A path planning method for robots in complex and narrow environments

Abstract

The invention discloses a path planning method for a robot in a complex and narrow environment, which combines the RRT-Connect algorithm and the reinforcement learning algorithm, retains a certain randomness of the random tree, and selects an appropriate step size according to the density of obstacles, both It can realize the rapid random exploration of the robot in the dense environment of obstacles, and the rapid passage in the sparse environment of obstacles, and increases the orientation of the random tree distribution, improves the convergence speed, and can improve the planning performance in the interaction process with the environment, so that the planning The path obtained is close to the optimal path and avoids falling into a local minimum; compared with the original standard algorithm, the improved algorithm of the present invention has a better planned path and takes less time.

Description

technical field [0001] The invention relates to a path planning method, in particular to a path planning method for a robot in a complex and narrow environment. Background technique [0002] Path planning is an important direction in the field of intelligent robot research. Robot path planning is divided into global path planning and local path planning. At present, the commonly used local path planning methods are mainly based on artificial potential field method, grid method, etc., and combined with genetic algorithm, fuzzy logic algorithm, neural network and other intelligent search algorithms to improve search efficiency, system robustness and adaptability. [0003] Fuzzy method is a kind of planning method usually adopted in online planning, including modeling and local planning. But fuzzy logic method fuzzy logic rules need to rely on human experience, which restricts the wide application of this algorithm in the field of robot path planning. [0004] Genetic algorit...

AI Technical Summary

Problems solved by technology

[0004] Genetic algorithm is an algorithm that simulates the genetic mechanism of nature and the theory of biological evolution to search for the optimal solution. It is widely used in the research of robot path planning, but the ability of genetic algorithm to adapt to a wide range of environments is weak, and the speed of model evolution is difficult to control. , so the algorithm has certain limitations. In addition, the genetic algorithm is easy to fall into the local optimal solution, which leads to the premature phenomenon of the algorithm.

[0005] The neural network method uses a structure similar to the synaptic connection of the brain's neural network to express and process information, but the neural network needs to train a large number of samples in the learning process, and needs to manually set more empirical parameters, so it is applied in the robot There are certain limitations in the field of path planning

Images