Real-Time Uncertainty Estimation of Autonomous Guided Vehicle Trajectory Taking Into Account Correlated and Uncorrelated Effects

This paper presents the description of a novel uncertainty estimation method employed for the navigation of autonomous guided vehicles. In the proposed algorithm, the uncertainty of the odometric navigation system is estimated as a function of the actual maneuver being carried out, which is identified by navigation data themselves. The result is a recursive method for estimating the evolution of spatial uncertainty, which takes into account unknown systematic effects and uncorrelated effects due to kinematic model uncertainty. The method is explained starting from the measurement models and its parameters as a function of the actual maneuvers. A verification of covariance propagation estimate due to systematic effects was carried out by means of a Monte Carlo simulation method. Experimental verification was carried out using an autonomous vehicle. Compatibility between a reference environment-referred system and the uncertainty estimated by the proposed method was achieved in 95% of the trials