Study on nerve-cell based real-time estimation method for GPS multi-path error

Multi-path effect is the main error source for GPS high precision positioning in near distance differential circumstance. This paper presents a novel method for vehicle navigation in the situation. The current state vector of a vehicle can be used to estimate the state vector in the next interval. Then residual error vector can be computed by using the estimated state vector and the state vector measured by GPS. Finally, two nerve cells are introduced to guarantee real-time estimation reliability: one nerve cell can real-timely distinguish whether the GPS signal includes multi-path error using residual error vector; the other real-timely diagnoses the current precision of vehicle state vector. Vehicle Experiments in the field show that nerve-cell based real-time estimation method can effectively eliminate the effect of GPS multi-path error on positioning precision.