In this repository, two sets of codes for intelligent control of cars in urban and suburban areas are introduced. The simulation environment of this project is based on the Unity simulator for self-driving cars and developed by the technical team of the FIRA International Competition. GithubRepo
The developed codes were eventually participated in the FIRA (Federation of International Robot-Sport Association) competition and won four places in the various sections of the competition.
In code developed for suburban or speed racing, concepts such as hough transform and canny edge detection have been used for line detection, digital PID, MPC, and instrumentation techniques such as the implementation of a digital encoder have been used for vehicle control. Also, ultrasonic sensors and image processing techniques have been used to detect obstacles and avoid collision with them. The use of behavioral cloning in the research and development process has also been investigated.
Example of code performance in the suburban environment:
Examples of Processes performed in the suburban environment:
In the code developed for urban competition, in addition to using all the concepts described in the race section, neural networks have also been used to detect traffic signs. YOLOV5 Object Detection Model
Also, due to the more performance sensitivity of autonomous cars in this part because of issues such as the possibility of pedestrians, speed was lower in the urban part and our focus was on higher accuracy and not hitting the obstacles by simulating the forced stop system.
Example of code performance in the urban environment:
Examples of Processes performed in the urban environment:
In this competition, our team was able to achieve four ranks in different categories as follows:
- First place in the technical report
- Second place in race section (suburban)
- Third place in the urban sector
- Third place in the total competition
Feel free to contact us via email or connect with us on linkedin.