Mars rover navigation using pseudolite transceiver arrays: Network-based ranging and extended self-calibration algorithm.

摘要

A Self-Calibrating Pseudolite Array (SCPA) is a self-deployable GPS pseudolite-based local-area navigation system that can be used on future robotic and manned planetary explorations. By utilizing bidirectional pseudolite transceivers, the SCPA can provide common global positioning to multiple agents working in a local designated area of a remote planet, including all the benefits of satellite-based carrier-phase differential GPS, such as drift-free, centimeter-level, and three-dimensional positioning, without requiring a satellite constellation above the remote planet.; Previous work has demonstrated that changing the relative array geometry by moving a roving transceiver unit enables the SCPA to self-calibrate both the array locations and the rover trajectory to centimeter-level accuracy. This self-calibration capability has overcome the difficulty of autonomous robotic deployment of the pseudolite-based navigation system on remote planets, eliminating the need for accurate a priori position information or precise placement of the array.; However, early field trials raised the issue of robustness due to pseudolite signal dropouts caused by multi-path fading, cycle slips, or losing line-of-sight. In order to complete the self-calibration process successfully, the SCPA was required to maintain all the signal locks over the entire calibration maneuver; the lack of necessary ranging measurements due to any signal dropout requires the process to start over.; This dissertation solves this robustness issue by two new methods: network-based ranging and an extended self-calibration algorithm. The combination of the two algorithms yields a dual-fault-tolerant system, tolerating at least any two simultaneous dropouts intermittently during the calibration process while still operating in the minimum one-mobile three-stationary transceiver configuration with single-frequency pseudolite signals. The resulting improved robustness has been demonstrated in field trials using the K9 Mars rover platform operated in the Marscape at NASA Ames Research Center. The experimental results validate that the new algorithms can reliably complete the self-calibration process while experiencing severe signal dropouts, yet still achieve centimeter-level calibration performance.