Vehicle Constraints Enhancement for Supporting INS Navigation in Urban Environments

Abstract

The complementary nature of INS and GPS can be used advantageously in navigation system design as long as GPS measurements are available. However, circumstances of complete GPS denial may occur in urban environments or in signal blockage scenarios. In such cases, the INS navigation solution drifts over time due to its inherent bias. To circumvent the INS drift, it is usually fused with external sensors. Nonetheless, use of external sensors increases the overall system cost. To reduce the need for external sensors, incorporation of vehicle constraints into the estimation process has been recently proposed. In order to improve the effect of these constraints, we propose implementing vehicle constraints both into the system and measurement models thereby enhancing the estimator performance. The paper demonstrates the contribution of the proposed approach via several field experiments. In all experiments, introducing system dynamics constraints reduced the navigation errors, compared to using only measurement constraints.