Augmenting Kalman Filtering with Parallel Cascade Identification for Improved 2D Land Vehicle Navigation

Abstract

Land vehicle positioning relies mostly on satellite navigation systems such as the Global Positioning System (GPS). However, GPS signals may be degraded or suffer from blockage in urban canyons and tunnels, and the positioning information provided is interrupted. One solution for such a problem is to integrate GPS with an inertial measurement unit (IMU) and the navigation solution is achieved using an estimation technique which is traditionally based on a Kalman filter (KF). In order to have a low cost navigation solution for land vehicles, MEMS-based inertial sensors are used. To further reduce the cost a reduced inertial sensor system (RISS) which consists of only one gyroscope and a speed sensor is integrated with GPS. The position and velocity errors can be estimated by a KF relying on RISS dynamic error model and GPS position and velocity updates. However, low-cost MEMS sensors suffer from complex error characteristics, which are difficult to model by the linearized KF models. The positional accuracy of the integrated system can be improved using Parallel Cascade Identification (PCI) that is cascaded with the KF. The proposed augmented KF-PCI method can handle both linear and nonlinear system errors as the linear parts of the errors are modeled inside the KF and the nonlinear residual RISS errors are modeled by PCI. The performance of this method is examined by road test trajectories in a land vehicle and compared to KF.