Research On Key Technologies And Fitness Experiments Of Search And Rescue Robot System Under Complex Terrain Conditions

Fast and effective battlefield rescue capability is a fundamental guarantee for the maintenance and regeneration of military combat power.Facing the casualty search and rescue(SAR)tasks in jungles,mountainous areas,urban and other terrain conditions,how to maximize the rescue capability and improve the efficiency of search and rescue mission execution through the research,enhancement,and application of key technologies of search and rescue robotic systems has become a hot issue of general concern to the world’s military powers and has important academic and engineering practical value.A series of key technologies,including the adaptation degree of SAR robot system under complex terrain conditions,the task assignment of SAR robot system based on the adaptation degree,the life detection algorithm of multi-sensor information fusion under complex terrain environment,and the path planning algorithm of SAR robot under multi-terrain conditions,are studied in this paper.Meanwhile,a SAR robot system for complex terrain conditions is built,and a series of fitness experiments are conducted.To quantitatively evaluate the adaptability of SAR subjects(rescuers,SAR robots,task loads,etc.)to SAR tasks in SAR tasks under complex terrain conditions,a SAR adaptability calculation method combining neural network and hierarchical analysis is proposed to quantitatively define and fuse multiple factors such as task,terrain,and distance that affect the adaptability of SAR tasks,and realize the adaptability of SAR subjects to SAR tasks quantitative analysis and calculation.To improve the task allocation efficiency of the SAR robot system under complex terrain conditions,a SAR robot system task allocation algorithm based on SAR adaptability is proposed,which fully considers various influencing factors such as task type,terrain conditions,and distance conditions.Firstly,the formal description method is used to describe each SAR subject and the tasks to be executed,then the adaptation degree of the SAR subject to a specific task is calculated by the SAR adaptation degree,and finally,the task is optimally assigned by the hierarchical reinforcement learning method.Compared with the traditional task assignment method,an 87% improvement in task assignment efficiency was achieved.To improve the accuracy of life detection in complex terrain environments,a life-detection algorithm with multi-sensor information fusion is proposed.Radar information,RGB video information,single-frame RGB image information,and infrared image information are used as inputs and Ordered Weighted Average(OWA)is used to fuse multiple information at the decision level.The information fusion is used to achieve high accuracy of life detection in complex terrain environments.To address the problem that current robot path planning algorithms cannot consider the adaptability of different robots to different terrains,an adaptation-based path planning algorithm for search and rescue robots under multi-terrain conditions is proposed.Based on the robot’s terrain adaptability,the robot path planning algorithm based on reinforcement learning and the potential field method is designed to design the most suitable movement path for different robots under multi-terrain conditions concerning their motion ability.The experimental results show that the adaptation-based robot path planning algorithm achieves a 23% and 31% improvement in mobility efficiency on tracked and wheeled robots,respectively,compared with the traditional method.Synthesizing the above research results,this paper builds search and rescue scenarios for complex terrain adaptation experiments under various terrain conditions such as urban and field environments.The experimental projects include life detection,identity recognition,and casualty evacuation.For each experimental project,three search and rescue modes,such as pure manual mode,human-machine assisted mode,and human-machine cooperative mode,are used for comparison and analysis.The experimental results show that the human-machine collaborative SAR mode achieves no less than 88% and 31% improvement in SAR effectiveness in urban and field environments,respectively.The relevant research and experimental results are of great reference significance for future search and rescue tasks in complex terrain conditions.