Biological excitation robot complete traverse path planning method based on backtracking search

Abstract

The invention discloses a biological excitation robot complete traverse path planning method based on backtracking search. The method is combined with the advantages of a biological excitation neural network algorithm, a backtracking algorithm and a D * (D star) algorithm to enable a robot to achieve complete transverse path planning in a complicated environment. The method specifically comprises the steps that 1, the biological excitation neural network model is used for simulating a dynamic environment to guide the reciprocating motion of the robot; 2, when the robot deadlocks, firstly, the backtracking algorithm is used for rapidly finding a target position for escaping from deadlocking, and then, and the D* algorithm is used for planning a shortest path from a current position to the target position. According to the biological excitation robot complete traverse path planning method based on the backtracking search, not only are the advantages kept that a path generated by the biological excitation algorithm is smooth and corner turning is less, but also the speed for the robot to escape from the deadlocking is increased, and therefore the robot is capable of covering a whole work space rapidly. According to the biological excitation robot complete traverse path planning method based on the backtracking search, a local minimum problem does not exist, the calculated amount is small, the implementation is simple, and a good adaptability exists in a dynamic unknown environment.

Description

technical field [0001] The invention designs a mobile robot complete traversal path planning method, in particular relates to a mobile robot complete traversal path planning method in a dynamic unknown complex environment. Background technique [0002] Complete traversal path planning in an unknown environment is an important problem in the autonomous navigation of mobile robots. This problem often exists in applications such as automatic sweeping robots, demining robots, and submarine mapping robots. Full traversal path planning requires the robot's footprint or sensor detection to cover the entire workspace. When the robot works in an unknown environment, in order to completely traverse the workspace, it is necessary to construct a map in real time and dynamically plan a path. [0003] Common methods for complete traversal path planning include artificial potential field method, template model method, A* algorithm, etc. The artificial potential field method is widely use...

AI Technical Summary

Problems solved by technology

Simon X.Yang(Chaomin Luo,Simon X.Yang.A BioinspiredNeural Network for Real-Time Concurrent Map Building and Complete CoverageRotot Navigation in Unknown Environment.IEEE TRANSACTIONS ON NEURAL NETWORKS,VOL.19,NO.7,JULY2008.) proposed A neural network algorithm based on biological stimulation. When the robot uses limited sensor information for path planning, it establishes an environmental map composed of square or rectangular units through neural dynamics, which can effectively solve the problem of point-to-point path planning. The method also has some shortcomings, such as high path overlap rate between separation areas, non-optimal path, and long waiting time when escaping from the dead zone.

Images