基于自控测距法的机器人位姿估计  

Improving estimations of a robot's position and attitude with accelerometer enhanced odometry

作　　者：王晓宇[1] 闫继宏[2] 徐莉红 

机构地区：[1]航天东方红卫星有限公司,北京100094 [2]哈尔滨工业大学机器人研究所,黑龙江哈尔滨150080 [3]304厂科研2所,山西长治046012

出　　处：《智能系统学报》2009年第2期169-174,共6页CAAI Transactions on Intelligent Systems

基　　金："863"国家高技术研究发展计划资助项目(2006AA04Z245)

摘　　要：针对两轮自平衡机器人运行过程中遇到打滑、越障、碰撞等异常事件,测程法进行位姿估计失效的情况,提出一种自控测距法.结合改进的Gyrodometry方法,通过融合码盘、陀螺仪与加速度计数据对机器人的位姿进行估计,实现了机器人的准确定位,解决了非系统定位误差对机器人位姿估计的影响,降低了陀螺仪、加速度计固有漂移的不利影响,提高了两轮自平衡机器人的定位精度.实验结果显示与测程法误差相比自控测距法位置误差降为1/3,方向误差降为1/6,验证了自控测距法的有效性.Unpredictable events can occur in the navigational processes of two-wheeled self-balancing robots, such as wheel-slippage, interaction with obstacles, and bumps. These invalidate position and attitude estimates. There-fore, an improved method was developed for accurate positioning of robots. The accodometry method adds accelerometer data to the improved Gyrodometry method, fusing encoder data with inertial sensor data. This method eliminates the effect of non-systematic errors onorobot position and attitude estimation and reduces ill-effects due to inherent drift of the gyro and accelerometer. As a result, positional accuracy is greatly improved. Experimental results showed that positional errors decrease by one-third and orientation errors decline by one-sixth, verifying the effectiveness of the accodometry method.

关 键 词：位姿估计 自控测距法 Gyrodometry法 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]