The Structure Design And Motion Planning Of Field Rescue Robot

The huge destructiveness and uncontrollability of the disasters cause huge economic losses.At present,the best way to deal with a disaster is to take rescue actions as soon as possible to reduce the loss and casualties when a disaster happens.However,there are some unsteady factors that are not suited for people to implement rescue.It's necessary that the rescue robots assist or replace the rescue team to do rescue work.But the rescue robots with fixed structure and single function are difficult to adapt to complex rescue environment.For the problem,the paper proposes a new modular self-reconfigurable rescue robot structure design and motion planning.The paper is structured as follows:(1)A new modular self-reconfigurable robot for search and rescue is designed.The modular robot is composed of two main mechanical structures,a planetary wheel for moving and the double locking system as the docking structure.Through the planetary wheel structure,the robot can cross obstacles flexibly in a disaster environment.The docking structure with double unlocking function not only makes the robot have self-repair ability and improve reliability,but also allows the modular robots to connect or disconnect according to the actual situation and the task requirements.(2)Heuristic path search algorithm is optimized in heuristic function.Through experiments,the most suitable heuristic functions of A star,JPS and bidirectional A star are selected.After the three optimized algorithms are compared,the bidirectional A star algorithm with the highest search efficiency and the shortest optimal path is selected.(3)An obstacle detection algorithm based on 3D point cloud is proposed.The Real Sense D435 as the sensor is used to gather the environmental information.The point cloud gathered is divided into two main types: ground point cloud and obstacle point cloud.Firstly,the ground point cloud plane model is generated from these point clouds.Based on the ground point cloud plane model,the clustering algorithm is used to classify different obstacles for the remaining point clouds.(4)In order to verify the feasibility of the above algorithm,a self-reconfigurable modular rescue robot hardware platform was fabricated and assembled and a ROS-based robot navigation system was built.Then,the motion planning of the robot was realized in the obstacle environment built in the laboratory.To sum up,the new self-reconfigurable modular rescue robot designed in this paper can adapt to various rescue environments and use Real Sense D435 to collect surrounding environment information and achieve path planning.