Research on INS/GNSS/UWB Adaptive Robust ESKF Kinematic and Static Filtering Based on Cost Function

Abstract

Multi-source autonomous navigation-dependable decision making has a crucial impact on the overall performance of navigation systems. To solve the problem of overall system robustness caused by the intelligent-dependable decision making difficulties of navigation systems from different sources, on an unmanned ground vehicle (UGV) as the carrier, a new multi-source fusion algorithm based on cost function is proposed in this paper. The algorithm uses INS/GNSS/UWB as the sensor data source and is solved by using error-state Kalman filter (ESKF)-based kinematic and static multi-source filtering. After the RSS, positioning residual and positioning stability are selected as parameters and weighted, and the cost function is constructed. The structure of the filtering can be adapted according to the cost function in complex environments. Through mathematical simulation and comparative experiments, the positioning accuracy of the algorithm is improved by 75.9% and 74.44%, respectively, compared to federated filter and traditional ESKF-based kinematic and static filtering. It also improves the reliability, decision-making ability, and robustness of multi-source autonomous navigation system.

Keywords

• UGV
• cost function
• ESKF
• filtering
• multi-source autonomous navigation