myasthenia
- Research tendon-driven existing mechanisms in research and make evaluations
- Learn from the Mechatronics project to form the basis for the development in this thesis
- Understand the operating principles of flex sensors and develop programs to read and filter signal noises using Butterworth low-pass filters and Kalman filters
- Design and fabricate the actuator mechanism
- Develop the software for the STM32 microcontroller to control the system Monitor device operation through the software ecosystem provided by ST
- Conduct experiments on the device
Structure of the report
This chapter presents the research motivation, objectives of the thesis, research subjects, scope, and contents of the study Finally, it outlines the thesis structure.
LITERATURE REVIEW
Introduction to the device's target users
Myasthenia Gravis is an autoimmune disease where the body produces antibodies against acetylcholine receptors (AchR) at the neuromuscular junctions, leading to a harmful misidentification of these healthy cells This condition affects approximately 5 to 20 individuals per 100,000 in both Vietnam and globally Despite its rarity, Myasthenia Gravis can have severe consequences, resulting in disability or even death from respiratory failure Early diagnosis and aggressive treatment are essential, particularly since the disease often strikes individuals of working age, impacting both personal health and workforce productivity.
In Myasthenia gravis patients, the main impairment occurs at the postsynaptic membrane of neuromuscular junctions, leading to a decrease in both the quantity and quality of acetylcholine receptors This reduction prevents the effective generation of action potentials when acetylcholine is released at the presynaptic membrane, ultimately hindering muscle contraction Research indicates that antibodies produced by the patient's autoimmune response damage the acetylcholine receptors at the postsynaptic membrane.
Myasthenia Gravis is marked by fluctuating muscle weakness that worsens with activity and improves with rest While this condition leads to significant muscle fatigue, it does not affect sensory perception or tendon reflexes, making it essential to differentiate Myasthenia Gravis from other disorders that present similar symptoms.
Myasthenia Gravis often begins gradually over weeks to months, which can make it challenging for patients to identify Nevertheless, some individuals may experience a sudden and severe onset characterized by generalized weakness Various factors, including systemic infections, thyroid surgery, stress, and pregnancy, can trigger or worsen the condition.
Myasthenia Gravis is a systemic disease that primarily affects specific muscle groups, notably the ocular, trunk, and limb muscles The ocular muscles are often the first to be impacted, resulting in double vision Following this, facial muscles may become involved, leading to changes in voice, slurred speech, and difficulties in articulation, as well as challenges with jaw closure that may necessitate manual assistance.
Injuries to the proximal regions of the limbs significantly impact patients' mobility, with lower limb damage affecting their ability to walk, stand, or navigate stairs, while upper limb injuries impede their capacity to grasp objects and carry out daily tasks This article focuses on providing support for patients with upper limb Myasthenia Gravis, particularly concentrating on enhancing hand function.
2.1.2 Overview of human hand anatomy
The hand, a crucial component of the human body, is located at the distal end of the upper limb and is vital for grasping and manipulation This region comprises soft tissue structures that surround the bones and joints of the fingers, extending from the wrist to the fingertips, including both the palm and the back of the hand It plays an essential role in executing daily tasks and complex hand functions.
2.1.2.1 The skeletal system of the human hand
Figure 2.1: The model of the human hand skeletal system
Caption: 1 Distal Phalanx; 2 Middle Phalanx; 3 Proximal Phalanx; 4 Metacarpal
The human hand includes 27 bones, which is divided as: 8 carpal bones in the wrist,
5 metacarpal bones in the palm and 14 phalanges in the fingers and the thumb Among which:
- Metacarpal bones: The palm includes 5 metacarpal bones, each corresponding to a finger Each metacarpal bone has three main parts: a head, a body and a base
The human hand consists of 14 phalanges, which are the bones that make up the fingers The thumb contains two phalanges, lacking a middle phalanx, while each of the other fingers features three phalanges These bones collaborate with the wrist to enhance the hand's flexibility and functionality, enabling a wide range of daily activities.
The hand joints play a crucial role in allowing individuals to execute intricate movements with objects and tools These joints facilitate the hand's ability to grasp, release, and perform various complex gestures, making them essential for dexterity and functionality.
