Self-driving rover that follows a car (with specific license plate number) and maintains a constant distance between the later using fuzzy control
The designed mechatronic system consists of a car-following rover that aims to follow a nearby moving obstacle and maintain a certain distance from it. To elaborate, a fuzzy controller determines the rover’s motors’ speed depending on the relative distance and distance increase with the obstacle to detect its speeding or deceleration, while accounting for any inclination the traversed path is subject to. To simulate how a cop car would operate, we also added the feature of detecting and registering the car’s license plate using a convolutional neural network.
The implemented features include:
- Detecting the presence of the nearby car using the distance sensors;
- Prompting the rover to follow it while maintaining a certain set reference distance from it;
- Adjusting the rover’s speed using fuzzy logic while accounting for any nonlinear path the obstacle may take by varying the speed of its left and right pairs accordingly;
- Capturing the license plate of the speeding obstacle using the CNN