Path planning method and device

Abstract

The disclosure provides a path planning method and device. The path planning method comprises the following steps: after determining a global path according to a starting point and an ending point, determining a plurality of local areas and a local starting point, a local ending point and a local scene of each local area; establishing an unconstrained optimization proposition corresponding to thelocal area according to the local starting point, the local terminal point and the local scene; generating a plurality of local paths as initial solutions of the optimization proposition comprising atrajectory curvature and a plurality of penalty functions; under each initial solution, setting the weights of a plurality of penalty functions to be the same and calculating a solution of the optimization proposition; when the solution does not meet a preset termination condition, adjusting the weights of a plurality of penalty functions and then continuously performing calculating by taking thesolution as an initial solution until an optimal solution meets the termination condition; and generating a plurality of local paths according to the optimal solution corresponding to each initial solution, and determining the optimal path of the local area in the plurality of local paths. The path planning method provided by the invention can reduce the collision risk of the robot.

Description

technical field [0001] The present disclosure relates to the technical field of artificial intelligence, in particular, to a path planning method and device. Background technique [0002] Local trajectory planning generally refers to the space-time related driving paths or control instructions generated on the basis of a given global planning rough guide line. The Timed Elastic Band algorithm (translated as Timed Elastic Band Algorithm, hereinafter referred to as TEB) is a well-known local trajectory planning algorithm in the ROS environment proposed in recent years. It pursues real-time calculation and is essentially a nonlinear model predictive control method. The mainstream method based on the internal penalty function optimization algorithm forms a sharp difference. [0003] The main shortcoming of TEB is that it adopts the idea of ​​external penalty function to construct the optimization problem of its kernel, which leads to the fact that the feasibility of the solutio...

AI Technical Summary

Problems solved by technology

[0003] The main shortcoming of TEB is that it adopts the idea of ​​external penalty function to construct the optimization problem of its kernel, which leads to the fact that the feasibility of the solution result cannot be fundamentally guaranteed-this is why most classical methods that draw on the idea of ​​internal penalty function dare not The reason why we are willing to adopt the idea of ​​​​external penalty function. This shortcoming leads to the kinematic feasibility of the solution quality of the TEB algorithm, and the inherent hidden danger that the collision avoidance ability cannot be guaranteed. It is easy to cause many strange phenomena such as passing through the wall during driving. occur

Images