- Interphalangeal articulations: These are the joints between the phalanges of the fingers, they allow flexion and extension of the fingers
- Metacarpophalangeal joints: These joints connect the metacarpal bones to the phalanges, they allow the fingers to flex, extend and rotate around their axis
- Intercarpal articulations: These joints connect the bones within the carpal bones, they help flexible movements of the hand
- Wrist: This joint connects the hand to the forearm, they allow diverse movements such as flexion, extension, rotation and tilting of the hand
2.1.2.2 Overview of the tendon and muscle system for the human hand
The anatomical structure of the palm region comprises skin, tissue, muscles, tendons, blood vessels, and nerves, with a focus on the hand's muscles and tendons essential for glove design and fabrication There are two primary types of muscles in the hand: extrinsic and intrinsic Extrinsic muscles, located in the forearm, facilitate strong but coarse finger movements through their tendons that extend to the fingers In contrast, intrinsic muscles, which originate and insert within the hand, enable delicate and precise finger movements These intrinsic muscles are further categorized into three groups: thenar muscles, hypothenar muscles, and interosseous muscles.
Figure 2.2: The anatomy of the human hand
2.1.3 A research overview of assistive gloves
Previous research on smart gloves has primarily concentrated on capturing finger movement data using bend sensors, often integrating various other sensors to assess hand configuration However, these gloves face challenges related to portability, rely heavily on physical connections for signal collection, and restrict natural hand movements.
Recent research has expanded beyond the challenges of collecting hand data to developing wearable devices that are practical and functional These studies employ a variety of combined methods, incorporating tactile stimuli such as pressure, contact, vibration, and softness, while utilizing different actuators including DC motors, pneumatic actuators, and dielectric elastomers.
Assistive gloves are designed to support essential hand movements, including flexion, extension, grasping, and releasing objects These gloves primarily utilize two types of mechanisms: active and passive structures.
An active mechanism utilizes common actuators like electric motors, pneumatic cylinders, or advanced artificial muscles to deliver dynamic support When integrated with smart processing software, these mechanisms offer enhanced convenience and responsiveness for users.
A passive mechanism relies on manual operations from the user, making it less effective due to the absence of actuators for control While it can provide assistance, its functionality is limited compared to active systems.
The challenges in research and development
2.2.1 The challenges in the transmission mechanism design
Smart assistive gloves must ensure 4 basic characteristics:
- Small, light, and easy to carry
- Comfortable and should not restrict the user's natural hand movements
- The shape and size can be easily adjusted to suit different hands and fingers
- Optimal and simple transmission system with good performance
When considering weight and portability, gloves with passive mechanisms are aesthetically pleasing and simpler in design but may lack user flexibility Conversely, gloves featuring active mechanisms can also be effective if their actuator systems are uncomplicated Research from Seoul National University indicates that a compact actuator system utilizing cables for transmission can fulfill the weight and portability requirements effectively.
When considering wearing comfort and freedom of movement for the human hand, external rigid skeletons are inadequate Instead, materials like fabric, silicone, polymer, or a blend of fabric and silicone offer superior comfort against the skin and facilitate natural movement during use.
Personalization is essential for achieving flexibility in shape and size in devices that utilize active mechanisms, particularly those with thimbles for finger actuation, which necessitate precise design dimensions tailored to each user's hand A straightforward glove fabrication process with minimal specific steps can effectively accommodate individual user sizes Additionally, a streamlined and efficient actuator system is crucial to minimize bulkiness while ensuring the device can facilitate hand grasping and flexion/extension Furthermore, optimizing the number of actuators within the system is vital for enhancing portability.
2.2.2 The challenges in designing control system
For the device's control system, two important criteria are required:
- Safety and reliability when controlling and troubleshooting during operation
- Flexible control based on user permission and input
Ensuring safety requirements and effective troubleshooting methods is essential for devices that impact user health The control system must operate predictably, allowing for easy monitoring of all operational states to facilitate prompt and safe responses in case of incidents.
When utilizing actively controlled devices, the fundamental requirements for control input signals are flexibility and user permission These elements are essential to ensure that the device operates correctly and safely, aligning with the intended operational goals.
RESEARCH, DESIGN, AND FABRICATION OF THE ASSISTIVE GLOVE
Design requirements
The authors of this thesis designed a glove using soft materials and tendon-driven mechanisms with monofilament strings to enhance finger movement This innovative glove facilitates the flexion and extension of the index and middle fingers, while keeping the thumb free for improved comfort and control during object grasping Additionally, the glove features detachable tendon ends for easy cleaning, ensuring convenience for users.
The authors are focused on creating a soft and safe assistive glove designed specifically for patients with myasthenia Drawing from various studies, they have identified key characteristics that the glove must possess, including an optimal size, lightweight material, ease of wear, and skin-friendly properties It is essential that the glove provides comfort and does not restrict natural movements, particularly in the finger joints Furthermore, the design allows for adjustments to accommodate different hand and finger sizes and shapes.
