Extended Particle-Aided Unscented Kalman Filter Based on Self-Driving Car Localization

Abstract

The location of the vehicle is a basic parameter for self-driving cars. The key problem of localization is the noise of the sensors. In previous research, we proposed a particle-aided unscented Kalman filter (PAUKF) to handle the localization problem in non-Gaussian noise environments. However, the previous basic PAUKF only considers the infrastructures in two dimensions (2D). This previous PAUKF 2D limitation rendered it inoperable in the real world, which is full of three-dimensional (3D) features. In this paper, we have extended the previous basic PAUKF’s particle weighting process based on the multivariable normal distribution for handling 3D features. The extended PAUKF also raises the feasibility of fusing multisource perception data into the PAUKF framework. The simulation results show that the extended PAUKF has better real-world applicability than the previous basic PAUKF.

Discriminative Parameter Training of the Extended Particle-Aided Unscented Kalman Filter for Vehicle Localization

“…However, this method adds to the vehicle computing unit's online computational burden. In previous research, we proposed the extended particle-aided unscented Kalman filter (PAUKF) for localizing a self-driving car based on a pre-defined map [15,16]. The basic PAUKF and extended PAUKF improved the performance by combining the advantages of the particle filter (PF) with those of the unscented Kalman filter (UKF).…”