Intelligent Integration of a MEMS IMU with GPS using a Reliable Weighting Scheme

摘要

The demand for civil navigation systems in harsh environments has been growing over the last several years. The Global Positioning System (GPS) has been the backbone of most current navigation systems, but its usefulness in downtown urban environments or heavily treed terrain is limited due to signal blockages. To help bridge these signal gaps inertial navigation systems (ENS) have been suggested. An integrated INS/GPS system can provide a continuous navigation solution regardless of the environment. For civil applications the use of MEMS sensors are needed due to cost, size and regulatory restrictions of higher grade inertial units. The Kalman Filter has traditionally been used to optimally weight the GPS and ENS measurements, but when using MEMS grade sensors the tuned parameters are not always the optimal ones. En these cases, the position errors during loss of the GPS signals accumulate faster than the ideally tuned case. To help compensate for imperfect tuning, neural networks were used to learn the residual errors and compensate for them during GPS signal outages. These neural compensations are capable of improving the Kalman Filter predictions when certain conditions are met, such as convergence of the training data on similar inputs to those used for prediction. To control the neural predictions, in cases where the learning has not yet converged or if the noise level of the neural predictions is larger than the Kalman Filter errors, an adaptive fuzzy inference system was used to weight the neural and Kalman Filter predictions. This fuzzy system ensured that the predictions were at worst the same as the Kalman Filter predictions, with improvements surpassing those of an ideally tuned filter.