Approach for designing the soft glove
Figure 3.1: The components of the glove
This section presents components and requirements for the glove, including:
(1) Thimble: Requirements: medium hardness, enough to transmit force for flexong or extending, but flexible to ensure comfort
(2a) Straps: Requirements: firm attachment to the finger body for stability and precise guidance, medium flexibility for easy bending according to finger shape
(2b) Upper straps: Requirements: sufficient stiffness, creating triangular prism shapes to securely fix at the upper back of the fingers
String holes must possess high stiffness and thickness to maintain their shape and minimize wear caused by friction The diameter should be carefully calculated to ensure it is not excessively large, which could hinder guidance, nor too small, which would lead to unnecessary friction.
(4) Glove body: Requirements: Soft and flexible to fit various hand sizes comfortably High flexibility adapts to the bending and stretching capabilities of the hand, avoiding balk
(5) Velcro: Requirements: It should be a high-quality product available on the market that meets adhesive strength requirements
Research by Kang et al [5] highlights that while the monolithic polymer KE-1300T used in glove production offers several benefits, the necessary design and mold modifications, along with the fabrication process, require considerable time and resources Additionally, the fabrication of string holes poses challenges, and the material's soft mechanical properties and inadequate surface contact lead to friction that hinders tendon movement The authors are committed to overcoming these limitations by leveraging Thermoplastic Polyurethane (TPU) and optimizing 3D printing technology for practical applications.
Thermoplastic Polyurethane (TPU) is an ideal material for gloves, thanks to its remarkable properties This flexible polymer boasts excellent tensile strength and high durability, allowing it to withstand pressures and impacts from artificial joints while offering superior abrasion resistance TPU is also waterproof, effectively absorbs impacts, and quickly returns to its original shape, conforming comfortably to the user's hand Additionally, TPU is health-safe, free from harmful substances, rigorously tested prior to production, and visually appealing.
Moreover, TPU, which is environmentally friendly, recyclable and reusable, helps to reduce plastic waste in the environment
The authors established clear objectives for glove design and approached the solution through a systematic process of design, fabrication, practical testing, and subsequent evaluation and refinement They introduce the initial glove concept used in the Mechatronics course project, detailing its strengths, weaknesses, and valuable lessons learned These insights provide a foundational framework that guides the direction of this thesis.
Figure 3.2: The glove design in the Mechatronics project:
(a) Front view of the glove and (b) Back view of the glove
The gloves offer significant advantages, including easy size customization and the innovative use of 3D printing technology with soft TPU plastic This material provides a compact and lightweight design that is easy to fabricate, ensuring fast production times while also being cost-effective Additionally, TPU plastic is skin-friendly and allows the gloves to flex and extend, thanks to its elastic properties and a tension-driven mechanism, which will be elaborated on in Chapter 4.
The arrangement of fibers on the back of the fingers, as shown in Figure 3.2, reveals several disadvantages, including suboptimal positioning and quantity Although the gloves can extend, they fail to provide the desired range of motion, resulting in folds that are hard to manage and causing material deformation that does not meet operational expectations Additionally, the bulkiness of the gloves, which remain attached to the structure, creates inconvenience for users and hinders ease of use.
3.2.4 Final design proposal - retaining advantages and addressing disadvantages
In this thesis, the authors draw on the operational principles of gloves from Seoul National University and incorporate insights from prior coursework to propose an optimal glove design that adheres to the specified design requirements.
Figure 3.3: Actual image of final design proposal:
The glove's design features both front and back views, showcasing its adherence to all specified criteria, including optimal dimensions This design effectively tackles the previously challenging issue of capital extension while ensuring a comfortable and enjoyable user experience, coupled with high aesthetic appeal The comprehensive design process will be detailed in the subsequent sections.
The kinematic model of the glove with human fingers
This article introduces a streamlined kinematic model for a soft glove featuring a single finger By optimizing the tendon lengths from the drive mechanism to the finger, this model significantly improves transmission durability Consequently, the authors can accurately predict the required tendon lengths for both bending and straightening the finger.
Optimizing the transmission system is crucial for adjusting tendon lengths to accommodate various hand sizes, allowing fingers to fully straighten and preventing the glove from encountering abnormal points that could cause contraction To minimize response movement, the authors have designed a fixed length from the drive to the body, utilizing Teflon tubes and pipe guides for enhanced safety Consequently, the primary focus for calculation and measurement is the distance from the user's arm to fingertip.
