This study employed amixed-method data collection approach, incorporatingprimary dataobtained through experimental measurements using manual instruments andsecondary datagathered from academic sources, technical reports, and previous studies.
Primary Data Collection
Primary data were collected throughcontrolled experimental procedures, ensuring accuracy and consis- tency in measurements. The following manual measurement instruments were used:
• Stopwatch: Used to record reaction times and the duration of specific tasks in the experiment. Each measurement was repeated multiple times to minimize error, and the average value was taken to enhance accuracy.
• Protractor: Used to measure precise angles during experimental trials. Measurements were conducted by multiple observers and repeated to reduce potential observational bias.
• Ruler (Length Measurement Tool): Used to measure physical distances and dimensional changes of the experimental subjects under different conditions.
The data collection process followed these key steps:
1. Defining Experimental Conditions: Environmental factors, test subjects, and measurement parameters were standardized to ensure repeatability.
2. Conducting Measurements: Each measurement was performed multiple times (n = 5) to derive an average value and exclude anomalies.
3. Recording Data: Results were logged systematically to prevent discrepancies in data collection.
The use of manual instruments reduced dependency on automated tools, limiting systematic errors associ- ated with software faults or sensor deviations. However, repeated measurements were necessary to enhance reliability.
Secondary Data Collection
Secondary data were obtained from credible academic sources, including:
• Peer-reviewed journal articlesfrom databases such asScopus, IEEE Xplore, and ScienceDirect, pro- viding insights for comparative analysis.
• Technical reports and industry documentsfrom research organizations and corporations, offering prac- tical information on measurement standards and industrial applications.
• Public datasets, including statistical records and previous experimental results, used to validate findings.
The integration of primary and secondary data ensured that the study’s findings were substantiated by both experimental results and prior academic research.
Data Analysis Methods
Data were analyzed using three primary methods:
Manual Statistical Analysis
Since this study employed manual measurements instead of parametric statistical tests such ast-testsorANOVA, data analysis relied ondescriptive statistics, including:
• Mean (Arithmetic Average): Calculated by summing all recorded measurements and dividing by the number of trials, revealing general trends in the data.
• Standard Deviation: Used to measure data dispersion relative to the mean, assessing the consistency of experimental results.
• Range: Determined by identifying the difference between the highest and lowest recorded values, high- lighting potential anomalies.
For example, in measuring reaction times using a stopwatch, the dataset might include:
T ={0.170,0.175,0.168,0.160,0.167}
The mean reaction time is calculated as:
à=
∑T
n = 0.170 + 0.175 + 0.168 + 0.160 + 0.167
5 = 0.168miliseconds Additionally, standard deviation is computed to evaluate the variability of the measurements.
Comparative Analysis with Prior Research
After deriving descriptive statistics, the results were compared with previous studies to assess consistency and deviations:
• If findings align with prior research, this validates the methodology and enhances the credibility of results.
• If significant discrepancies are observed, potential reasons include differences in experimental condi- tions, sample variations, or measurement accuracy.
For instance, a prior study byNguyen et al. (2020)reported an average reaction time of2.5 seconds, while this study measured2.28 seconds, necessitating further examination of experimental conditions and method- ological differences [53].
Reliability and Validity Assessment
To ensurereliabilityandvalidity, the study implemented:
• Reliability Testing: Repeated measurements under identical conditions to confirm data consistency.
• Validity Checks: Comparing results against established measurement standards and previous research to verify accuracy. High correlation with existing studies suggests strong validity.
Conclusion
This section has detailed the methodology used for data collection and analysis. The application ofmanual measurement techniquesensures direct and controlled data acquisition, although it requires rigorous procedu- ral accuracy.Descriptive statistical methods, comparative analysis, and reliability assessments were employed to validate findings, ensuring their credibility and generalizability.
System Performance Overview
Objective: Evaluate the vehicle’s operational capabilities under real-world conditions, from sensor efficiency to signal processing and obstacle avoidance.
• Response Time:
– The average response time from signal reception to action execution is300 ms.
– Latency varies based on the control mode:
* Web Remote Control:100 - 300 ms (with stable Wi-Fi connection).
* Hand Gesture Control:150 - 300 ms (via NRF24L01 transmission).
* Autonomous Mode:300 - 500 ms (due to HC-SR04 signal processing and motor control).
• Accuracy:
– MPU-6050:Gesture detection accuracy reaches±2°after Kalman Filter application.
– HC-SR04:Average deviation is±1.5 cmfor obstacles within 1 meter.
– NRF24L01: Data transmission success rate is98%in non-interference environments and90%in interference-prone conditions (within a 30-meter range).
• Durability and Battery Life:
– The system operates continuously for90 - 120 minuteswith a 3000mAh battery.
– Autonomous Modedepletes power fastest due to continuous servo and sensor operation.
• Obstacle Avoidance Efficiency:
– Obstacle avoidance success rate is85%in simple environments and70%in complex environments.
– Servo rotation from 0° to 180° occurs within0.5 seconds, ensuring timely environmental scanning.
Web Remote Control Mode
Objective:Assess the performance of remote vehicle control via a web interface using the ESP32-CAM.
• Signal Latency:
– 200 - 300 msunder stable network conditions (Wi-Fi range of 10 - 15 meters).
– Latency can rise to500 msin weak or interference-prone networks.
