Design And Implementation Of Autonomous Navigation System For Two-Wheel Differential Robot

With the improvement of computer performance and the improvement of sensor technology,robot navigation technology has been widely used in various industries.Although the robot navigation technology has been initially applied in the daily environment,there are still some problems that are difficult to solve,and there is still a certain distance from the real commercial landing.For example,robot kinematics and dynamic constraints can make the robot motion planning problem rise in dimension and difficult to solve.The unmanned system on the highway transforms the state of the vehicle to the frenet coordinate system centered on the lane line to turn the planning problem into a simple mathematical form.However,in an unstructured environment,smooth lane lines with less curvature are often difficult to obtain,and highway unmanned frames are not suitable.In order to achieve the real-time nature of the navigation system,the unstructured environment usually uses a reactive path planning method to control the robot,but the reactive planner is easy to fall into the local optimum.Aiming at the two problems of real-time and obstacle avoidance flexibility of robot navigation system,this paper designs a robot navigation system that combines two functions of autonomous following and autonomous line tracking with the two-wheel differential robot as the research carrier.The main contents of this paper are as follows:(1)The flow structure of the two navigation modes of robot following navigation and linetracking navigation is introduced,as well as the required software platform and hardware carrier.(2)A robotic following navigation system for dynamic environment was designed and implemented,including the detection and tracking of the target personnel and the robot following navigation.Aiming at the problem that the target personnel in the occluded environment are difficult to identify,a laser and AOA(Angle of Arrival)positioning tag are proposed as the sensor of target tracking algorithm.The experimental results show that the method can solve the problem of target recognition under occlusion.And the target tracking result is more stable than the single sensor method.Aiming at the problem that the evaluation function has local optimum in DWA(Dynamic Window Algorithm)algorithm,a FMM＿DWA algorithm considering the heading angle of the robot is proposed.The experimental results show that the FMM＿DWA algorithm can also flexibly avoid obstacles in narrow areas,and the follow-up control effect is smoother than existing ones.(3)A robotic line-tracking navigation system for unstructured environment is designed and implemented,including path smoothing,path tracking and obstacle planning.Path smoothing converts discrete path points to Bezier curves,limiting the maximum curvature of the global path.Path tracking combines forward-looking speed planning algorithm and Pure Pursuit algorithm to ensure the safety of navigation;in addition,by calculating the curvature and velocity constraints,the tracking path can be avoided by satisfying the kinematics and dynamics of the robot.Aiming.at the problem that the existing robot navigation system is difficult to meet in real time,the obstacle planning method expresses the path planning problem as convex quadratic programming and solves it quickly to meet the real-time navigation requirements.