To make the design and calculations more efficient than, the authors have set three main objectives:
To ensure optimal finger function, it is crucial that tendon length during contraction exceeds that of the straightened position, preventing potential anomalies Additionally, minimizing tendon stretchiness during extension is important for reducing response time Proper design of the pulley radius in relation to tendon length is essential for effective finger movement transmission Consequently, conducting kinematic analysis is vital to achieve these objectives.
The authors introduce a kinematic model hypothesis where finger segments are rectangular, and joints can be adjusted vertically along each segment's path They utilize flexible guide wires with strong load-bearing capabilities, ensuring an even distribution of distance between the wires from the proximal to distal phalanx, including the thimble attachment position Additionally, the extended wires are arranged along the palm, forming an arc with a diameter equal to the finger's thickness As the finger transitions to different states, the wire path length adjusts accordingly, necessitating changes in the transmission tendon length to accommodate the desired posture Key parameters from the study are illustrated in the accompanying figure.
Figure 3.4: The Kinematic model of the soft glove
In the kinematic model, the x-axis corresponds to the central axes of the finger joints, whereas the y-axis is located at the endpoint of the palm, where it meets the wrist Each finger segment's dynamics are defined by the unit vector \( u_i \), and the lengths of these segments are represented by \( l_i \).
(i=1,2,3,4) correspond to the lengths of the joints between the palm and proximal phalanx, proximal phalanx, middle phalanx, and distal phalanx, respectively The authors define the vectors as follows [5]:
In the kinematic model, tendons are strategically placed on the fingers, with m1 representing the palm's endpoint, m2 to m4 indicating the strap locations, and m5 marking the fingertip's endpoint The distances Dk from m2 to mk, where k equals 3 or 4, are illustrated in the accompanying figure, with specific definitions for the distances between the points m2s.
Thus, it has the coordinate form of the point m 1 relative to the origin as follows:
The authors perform transformations as in (3.7), resulting in (3.8), then the authors determine m 2 relative to the origin in the coordinate system, resulting in (3.8):
By using the vector addition method, the authors obtain from (3.9) to (3.11):
TABLE 3.1PARAMETERS FOR KINEMATIC MODEL
Angles between phalanges and hands together
:1 4 u i i Z Phalanx unit vector s Distance from center line of phalange to strap end
:1 5 m i i Z m i = End point of the flexion wire on the hand
2 4 m m = Strap end points m 5= End point of the flexion wire on the thimble t 1/2 finger thickness
In the kinematic model, the flexion wire path is defined by the distance between points m1 and m5, with the required wire length for flexion determined by the changes in this distance The necessary tendon lengths for contraction are established by the authors, while tendon extension is guided by the cumulative angular changes at each joint.
DESIGN AND FABRICATION OF THE MECHANICAL TRANSMISSION
The operating principle of the gloves using tendon-driven transmission
The glove's design utilizes soft materials to simplify its structure, eliminating the need for complex joint alignment between the user and the rigid exoskeleton Tendons are anchored at the fingertips and along the phalanxes, allowing motors to pull them for finger bending and straightening (see Figure 4.1) Additionally, the glove's back is engineered to stretch, enhancing the flexibility of finger movements.
The innovative glove utilizes a single actuator to flex and extend the index and middle fingers, with dual wires controlling the flexion and a single wire managing the extension at the upper back of each finger This design allows the remaining fingers to move naturally in response to the actions of the index and middle fingers, reflecting the hand's biological characteristics Additionally, the thumb remains unconstrained and can accommodate accessories for specialized tasks, such as writing with a pen.
Technical requirements for transmission mechanism
In this project, the authors implement a tendon-driven transmission system utilizing monofilament wire, controlled by a DC motor and user-sent encoder signals This mechanism facilitates the flexion and extension of the index and middle fingers Additionally, the tendons from the mechanism connect to those extending from the glove, allowing for easy removal for cleaning purposes.
The authors seek to develop a mechanism that guarantees user safety while being compatible with soft gloves Based on insights from studies [2], [3], [4], and [5], they have outlined the technical requirements necessary for this mechanism.
- The mechanism should be compact in size to combine with the electrical part, creating a controller for the glove that can be carried along
- There should be no pre-tension forces on the system to ensure safety and not cause discomfort to the user
- The mechanism ensures that the winding and unwinding of the tendon wires does not creat slack and cause a jammed mechanism
- Achieve the necessary tension in the right amount of time to be able to control the fingers quickly and accurately.
