Abstract: The paper concentrates on the study of dangerous environment fire reconnaissance robot which developed by the laboratory. The robot is a tracked body, loaded an arm with a 3 degree of freedom, and the HD reconnaissance camera is attached to the end of the robotic arm. The contents of this paper are as in the known environment independent planning the manipulator obstacle avoidance and adjusting the camera sight direction to achieve lock on the environment in key areas. This paper firstly establishes manipulator kinematics model using exponential product, Minkowski and 3D convex hull algorithms are used for detecting the collision, Monte Carlo sampling method is used to sampling configuration space. It establishes virtual links at the end of camera and uses the collision detection algorithm to ensure the realization of sight without shelter. Simulation results show that using this algorithm can find the best camera pose to lock the key area and to ensure that view has no occlusion.

Key words: fire reconnaissance robot, collision detection, Monte Carlo, Line Of Sight（LOS） planning

摘要： 针对自主研发的危险环境消防侦察机器人展开研究。此机器人为履带式车身，装载一3自由度机械，高清侦察相机固定在机械臂末端。研究内容为在已知环境下自主规划机械臂规避障碍物，并调节相机视轴方向，实现对环境重点区域的锁定。首先对机械臂建立指数积（POE）正运动学模型，使用Minkowski和与三维凸包检测算法作为碰撞检测算法，使用蒙特卡洛方法在构型空间采样。再通过采样数据匹配末端视角方向，在末端建立虚拟连杆，使用碰撞检测算法保证实现视线无遮挡。仿真实验证明，使用此算法能够快速将相机视轴锁定至重点区域并保证无遮挡，且机械臂末端相机处于最佳位姿。

关键词: 消防侦查机器人, 碰撞检测, 蒙特卡洛, 视轴规划