Particle Filter for State Estimation in LHD Vehicles

Abstract

Abstract LHD (load-haul-dump) vehicles are used extensively in underground mining operations for transporting ore, primarily in tunnels where access is difficult or dangerous. To ensure LHD's performance underground is efficient and safe, a fault detection and isolation system (FDI) is required that will maintain control in the face of a fault in an element such as a sensor or actuator, or in some other vehicle component. A fundamental element in an FDI system based on analytic redundancy is the state estimator, which must take into account the vehicle's non-linear behavior. This paper compares two non-linear estimator alternatives for LHDs: the extended Kalman filter and the Bootstrap version of the Particle filter. Simulation results indicate that on most tests, the extended Kalman filter exhibits a better behavior.