Mobile robot path planning method in unknown dynamic environment

Abstract

The invention discloses a mobile robot path planning method in an unknown dynamic environment. The method enables each position to be converted into a nerve cell, employs the activity values of nerve cells to represent environment changes, and ingeniously employs a biological motivation nerve network model to simulate a dynamic environment. The biological motivation nerve network model guarantees that the activity values of positive nerve cells can be transmitted outside to affect the whole state space, and the activity values of negative nerve cells just act on a local part. A decision at the next position point comprises a maximum nerve cell activity value and minimum robot corner factor, and an obtained path is the shortest one or about the shortest one. Moreover, the path is smooth, and the number of corners is smaller. The method does not have a problem of local minimum points, is small in calculation amount, is simple in implementation, and is good in adaptability in the unknown dynamic environment.

Description

Technical field [0001] The invention relates to a path planning method for a mobile robot, in particular to a path planning method for a mobile robot in a dynamic and unknown complex environment. Background technique [0002] With the development of technology, the application range of mobile robots is becoming wider and wider. Mobile robots have gradually expanded from regular indoor environments to more complex outdoor unstructured environments with irregular features. Path planning is an important part of robot navigation technology. It searches for an optimal or nearly optimal non-collision path from the initial state to the target state according to a certain performance index. [0003] Common path planning methods mainly include artificial potential field method, A* algorithm, neural network, etc. However, these methods always have certain defects for different practical problems. The artificial potential field method has a local minimum point problem, which will cause the...

AI Technical Summary

Problems solved by technology

Some neural network models can only deal with static environments, and BP neural network models can generate obstacle avoidance trajectories, but require a learning process and are slow

Glasius proposed a dynamic real-time obstacle avoidance model based on Hopfield network, which can solve the local minimum problem, but it is difficult to adapt to the high-speed dynamic environment

The bio-inspired neural network method proposed by SimonX.Yang in the article NeuralNetworkApproachtoDynamicCollision-FreeTrajectoryGeneration provides an effective solution for path planning, but under certain circumstances, there will be path misjudgment problems, and the path is not smooth enough

Images