| With the increasing demand of indoor intelligent service robot application,mobile robot as its mainstream has become a research hotspot.Autonomous navigation of mobile robot is the basic requirement of intelligence,and mobile robot environment perception and path planning are the core of autonomous navigation technology.Sensors are the main tools for mobile robots to perceive the world.However,a single sensor has limitations and is difficult to meet the complex and changing indoor environment.The positioning error of mobile robot in indoor environment is great.In path planning,the path planned by traditional method is difficult to adapt to the motion mode of mobile robot.In this paper,the method of laser and vision integration mapping is used as the way of environment perception in the navigation process of mobile robot.The extended Kalman filter algorithm(EKF)is used as a real-time localization method for mobile robots.A hybrid optimization A* algorithm and dynamic window algorithm(DWA)are used as navigation path planning methods for mobile robots.The main research work is as follows:(1)To address the problem that it is difficult for a single sensor to obtain real-time accurate environmental information in the process of mobile robot navi gation.By converting the 3D information collected by the depth camera into 2D lidar data and using Bayesian method at the decision level,the lidar data is fused with the depth camera data to build a map to compensate for the lack of information caused by the limitations of the single sensor itself.The experiment shows that the method effectively improves the detection accuracy of local environment.(2)The indoor environment is addressed to the problem that the wheeled odometer localization method has serious non-systematic cumulative errors,resulting in the mobile robot not being able to locate accurately.The real-time localization of the mobile robot is achieved by extending the localization method of probabilistic estimation of Kalman filter algorithm,based on the kinematic model and observation model of the mobile robot.The experiment shows that the method can effectively reduce the non-systematic errors and improve the self-positioning accuracy of the mobile robot.(3)In the path planning algorithm,the traditional A* algorithm is limited by the search method,and the planned path has a lot of meaningless inflection points and a large Angle of the inflection points,which leads to the mobile robot’s frequent turns in the process of motion,consuming a lot of time and even getting out of the path.The number of inflection points is reduced by optimizing the search method.The cubic uniform B-spline algorithm was used to optimize the remaining inflection points and reduce the Angle of the inflection points.In complex environment,the evaluation function of DWA algorithm is improved by optimizing A* algorithm,and A mixed A*＿DWA algorithm is formed,which solves the problem that local path planning DWA algorithm is easy to fall into local optimal and cannot find the target point.Experiments show that this algorithm has efficient path planning and obstacle avoidance ability.(4)Build the experimental environment in the simulation experimental platform to verify the feasibility of the navigation system.In the real environment,the effects of laser SLAM building and fused SLAM building are compared and analyzed,as well as the ability of A* algorithm,DWA algorithm and hybrid A*＿DWA algorithm for path planning and obstacle avoidance.The experimental results show that the navigation approach proposed in this paper has significant superiority and reliability. |
