희소 앵커 환경에서 UWB 와 IMU 센서 융합기반 실내 측위 연구

Abstract

High precision location information plays an important role in military, industrial and commercial applications. In fact, indoor positioning demand is much larger than outdoor positioning demand, [Figure 1. 1] shows part of the application of indoor positioning. But, because UWB positioning accuracy will decrease in a non-line-of-sight (NLOS) environment, the number of base stations needs to be increased. Also since the deployment of Anchors requires a lot of manpower and material resources, reducing the number of anchors has become a specific requirement. As MEMS technologies continue to advance more and more IMU sensors are integrated into the phone. So, in order to achieve a simpler setup and a more widely applicable indoor positioning scheme, this paper proposes a method on the UWB/IMU sensor based indoor positioning system in a sparse environment. For the PDR only positioning, the various estimation model of step size and direction were compared. Although the gyroscope positioning has the best accuracy, due to the accumulation of time integration errors, magnetometer, accelerometer and gyroscope were fused by the Kalman Filtering. With these efforts, under the 3.63m×7.67m 2-D environment, the average positioning error could be obtained up to 0.6m.To fuse IMU with UWB, particle filtering was employed. Two positioning experiments are conducted: a single anchor & 2-D PDR particle filter (unknown start position) and sparse anchors & 3-D PDR particle filter (known start position). In the first experiments, under the 3.63m×7.67m 2-D environment, the best accuracy can reach up to 0.5 m by using particle filtering. In the second experiments, under the 60.8m×92.4m×1.5 m 3-D environment, the average accuracy of PDR only is 8m because the magnetometer is affected by the indoor environment. By fuse UWB through particle filter, the positioning accuracy reach up to 3m.