Loosely coupled GPS/INS integration with Kalman filtering for land vehicle applications

Nowadays, global positioning system (GPS) has been used widely in land vehicles to provide positioning services. However, information from this stand-alone device may be interrupted under some circumstances such as: urban environment, under the tunnels, etc. In addition, low-rate sample time typically 1Hz is another drawback of GPS. Therefore, GPS receiver can be augmented with the inertial navigation system (INS) to provide faster positioning information. By fusing GPS and INS data, the errors are bounded and accuracy increases considerably even when using low-cost INS and GPS. This paper presents a method of INS/GPS integration where a loosely coupled model is formulated and an extended Kalman filter is then applied to estimate information about position, velocity, and acceleration. Experiment results show that the processing time is reduced significantly but the estimated errors are acceptable (less than 1 meters) when applying the proposed integration method under the good signal of GPS. The performance evaluation is also implemented on several road trajectories in urban city that the 3 meters precision could be reached. In addition, accurate positioning and navigation results are still available from 9 to 14 seconds of GPS outages with the position errors spread from 3 to 10 meters (RMS).