月球机器人仿真中一种消除轮地交互误差的方法

利用虚拟现实技术虚拟出月球机器人在月面上的作业环境和作业过程,是提高机器人作业的安全系数和工作效率的一条有效途径。在3D重建得到的虚拟月面环境中,如果采用通常的单纯基于运动学(或者动力学)模型的仿真方法,对机器人的作业和运动进行虚拟,那么机器人与地形交互的过程中容易产生接触偏差。而且,随着仿真时间的推进,这种接触偏差会逐渐积累并不断增大,进而严重影响仿真测试的精度和效果。为了消除月球机器人仿真中的轮地交互误差,在分析误差来源的基础上,提出了基于运动学优化的解决方法。最后利用实际的虚拟现实仿真系统,验证了所提出方法的有效性。

In order to secure the safety and efficiency of the lunar rover, it is an effective method to simulate the working process and the environments on lunar surface with virtual reality technology. However, there are errors in the interaction between the rover and the virtual lunar surface based on 3D-reconstruction, when using a general simulation method only based on kinematics (or dynamics) models. In particularly, the errors will be accumulated with the advance of time. As a result, the simulation precise and visual effect will be deteriorated. In order to eliminate the wheel-terrain interaction errors in the lunar rover simulation, the error causes were analyzed, and a method of optimization based on kinematics was described to solve the problem. The experiment results on the lunar rover simulation system validate the approach.

中科院机器人学研究室基金课题宜人化地面遥操作系统(A040118)