Robust adaptive SINS/CNS integrated navigation algorithm for Mars rovers with experimental verification

• Published in: Measurement : journal of the International Measurement Confederation Vol. 236; p. 115087
• Main Authors: Li, Zhihao, Zhan, Yinhu, Du, Haijian, Shen, Yang, Chen, Shaojie, Zhang, Chao
• Format: Journal Article
• Language: English
• Published: Elsevier Ltd 15.08.2024
• Subjects: Adaptive factor; Mars rover; Mounting error angle; Robust estimation; SINS/CNS integrated navigation; Robust estimation; Mounting error angle; Adaptive factor; SINS/CNS integrated navigation; Mars rover
• ISSN: 0263-2241
• DOI: 10.1016/j.measurement.2024.115087

•RASCIN algorithm created for integrating SINS/CNS data in Mars rover navigation.•Mitigates gross errors using robust estimation theory and adaptive factors.•Achieves 92.48 % reduction in position errors versus pure inertial navigation.•Enhances three-axis attitude accuracy and real-time navigation capabilities.•Validated algorithm effectiveness via field trials, with errors at 1.22 % distance. High-accuracy, real-time autonomous navigation technology is crucial for enabling Mars rovers to successfully conduct scientific exploration missions. However, previous studies typically use only the single-axis attitude information output by the Mars rover and do not consider the significant impact of astronomical attitude measurement gross errors and kinematic model gross errors on navigation accuracy. Therefore, this paper introduces a robust adaptive strapdown inertial navigation system (SINS)/celestial navigation system (CNS) integrated navigation algorithm, denoted as the robust adaptive SINS/CNS integrated navigation (RASCIN) algorithm. This algorithm integrates the highly accurate three-axis attitude information provided by the CNS with the acceleration and angular velocity data obtained from the SINS. It also introduces robust estimation theory to mitigate the influence of gross errors in celestial measurements by adjusting the weight matrix. In addition, the algorithm enhances system robustness by incorporating adaptive factors designed to mitigate errors originating from the motion model within the SINS/CNS fusion. Simulation experiments yield compelling results, showing that compared to pure inertial navigation and extended Kalman filter (EKF) solutions, the RASCIN algorithm significantly improves the accuracy of three-axis attitude angles; it also reduces the closed-loop position errors by 92.48 % and 72.92 %, respectively, compared to these solutions. Furthermore, experimental trials conducted on Earth’s surface affirm the effectiveness of the algorithm in controlling the daily navigation error of the Mars rover, limiting it to within 1.22 % of the distance traveled. Moreover, the RASCIN algorithm substantially reduces closed-loop position errors, achieving a reduction of 95.10 % compared to pure inertial navigation and 90.63 % compared to the EKF solution. Importantly, the algorithm also greatly enhances real-time navigation capabilities.