Flight Trials of the Continuous Visual Navigation System

Robin J. Handley, Lawrence Dack and Paul McNeil

Abstract:	Accuracy and robustness are vital to present and future military air navigation. Both aircraft and medium/long-range missiles require precise position, velocity, and attitude data at all times in order to ensure mission success. Such data can be provided by a modern, integrated Inertial Navigation System (INS)/military-GPS navigation system - optionally equipped with Controlled Radiation Pattern Antenna technology, and employing tightly-coupled integration strategies, to minimise the likelihood and impact of GPS unavailability. However, the knowledge required to jam GPS is becoming public, and since the hardware to achieve this goal is basic, there is a significant risk to mission success even when the most sophisticated (and expensive) anti-jamming measures are employed. This leaves the integrated INS/GPS system to propagate its calibrated INS solution without further GPS fixes. Whilst this can be sufficiently accurate for a limited period of time (depending on the quality of the inertial sensors and of the integration algorithms), if jamming has occurred early enough in the mission, navigation accuracy may degrade too much to ensure mission success. One solution to this problem is Terrain Contour Navigation (TCN). However, current TCN systems™ position accuracy is poor relative to GPS and will only function over undulating terrain, which creates a heavy mission planning workload as well as reducing the number of viable routes that can be taken. In order to fill the performance/availability gap in INS/GPS or even INS/GPS/TCN systems, the Continuous Visual Navigation (CVN) System was developed by the navigation specialists at DERA (Farnborough) and Hi-Q Systems. This work has been funded by the United Kingdom Ministry of Defence, through the Applied Research Programme. CVN is a linear feature matching navigation technique, which overcomes the weaknesses in traditional Scene Matching navigation systems. It has been developed around the perceived way in which a human navigator might perform visual navigation; i.e. a combination of feature prediction and feature matching. As a result, CVN has no need for prior mission planning to identify way- points. In addition, in order to be robust to ‚bad fixes™, CVN utilises multiple, parallel Kalman Filters, which enable all likely potential matches for a given feature to be modelled and modified in the light of later measurements. Since first being proved as a feasible and accurate technique several years ago in off-line research, CVN has been improved and flown live on the DERA Tornado aircraft. During flight trials in August 2000, CVN™s preferred filter maintained accurate navigation over many passes of the UK database areas. On one flight, CVN was supplied with medium (aircraft) grade INS data; however on another flight CVN was instead supplied with low (missile) grade INS data, provided by a closed loop INS/GPS system using measurements from an LN200 Inertial Measurement Unit (IMU). In order to simulate GPS jamming, the GPS fixes were manually disabled prior to the aircraft entering the CVN database area. Partly as a result of the recent improvements to the hypothesis scoring scheme, and partly the improvements to the data interface logic, the real-time CVN results are the best yet. The most exciting aspect was the small degradation in CVN accuracy whilst using the missile grade data compared to when using the aircraft grade data. This was despite the fact that the missile grade position drift was two orders of magnitude higher than that of the aircraft grade system.
Published in:	Proceedings of the 2001 National Technical Meeting of The Institute of Navigation January 22 - 24, 2001 Westin Long Beach Hotel Long Beach, CA
Pages:	185 - 192
Cite this article:	Handley, Robin J., Dack, Lawrence, McNeil, Paul, "Flight Trials of the Continuous Visual Navigation System," Proceedings of the 2001 National Technical Meeting of The Institute of Navigation, Long Beach, CA, January 2001, pp. 185-192.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In