Research On Multi-floor Path Planning Of Security Robot

With the development of economy and the improvement of labor value,the traditional security has gradually shown its disadvantages.The introduction of robot technology and artificial intelligence will be the development direction to improve security efficiency.On the basis of analyzing the path planning technology of security robot at home and abroad,this paper makes a further study on improving the multi-floor patrol capability of security robot.The main contents are as follows:1.An multi-floor navigation system based on ROS is proposed.After the completion of the patrolling,the security robot reaches the transport point such as elevator or escalator and transmits to the target floor,and then switches the map of the target floor to realize multi-floor patrolling.2.An improved A~* algorithm named BCA~* for variable parameter Bezier curves is proposed in view of the shortcomings of A~* algorithm in global path planning,such as many broken lines,many turning times and large cumulative turning Angle.On the basis of the traditional A~* algorithm,each node is evenly divided to obtain A series of path node coordinates.The planned path is shortened by removing the redundant nodes in the middle.Finally,variable parameter Bezier curve is used to curve the path inflection point to plan A smooth and shorter path.Simulation and experimental results of Turtlebot2 robot show that the path length of this algorithm is reduced by 20.75% compared with the traditional A~* algorithm.After solving the problem of turning Angle,the running time of the robot was shortened by 43.28%.3.On the basis of global path planning,an improved TEB local path planning algorithm is proposed.The traditional Timed Elastic Band(TEB)algorithm does not consider the constraints of the robot's stability during operation and the minimum turning radius of the wheeled robot.In this paper,the constraints of the minimum turning radius and impulse limit of the robot are added to improve the stability of the robot in the process of running.At the same time,aiming at the problem that traditional TEB is easy to fall into local minimum,a path selection strategy considering the right-handed behavior criterion is proposed to satisfy the security patrol of security robot in the densely populated area.According to the RVIZ simulation experiment analysis based on ROS,the cumulative rotation Angle is58.27% lower than the traditional TEB algorithm,the kinetic energy consumption is54.95% lower than the traditional TEB algorithm,and the probability of right row is40% higher than the traditional TEB algorithm.4.Based on the environment of the experimental building,the Gazebo physical simulation environment and ROS robot experimental platform were built to complete the architecture construction and function realization of the improved algorithm.At the same time,the multi-floor navigation system was used to realize cross-floor navigation.Finally,the experimental results were analyzed.The practice shows that the constructed multi-floor navigation system can enable the security scout robot to realize multi-floor patrol,and the improved algorithm can efficiently provide the robot with path planning,with high robustness and real-time performance.