Design of A Two-point Steering Path Planner Using Geometric Control.

For lateral vehicle dynamics, planning trajectories for lane-keeping and lane-change can be generalized as a path planning task to stabilize a vehicle onto a target lane, which is a fundamental element in nowadays autonomous driving systems. On the other hand, two-point steering for lane-change and lane-keeping has been investigated by researchers from psychology as a sensorimotor mechanism of human drivers. In the first part of this paper, using knowledge of geometric control, we will first design a path planner which satisfies five design objectives: generalization for different vehicle models, convergence to the target lane, optimality, safety in lane-change maneuver and low computational complexity. Later, based on this path planner, a two-point steering path planner will be proposed and it will be proved rigorously that this two-point steering path planner possesses the advantage--steering radius of the planned trajectory is smaller than the intrinsic radius of reference line of the target lane. This advantage is also described as "corner-cutting" in driving. The smaller driving radius of the trajectory will result in higher vehicle speed along the winding roads and more comfortness for the passengers.

[1]  E Weinan,et al.  Machine Learning Approximation Algorithms for High-Dimensional Fully Nonlinear Partial Differential Equations and Second-order Backward Stochastic Differential Equations , 2017, J. Nonlinear Sci..

[2]  Francesco Borrelli,et al.  Kinematic and dynamic vehicle models for autonomous driving control design , 2015, 2015 IEEE Intelligent Vehicles Symposium (IV).

[3]  Eric W. Justh,et al.  Boundary following using gyroscopic control , 2004, 2004 43rd IEEE Conference on Decision and Control (CDC) (IEEE Cat. No.04CH37601).

[4]  Rob Gray,et al.  A Two-Point Visual Control Model of Steering , 2004, Perception.

[5]  E Weinan,et al.  Deep Learning Approximation for Stochastic Control Problems , 2016, ArXiv.

[6]  C. Tracy,et al.  A Fredholm Determinant Representation in ASEP , 2008, 0804.1379.

[7]  The universality classes in growth of iron nitride thin films deposited by magnetron sputtering , 2003 .

[8]  C. Tracy,et al.  Integral Formulas for the Asymmetric Simple Exclusion Process , 2007, 0704.2633.

[9]  Roux,et al.  Roughness of crack interfaces. , 1991, Physical review letters.

[10]  Shane Legg,et al.  Human-level control through deep reinforcement learning , 2015, Nature.

[11]  Stefan Schaal,et al.  A Generalized Path Integral Control Approach to Reinforcement Learning , 2010, J. Mach. Learn. Res..

[12]  Arnulf Jentzen,et al.  Solving high-dimensional partial differential equations using deep learning , 2017, Proceedings of the National Academy of Sciences.

[13]  R. Jazar.  Vehicle Planar Dynamics , 2019, Advanced Vehicle Dynamics.

[14]  Zhang,et al.  Dynamic scaling of growing interfaces. , 1986, Physical review letters.