It's a prototype of a driverless car that processes imagery and ultrasonic sensor data to avoid it from colliding into leading vehicle/object and getting off the lane line. This system uses RaspberryPi 4B as the mainbboard to process all the input data and outputs the steering angle and speed (PWM) values that the car should perform. Then, those values are sent to an arduino uno using serial communication to be executed on the car using ON/OFF control for the speed and PD (Proportional-Differential) control to control the steering angle.
The system begins by taking image and reading ultrasonic sensor data with multithreading to ensure the reading of the same situation by both sensors. The image captured by the camera then undergoes some image processing techniques to read the distance of the car to the leading car that's attached with a plain color object as a distance reading reference. Then the distance information of both sensors are fused into one final distance reading using a Fuzzy Logic Algorithg which purpose is to increase the distance reading robustness and accuracy due to the strengths and weaknesses of camera sensor and ultrasonic sensors. This data then used to avoid the car from colliding into the leading vehicle. After that, the system process the image data used before to detect the position of the car in reference to the lane lines. This data is used to avoid the car from getting off the lane.
The distance reading data, car position data, and the speed input of the car are then used as an input of another Fuzzy Logic Algorithm that outputs the speed and steering angle values that the car should execute. In the end of the experiment, all input, processed, and output data then stored in SD card for further analysis.