Learning from Experience for Rapid Generation of Local Car Maneuvers
This work provides a faster and more reliable local path planning solution for autonomous vehicles, which is crucial for real-time responsiveness in dynamic traffic situations.
This paper addresses the challenge of rapidly generating feasible local paths for autonomous cars. They trained a deep neural network (DNN) to plan paths in approximately 40 ms, outperforming existing planners in task completion rates while generating smooth, nearly-optimal paths.
Being able to rapidly respond to the changing scenes and traffic situations by generating feasible local paths is of pivotal importance for car autonomy. We propose to train a deep neural network (DNN) to plan feasible and nearly-optimal paths for kinematically constrained vehicles in small constant time. Our DNN model is trained using a novel weakly supervised approach and a gradient-based policy search. On real and simulated scenes and a large set of local planning problems, we demonstrate that our approach outperforms the existing planners with respect to the number of successfully completed tasks. While the path generation time is about 40 ms, the generated paths are smooth and comparable to those obtained from conventional path planners.