2D Vehicle Motion Simulator with Ackerman Steering

Recently, autopilots are increasingly being used for controlling cars and robots. Proper quality control of self-driving vehicles is critically important for ensuring road safety. Additionally, some autopilots employ neural network technologies. A motion simulator is required to test the correctness of the path planning module and to train the control model. The Robotics Group of the Faculty of Computer Science at the National Research University Higher School of Economics (HSE) is developing a real robot in the form of a small four-wheeled car that is capable of moving in a static indoor space in autopilot mode. The robot features Ackerman steering, rear-wheel electric drive, and a differential. A 2D LiDAR, a stereo camera and an inertial measurement unit are used for navigation. This paper presents a program that simulates the movement of this car. This paper contains 33 pages, 3 chapters, 11 figures, 1 table, 9 listings and 28 sources. Keywords: robotics, autopilot, self-driving, simulator, Ackerman steering geometry, collision detection.