Safe Motion Planning for Multi-Vehicle Autonomous Driving in Uncertain Environment

In the field of motion planning for autonomous driving systems, ensuring the safety of multi-vehicle navigation is one of the crucial topics. An unavoidable problem in practice is that the noise-induced uncertainties in real-world applications highly degrade the safety of multi-vehicle navigation. It is also challenging to guarantee the required computation efficiency of motion planning algorithms in such uncertain environments. In this work, we present a novel motion planning framework to enhance the safety and computation efficiency of multi-vehicle navigation. This framework utilizes the iterative linear quadratic Gaussian (iLQG) algorithm to deal with the nonlinearity of the vehicle dynamics and overcomes the difficulties in handling inequality constraints (e.g., collision avoidance constraints). Furthermore, we propose an innovative Alternating direction method of multipliers based Linearized Chance Constraint (ALCC) method to address collision constraints in noisy uncertain environments. Simulation experimental results demonstrate that our method achieves higher safety with high computational efficiency compared to other methods in various multi-vehicle motion planning and navigation scenarios.