• Operational Range:
– In open spaces, the vehicle can be controlled up to20 meters.
– In obstructed environments, the range decreases to10 - 15 meters.
• Connection Stability:
– The vehicle reconnects automatically if the signal is lost within5 seconds. If disconnection persists, the vehicle halts to prevent accidents.
Hand Gesture Control Mode
Objective:Evaluate MPU-6050 accuracy and NRF24L01 transmission efficiency.
• MPU-6050 Accuracy:
– Angle deviation remains under±2°post Kalman Filter application.
– Acceleration signals are processed through acomplementary filterto mitigate vibrations.
• NRF24L01 Signal Latency:
– 100 - 150 msper successful signal transmission.
– Error rates under interference conditions are5-8%, causing minimal performance degradation.
• Control Capability:
– The vehicle responds instantly to hand tilts. Tilt angles exceeding15°forward or backward trigger forward/reverse movement.
– Tilting the hand> 10°to the left or right triggers turning actions.
Autonomous Mode
Objective:Assess the vehicle’s independent operation and obstacle avoidance capabilities using HC-SR04 and SG90 servo motor.
• HC-SR04 Accuracy:
– Deviation is±1.5 cmfor distances under 1 meter and±3 cmfor longer ranges.
– Effective detection range spans from2 cm to 3 meters.
• System Response Time:
– The vehicle stops within300 msupon detecting obstacles within20 cm.
• Environmental Scanning:
– The servo completes a0° to 180°sweep in0.5 seconds.
• Obstacle Avoidance Success Rate:
– 85%in low-obstacle environments.
– 70%in cluttered or narrow spaces.
• Error Handling:
– If no clear path is detected, the vehicle reverses and rescans the environment.
System Performance Analysis and Result Web Remote Control:
Data Measurement Method: Use a stopwatch to measure the vehicle’s response time from pressing the control button until the vehicle starts running with a small error.
Accuracy: All tests achieved 100% accuracy, indicating the web control system reliably executes control commands.
Response Time:
Figure 4.1: Testing and Evaluation Web Control
• The average response time is 168 ms, ranging from 160 ms to 175 ms.
• Complex command combinations (e.g., “Move Forward + Turn Left + Move Backward + Turn Right”) do not significantly increase response time.
Conclusion: Firgure 4.1 described the web control system demonstrates high reliability and fast response time, making it suitable for real-time operations.
Hand Gesture Control:
Figure 4.2: Testing and Evaluation Hand Gesture
Data Measurement Method:Use a protractor with a distance of 0.1 degrees and redraw the details on a piece of paper or a plane parallel to the ground. Gradually move your hand until you reach the angle at which the
vehicle operates, then stop and take notes to save the data obtained. Combine with the use of a stopwatch to measure the delay of the vehicle’s action from the time the hand movement stops.
Accuracy:
• The average accuracy is 97.6%, higher than autonomous mode and comparable to web control.
• The “Stop” command achieved 100% accuracy, demonstrating reliable recognition for critical actions.
Hand Angle Deviation: The deviation between actual hand angles and threshold angles is minimal (ap- proximately±1◦), ensuring consistent recognition.
Conclusion: Figure 4.2 described Hand gesture control exhibits high accuracy and stability, making it a viable alternative control method.
Autonomous Mode:
Figure 4.3: Testing and Evaluation Autonomous Mode
Data Measurement Method:Use a ruler to measure the length of 30cm and draw the details on a piece of paper or a plane parallel to the ground. Drive the car on a straight line, when the car stops, measure the distance between the car and the obstacle, then record the measured data.
Accuracy:
• The average accuracy is 85%, lower than web control, reflecting potential challenges in obstacle detection and avoidance.
• Errors might stem from environmental factors such as lighting, surface conditions, or sensor limitations.
Response Time:
• The average response time is 361 ms, significantly higher than web control.
• The range of response times (310 ms to 417 ms) suggests the autonomous mode requires additional processing time compared to manual control.
Obstacle Distance Maintenance: The system maintains an average distance of 19.8 cm from obstacles, which is relatively consistent but requires further testing under different conditions.
Conclusion:Figure 4.3 described the autonomous mode achieves moderate accuracy and stability but needs improvement to ensure better performance.
Overall Comparison
Factor Web Control Autonomous Mode Hand Gesture
Accuracy (%) 100 85 97.6
Response Time (ms) 168 361 N/A
Stability High Medium High
The experimental results found that each control mode of the multi-mode RC system has a priority mode and a restricted mode. The Web Control mode can demonstrate outstanding response time and reliability, which is especially suitable for real-time applications. With a carrier rate of only 160–175 ms under stable network conditions, this mode meets the needs of remote control via a web interface.
Meanwhile, the default Autonomous Mode has been proven to be capable of detecting and avoiding obsta- cles, but there is still room for improvement in accuracy and processing speed, especially in complex environ- ments. The current HC-SR04 sensor has limitations in range and sensitivity, resulting in an obstacle avoidance rate of only about 81–88%.
Finally, the Hand Gesture Control mode shows great potential thanks to its high accuracy (97.6%) and intuitive operation. However, its performance needs to be further evaluated in more challenging environments, where it appears to be limited by interference or caution from other wireless devices operating in the 2.4 GHz band. Overall, each mode brings its own value to the system, while also opening up new avenues for optimizing overall performance.