基于UKF数据融合的勘测机器人定位方法

Location Method of Survey Robot Based on UKF Data Fusion

传统里程计定位方法在室内环境下误差大,难以通过单一传感器对机器人精准定位。以四轮侧滑移动机器人(SSMR)为研究对象,提出一种无迹卡尔曼滤波(UKF)组合导航方法。该方法融合里程计、惯导、超宽带(UWB)传感器数据。利用超宽带定位精度高的特点,减少里程计航迹推算的打滑及累计误差影响。实验结果表明,基于UWB与里程计融合的多传感器定位方法误差明显降低,提高了定位准确性,能够满足勘测机器人在地下非GPS环境精确定位的要求。

The traditional odometer positioning method has a large error in an indoor environment,and it is difficult to accurately locate the robot using a single sensor.Taking the four-wheel skid-steering mobile robot(SSMR) as the research object,an unscented Kalman filter(UKF) integrated navigation method is proposed.This method integrates odometer,inertial navigation,and ultra-wideband(UWB) sensor data.The high accuracy of ultra-wideband positioning is used to reduce the influence of slip and accumulated errors in the odometer track calculation.The experimental results show that the multi-sensor positioning method based on the fusion of UWB and odometer significantly reduces the error,improves the positioning accuracy,and can meet the requirements of accurate positioning of the surveying robot in the underground non-GPS environment.