| With the popularity of wheeled robots in daily life,people pay more and more attention to the safety of their work.How to efficiently and autonomously avoid obstacles in the environment has always been a hot and difficult point in the field of mobile robots.The quality of obstacle avoidance results will also indirectly affect the results of navigation and positioning.In order to achieve accurate obstacle avoidance,it is necessary to have a good perception of the surrounding environment.Using multisensor information fusion technology,the advantages of each sensor can be better utilized,and more comprehensive external environment data can be obtained,thereby improving the accuracy and rapidity of the response during obstacle avoidance.Therefore,in this paper,in complex environments,wheeled robots may not only meet the needs of obstacle avoidance in static environments,but also need to conduct realtime avoidance of unknown obstacles.It has important theoretical significance and practical value.The research contents of this paper are as follows:(1)Build a wheeled robot obstacle avoidance system platform by selecting different hardware modules to combine with each other and designing corresponding software programs,and design an overall scheme of the obstacle avoidance system.(2)Due to the diversity of obstacles and the limitations of the sensor itself,a single sensor cannot comprehensively detect the information of surrounding obstacles.This paper proposes a method based on the combination of RBF neural network and regional fusion rules.The obstacle detection method can effectively identify obstacles in the environment and obtain accurate distance information.The simulation experiment proves that the accuracy of obstacle distance measurement can be improved.(3)In view of the shortcomings of the traditional A* algorithm,such as long search time,many path turning points,and running too close to obstacles,an improved method is proposed,which reduces the number of search directions from 8 to 5,and speeds up the search time;The key point strategy is used to delete unnecessary turning points and improve the smoothness of the path;a safety factor is added to the evaluation function to improve the safety of obstacle avoidance.And a new evaluation function is designed to integrate the improved A* algorithm with the DWA algorithm,so that the robot can avoid unknown obstacles while obtaining a global safe obstacle avoidance path.Simulation and comparison experiments are carried out to verify the effectiveness of the fusion algorithm.(4)For complex indoor scenes,experiments are carried out with the self-built obstacle avoidance system,which shows that the obstacle avoidance system based on multi-sensor information fusion designed in this paper can detect obstacles comprehensively and efficiently.On the basis of the safe path,unknown static and dynamic obstacles can be avoided. |
