Efficient Speed Planning in the Path-Time Space for Urban Autonomous Driving

In this article, an algorithmic speed planning method for an autonomous vehicle dealing with moving obstacles is presented. The path-time space is used to represent predicted collision zones with the moving obstacles, and algorithms are proposed to build an appropriate speed profile to pass before and after these obstacles while ensuring the respect of safety distances, by leveraging the path-time representation. These algorithms are implemented in MATLAB/Simulink and simulations with SCANeR Studio are conducted to provide realistic results of the proposed method. The results show that the computed speed profiles are suited to the situation and obtained with acceptable processing time.