Approaches to design of transmission mechanism
4.3.1 Approach 1 – The string-tension mechanism using spring elastic force
In a prior project, the authors developed and evaluated a string-tension mechanism that utilizes the elastic resistance of springs to generate tension in tendon wires This innovative mechanism successfully fulfilled the primary requirement of eliminating pre-tension while providing effective tension, enabling rapid finger contractions.
(b) Figure 4.2: String-tension mechanism using spring elastic force
(b) Motor position and finger flexion response (measured by flex sensor)
The finger contraction response graph closely resembles the motor's position graph, as illustrated in Figure 4.2b This similarity arises when the motor winds the string, resulting in a reduction in length while simultaneously lifting the guide wheel, which stretches the spring The spring resists this stretching, exerting force on the string, and reaches maximum tension at the point of greatest stretch.
(c) Figure 4.3: The design of a string-tension mechanism in the Mechatronics project:
(a) The outer part, (b) The inner part, and (c) The system layout
The string-tension mechanism generates temporary tension in the system as the motor winds the string, guiding it through the pulleys As the string shortens, it pushes the main pulley upward, which is mounted on a shaft linked to two springs that resist this vertical movement This elastic force counteracts the tension from the tendon string on the main pulley, establishing overall tension within the system.
The mechanism offers the advantage of eliminating pre-tension, allowing for effective tension that enhances finger responsiveness However, the string arrangement, illustrated in Figure 4.3b, presents challenges with three wire kinks that occur when the motor rotates against the coil, resulting in the actual string return being slower than the motor's return speed This creates a coil slack issue when the mechanism ejects the strings, leading to an undesirable response Additionally, as shown in Figure 4.3c, the transmission setup for both the middle and index fingers necessitates two separate mechanisms, resulting in a larger, less compact transmission design.
4.3.2 Approach 2 – The tendon-controlling transmission mechanism
In this thesis, the authors analyze the operating principles of a mechanism developed at Seoul National University and draw insights from a previous project to create a mechanical design that fulfills the necessary technical specifications.
Figure 4.4: Drive diagram of the transmission mechanism:
(a) The drive diagram and (b) The tendon wiring method
The proposed transmission mechanism features key components illustrated in Figure 4.4a, with a main coil designed to wind and return wire to the fingers It incorporates four smaller coils of equal diameter, specifically two for flexing the middle and index fingers and two for extending them This design leverages the motor's rotational capabilities for both finger contraction and extension Additionally, the feeding coil is divided into four separate coils to prevent string entanglement, simplifying the design and calculations The uniform diameter of the small coils enhances control over finger movements, while the wire length adjustment is conveniently managed from the glove side.
Figure 4.4b demonstrates the operation of the tendons and coils in finger movement When the finger flexes (indicated by the red wire), the motor winds the strings to pull the finger, causing the transmission gear to rotate the feeding coil in the opposite direction This action simultaneously pushes the flexion strings while retracting the extension strings (represented by the blue wire), preventing resistance during movement The mechanism for releasing the tendon strings from the coils includes not only the feeding coil but also passive guide wheels and a 2000-grit rough surface on the coil, which aids in securely clamping and releasing the wires.
The initial evaluations of this structure indicate that it meets size requirements, eliminates the need for pre-tensioning, and minimizes unwinding issues on the main coil Additionally, the actuator's tension response will be assessed experimentally through the analysis of signals from flex sensors.
Calculate and design the transmission mechanism
The gear ratio in the transmission mechanism is a crucial input condition, as it determines how effectively the feeding coil pushes the coils outward during rotation To prevent the unwinding of the coil when retracting the tendon strings, specific conditions must be met.
Where ω main , ω feeder are the rotation speeds of the main and feeding coils, respectively, in revolutions per minute To ensure the above conditions, the authors chose a ratio of
2 gears of 3:2 due to detailed supply conditions with gear parameters in Table 4.1:
TABLE 4.1: THE DIMENSIONS OF THE GEARS
In transmission mechanisms, wires must meet high standards for load-bearing capacity, low elasticity, and durability to ensure precise control and minimal replacement The ideal wire should also be flexible enough to bend easily, allowing it to be rolled closely to the surface of the main coil Therefore, the authors selected a 0.7 mm diameter monofilament line, which offers a maximum pulling capacity of 20 kg while maintaining excellent flexibility.
4.4.3 Calculate the dimension of the main coil and the feeding coil
To prevent the mechanism from unwinding the coil, the exit speed of the wire from the feeding coil must exceed the release speed of the wire from the main coil.
