2D lidar and ultra-wideband fusion location algorithm based on landmark assistance

In an indoor environment where global positioning system (GPS) signals are severely attenuated, ultra-wideband (UWB) and 2D lidar are widely used in the autonomous positioning of mobile platforms. However, the presence of nonline-of-sight (NLOS) environments can lead to large errors in UWB positioning, and 2D lidar will increase the cumulative error due to the loss of accuracy in sparsely textured scenes. In order to reduce the positioning error, a UWB and 2D lidar fusion positioning algorithm based on the assistance of a few landmarks is proposed in this paper. Considering the colored noise of lidar location data, a Kalman filter algorithm based on cumulative error analysis is proposed. First, the lidar error curve is fitted by the least-square method, and then the relationship between the noise covariance matrix and the lidar cumulative error function is established by introducing the scale factor, which is substituted into the Kalman prediction equation. Experimental results show that the proposed multi-sensor fusion localization algorithm is feasible, and compared with the single localization method, the proposed fusion algorithm can significantly improve the localization accuracy; matching landmarks can achieve a positioning accuracy of 0.15 m, which is about 24.4% lower than the root mean square error of traditional Kalman filter.