Research And Implementation Of Path Planning Algorithm For Mobile Robot In Complex Environment

With the continuous progress of science and technology,the intelligent level of mobile robots is also getting higher and higher.Today,mobile robots have been widely used in many fields.Path planning is one of the core contents in the field of mobile robot research and is an important manifestation of its level of intelligence.In this thesis,starting from the safety and effectiveness of mobile robots,in-depth research on the path planning of robots is researched in the face of complex and changing working environments.First,aiming at the problem of unreachable target and local minimum value in the artificial potential field method in specific environment,a fuzzy artificial potential field algorithm is designed by fusing fuzzy control with the artificial potential field method.The artificial potential field method is used to simply model the local environment of the robot,and then use the obtained repulsive force and the angle between the target,the obstacle and the robot as the input of the fuzzy controller,for specific situations,formulate expert rules to obtain the deflection angle output by the fuzzy controller,and the sum of the deflection angle and the gravitational angle of the artificial potential field method is the final movement direction of the robot.Experiments verify that the fuzzy artificial potential field algorithm effectively solves the problems of unreachable targets and local minimums in artificial potential field method,compare with grid method and genetic algorithm in multi-static obstacle environment,prove that the proposed algorithm has better effectiveness.Then,using the designed fuzzy artificial potential field algorithm to generate a training data set for neural network supervised learning.At the same time,a new scalable neuron network is designed,a neural network model is built,and the neural network model is trained using the data set generated by the fuzzy artificial potential field algorithm.Finally,experiments are conducted in the built dynamic complex simulation environment,and compared with dynamic window method and ant colony algorithm.The experimental results show that the trained neural network model has good performance.Further,aiming at the unknown uncertain obstacles and high-speed approaching emergency obstacles in the dynamic and complex working environment of mobile robots,this thesis combines neural network model to design a real-time decision-making system under the consideration/reaction hybrid robot architecture.The real-time decision-making system classifies obstacles according to the specific situation encountered by the robot,and decides the direction and speed of the robot's movements according to different types of obstacles,improving the safety of the robot to a certain extent.Finally,building a targeted dynamic and complex environment to verify the good performance of the proposed real-time decision-making system.In the same complex environment,it is compared with the dynamic window method,ant colony algorithm and the designed neural network model,the experimental results show that the real-time decision system has better security performance.Finally,building a mobile robot experimental platform for experimental verification.This thesis chooses RikiRobot robot to build the experimental platform,and set up a targeted experimental environment.The experimental results show that the robot can effectively avoid all types of obstacles set in the environment and reach the target position safely.The experiments verify the effectiveness of the proposed path planning method and its feasibility in practical applications.