Multisensor fusion for an experimental airship based on strong tracking filter

摘要

To improve the independent ability of attitude determination and positioning for an unmanned experimental airship platform, a micro inertial measurement system (MIMS) is expected to integrate with the existing system, which incorporates a digital magnetic compass and a differential pseudorange GPS receiver. The navigation error of the low-precision MIMS will be calibrated using nondrift DGPS receiver and magnetic compass. This paper proposes an adaptive strong tracking filter to perform multisensor fusion to assure state-error estimation of convergence under some uncertain conditions. These uncertainties include model simplification, unknown microsensor stochastic characteristics, a large-scale initial filtering parameter variation, and state sudden change. Monte Carlo simulations demonstrate the filter has strong robustness to all the uncertainties mentioned above. By this filtering approach, the navigation errors of MIMS are limited to a certain range. Accordingly, the whole integrated measurement system will respond to dynamics, and its automotive navigation ability is also enhanced.