Figure 4.5: An illustration of when the system is pulling the fingers
From Figure 4.5, considering the state when flexing the finger (the direction when stretching is similar), the condition for the coil unwinding to not occur is: v m in a v feeder (4.2)
Where v main and v feeder respectively, is the linear velocity of the wire points at the time being considered Expanding velocities expression in (4.2), the authors have: a 1 2 a 1 1 a
(1.5) 1.5 m in feeder m in feeder m in feeder
To ensure proper engagement between transmission gears, the diameter of the steps separating the coils must be smaller than the diameter of the gear ring Specifically, the step diameter, referred to as D feederstep, must meet the condition of being less than 20.
Choose D feeder to be 15mm, D feederstep to be 17mm From then, the authors define the condition for the main coil’s diameter (4.3):
Choose R main to be exact 22mm for ease in 3D printing the details
4.4.4 Calculate and choose the motor
Technical resistance forces in the mechanism include the resistance encountered when folding the finger and the friction on the feeding coil's surface The feeding coil is coated with a layer of sandpaper with a 2000 grain size, providing a smooth yet durable surface, making the friction negligible Therefore, the primary resistance comes from the finger itself According to documentation [13], the finger-pulling force is considered equivalent to the average grip force in humans, which is approximately 15N.
M finger F finger d main 15 22 330[N.mm] ≈ 3.37 [kg.cm] (4.6)
The authors define the conditions to both be able to bend the finger and ensure power assistance so that the motor's torque meets the following conditions:
The authors have selected the JGB37-520 motor, which meets the specified condition (4.3) and other essential motor requirements, including availability and low-speed operation for safety during driving This motor is equipped with an encoder and operates at a voltage of 10-12V, as detailed in Table 4.2.
TABLE 4.2: TABLE OF JGB37-520 DC MOTOR CHARACTERISTICS
No-load maximum speed (RPM) 60-69
Assuming an ideal transmission from the motor to the flexible coupling, the shaft torque of the main coil's shaft can be determined from equation (4.6) when the system is balanced.
The shaft diameter is preliminarily determined according to formula 10.9, based on document [14]: d main 3 0.2 T main (4.9)
In which, for the steel shafts, then [τ] = 15 30 MPa, choose [τ] = 15 MPa, replace the
(4.9) with T main = 0.330 N.m, the authors have 3 0.330 6 4.79
To ensure compatibility with the gear's inner diameter, select a main shaft diameter of 6mm, which will also be used for the feeder coil For the idling wheels, opt for 5mm shafts.
When selecting bearings for coil support, it is essential to choose the right shaft diameter, as high precision is not critical For the main coil support and transit, opt for 608ZZ bearings, which feature an 8mm inner diameter, a 22mm outer diameter, and a thickness of 7mm For idling wheels, select 625Z bearings, characterized by a 5mm inner diameter, a 16mm outer diameter, and a thickness of 5mm.
THE DEVELOPMENT OF CONTROL PROGRAM AND DEVICE OPERATION
Device control purposes
The authors introduce a semi-automated control system for an assistive glove device, which necessitates human intervention for safety This system is designed to automatically respond to user requests while ensuring that control inputs are managed with user permission.
The control system utilizes a rotary encoder, enabling users to precisely adjust the motor shaft position in small angular increments for enhanced tension control To establish the maximum revolutions of the motor as a control limit, the authors implement a calibrated measurement of finger contraction and extension at the device's starting point, ensuring safe and restricted movement of the glove.
Actuator controlling system
The actuator system's control component performs essential functions, including processing user signals from push buttons and encoders, interpreting sensor data from flex sensors and magnetic motor encoders, displaying device status through LED indicators, controlling the DC motor actuator, and facilitating communication with data visualization software.
Figure 5.1: Block diagram overview of the control system
5.2.2 Components in the control system
As shown in Figure 5.1, this system includes main blocks: User input block, Sensor block, LED display block, Motor control block, and Main microcontroller block
Figure 5.2: Schematic diagram of the user input block
The user input block features two buttons for calibration and reset calibration, necessitating the addition of capacitors to mitigate button bouncing and alleviate software load The encoder is equipped with built-in resistors and capacitors that filter the output pins, which connect to the microcontroller (see Figure 5.2).
Figure 5.3: Indicator Block and Flex Sensor Block
The indicator block features three primary LEDs: red for system initialization, blue for calibration, and green for operational status, as outlined in the flow chart Additionally, the sensor block is equipped with two sensors positioned on the index and middle fingers to detect finger contractions and extensions through sensor deformation.
