Autonomous Exploration For Mobile Robots In Rescue Environments

Urban search and rescue(USAR)robots face a very complex and urgent task.The robot must explore the entire unknown environment in a short amount of time and locate survivors after a disaster.One goal in rescue robotics research field is to make the robot have the ability to perform urban search and rescue autonomously.In this paper,autonomous exploration is researched in rescue environments and corresponding algorithms are proposed for mobile robots.The algorithms of 3D pose estimation and elevation mapping,autonomous exploration decision and path planning in rescue environments have been realized and tested in a variety of typical terrains,and applied on the autonomous exploration system of our NuBot rescue robot,which achieved outstanding performance in the RoboCup Rescue Robot League's autonomous exploration competition,won the championship of RoboCup 2016 China Open Rescue Robot League,and entered the finals of RoboCup 2016 Rescue Robot League and received the Best in Class Small Robot Mobility.A 3D pose estimation and elevation mapping algorithm based on multi-sensor fusion was proposed in this paper,by combining the single line laser range finder,RGB-D camera and inertial measurement unit.The deficiency of 3D information in the 2D SLAM approach is made up with the robot's 3D location and elevation map.The frontier based autonomous exploration decision algorithm was improved in this paper.A detection range based frontier detection algorithm was proposed for urban search and rescue,and a path cost calculation method using gradient cost was proposed to effectively improve the adaptation of 2D path planning in the complex terrain,which makes sure that the autonomous exploration is safe and effective for our NuBot rescue robot.