Research on Autonomous Driving Simulator Control and Decision Algorithms based on Computer Vision Methods

摘要

Researches on perception and control technologies are the most crucial area in autonomous driving systems and is the core to achieve fully autonomous driving. Simulations, especially hardware-in-loop and algorithms-in-loop could accelerate the testing process. In this article, based on the second World Intelligent Driving Challenge took place in Tianjin, held by CATARC in May 2018, with the combination of computer vision, edge detection, objection detection, Kalman filter and vehicle dynamics methods, constructed in the Automotive Artificial Intelligence Simulator environment, aiming at paving the road to the better research on simulation testing and autonomous driving.