Figure 5.4: Motor and Motor Driver Blocks
The authors implemented an H-bridge circuit utilizing the TB6612FNG IC, which supports a maximum voltage of 1.2A, proving to be more efficient than the widely used L298N This advantage is attributed to the TB6612FNG's MOSFET switching capabilities, enabling effective position control Previous experiments indicated that while the L298N struggles to respond with a PWM below 10%, the TB6612FNG maintains reliable performance even at PWM levels as low as 5-10%.
Figure 5.5: The H-bridge circuit using IC TB6612FNG
The main controller block features an STM32F103C8T6 microcontroller, which efficiently manages the PID controller, filters noise from bend sensors, and processes user inputs Additionally, it utilizes the Serial Wire Debug protocol to seamlessly connect with data visualization software, eliminating the need for manual feature construction with other serial protocols for computer connectivity.
Figure 5.6: The main controller block
Finally, the power block includes a 12V - 2000mAh battery and LM2596 DC-DC voltage reduction Buck modules to lower the voltage to 3.3V for the buttons and 5V to power the Microcontroller
Figure 5.7: The power supply block
5.2.3 Interface with the STMCubeMonitor application
STM32CubeMonitor is a powerful software tool designed for fine-tuning and diagnosing applications for the STM32 chip by providing real-time data visualization Its non-intrusive monitoring feature allows users to access and display data without modifying the main program, utilizing data transfer protocols such as SWD, UART, I2C, or SPI for seamless communication with a computer By employing the ST-Link device with Serial Wire Debug (SWD) protocol, developers can read, modify, and visualize variable values efficiently.
To effectively monitor various variables across different graphs, it is essential to establish distinct monitoring groups In this project, the authors focus on four primary data groups: the Flex Sensor Data Group, the Motor Position Data Group, the System Status Data Group, and the System Control Data Group, which encompasses user inputs.
The STMCubeMonitor's block organization, as illustrated in Figure 5.8, consists of an overview setup and monitoring variables within the program Initially, the Monitor groups block initializes four data groups, from which the authors identify the necessary variables to track in the Variables edit nodes blocks, specifically focusing on the flex sensor and DC motor, as depicted in Figure 5.9b Following this, signal processing blocks (Processing edit nodes) are established to facilitate data export to log files, create virtual variables, and utilize program variables for post-processing calculations and visualization The final step involves configuring graph blocks for display size, shown in Figure 5.9a, where these blocks are monitored collectively through a debugger device, connecting to Start/Stop Acquisition, DeviceProbeIn/Out, Clear, and Notification blocks Upon completing the settings, executing the monitoring software yields results similar to those in Figure 5.9.
Figure 5.9: The project’s data monitoring window.
Responses of the system
5.3.1 Read and filter signal noise from bending sensors (Flex sensors)
The authors conduct an experimental evaluation of two filtering techniques, Butterworth and Kalman, to assess their effectiveness in noise reduction and change tracking during contraction and extension The study aims to ensure minimal value fluctuations in these states, highlighting the suitability of each filter for the intended application.
5.3.1.1 Noises filtering with the Butterworth low-pass filter
To address noise in the raw signal from the flex sensor, the authors initially considered implementing a low-pass filter to eliminate high-frequency disturbances, particularly when the hand is relaxed During peak finger contraction, the sensor values are expected to stabilize They utilized a 5th-order Butterworth low-pass filter, employing a discrete update formula for optimal performance.
Where y is the filtered output at state (n), y k (k=1, 2…4) is the filtered output at (n
– k) th state, and x m (m=0, 2…5) is the raw reading data at (n-m) th state
The coefficients a_i (i=0, 1 … 4) and b_j (j=0, 1 … 5) are determined through coefficient matching, influenced by the selected cutoff frequency for the filter and the sampling period To maintain brevity, the authors will provide the formula for calculating these coefficients rather than specific numerical values, as extensive experimentation with various cutoff frequencies has been conducted.
Where (5.3) contains the middle-process calculation coefficients:
The following formulas are used to calculate a i and b j (i=0, 1 … 4; j=0, 1 … 5):
The authors evaluate the performance of the Butterworth filter at various cutoff frequencies, specifically 10Hz, 5Hz, 2.5Hz, and 0.95Hz, to determine its suitability for the system The accompanying graph illustrates a decrease in value, indicating that the finger is flexed.
(b) Figure 5.11: Filter at 10Hz cutoff frequency, where:
(a) Raw (left) and Filtered data (right) and (b) Fluctuating range of value of relaxed hand
At this frequency, the oscillation range when the user relaxes their hand is high (25 units) will greatly affect the safe control to stop at the right time
(b) Figure 5.12: Filter at 5Hz cutoff frequency, where:
(a) Raw (left) and Filtered data (right) and (b) Fluctuating range of value of relaxed hand
At this frequency, the oscillation range when the user relaxes their hand is still high
(b) Figure 5.13: Filter at 2.5Hz cutoff frequency, where:
(a) Raw (left) and Filtered data (right) and (b) Fluctuating range of value of relaxed hand
At this frequency, the oscillation range is significantly reduced to just 10 units, maintaining favorable phase characteristics Nevertheless, the authors are continuing their research to further minimize this oscillation range.
(b) Figure 5.14: Filter at 0.95Hz cutoff frequency, where:
(a) Raw (left) and Filtered data (right) and (b) Fluctuating range of value of relaxed hand
At this frequency, the oscillation range remains largely stable, but the phase characteristics start to change If the cutoff frequency is further decreased, the filter's performance will decline, leading to a gradual deterioration of the phase characteristics.
The analysis reveals that oscillation frequencies of 2.5Hz and 0.95Hz can be considered, yet the filtering outcomes depicted in Figures 5.13a and 5.14a indicate that these results still exhibit fluctuations consistent with the raw data To address the issue of unknown noise and enhance the accuracy of readings from a relaxed hand, the authors propose utilizing a Kalman filter.
5.3.1.2 Noises filtering with the Kalman filter
In this section, the authors present experiments utilizing the Kalman filter for sensors, building on the theoretical foundation established in Chapter 2 They focus on two critical values: r_n, representing measurement uncertainty, and q_n, indicating process noise An increase in r_n leads the filter to rely more on the previous state’s extrapolated value, while a decrease allows it to align more closely with the measurement results For their experiments, the authors adjust r_n and set q_n to 0.035.
(If q n is close to 0, the filter’s performance will be poor)
The authors present representative experiments with uncertainties including 20 2 ,
10 2 , 8 2 and 7.5 2 to determine the suitable one for the system
The initial test results shown in Figure 5.15 indicate that while fine filtration has been applied, the phase characteristics remain suboptimal This is primarily due to the filter's inability to accurately track the measured values, which can be attributed to a persistently high r n value.
(b) Figure 5.16: Kalman filter result with r n = 10 2 , where:
(a) Filtered output and (b) Oscillation range of the value
The results depicted in Figure 5.16a indicate a significant enhancement in phase characteristics, with a minimal fluctuation of only 2 units This represents a substantial improvement in filtering outcomes compared to earlier experiments The authors proceeded with further experimentation to continue enhancing phase characteristics.
(b) Figure 5.17: Kalman filter result with r n = 8 2 , where:
(a) Filtered output and (b) Oscillation range of the value
As the value decreases, the phase characteristic results improve, and the oscillation range stabilizes However, lower values attract more noise, as the filter increasingly aligns with the measured value.
(b) Figure 5.18: Kalman filter result with r n = 7.5 2 , where:
(a) Filtered output and (b) Oscillation range of the value
This experiment concludes that by changing the r n value, it can be seen in Figure
5.18a that the phase characteristics become better but at the cost of a larger oscillation range (4 units) and the value fluctuates to noise Therefore, if it continues to decrease, the set goal will not be achieved Therefore, the above experiments have helped the authors to choose the appropriate value of r n = 8 2
5.3.2 Finger’s range of movement calibration feature
The authors present a calibration feature for finger flexion range that activates when the user first opens the device This functionality is integrated into the device control program, enhancing user experience and precision.
This feature is essential for defining the user's movement limits, preventing the winding motor from causing excessive finger flexion or extension Additionally, as the conductive ink on the sensor ages, its bending sensor values will change over time, making the initially calibrated default value unreliable for prolonged use.
5.3.3 DC motor position control responses
The objective of the control system is to achieve minimal fluctuation in position response, maintaining it within a narrow range of 1-5% of one full revolution This section details the experiments conducted to optimize the proportional-derivative (PD) parameters of the PID controller for precise motor position control, utilizing the coefficient determination method outlined in reference [16].
To define K D , the authors conducted experiments of K D = 0.035 and decreased 10% time after time The authors’ experimental part will focus on presenting representative values for the experiments
(b) Figure 5.20: Responses of motor shaft position control (by number of pulses): (a) With K D = 0.035 the steps respectively are 1% and 5% of 1 round
(b) With K D = 0.025 the steps respectively are 1% và 2.5% of 1 round