Practical Implementation of KalmanNet for Accurate Data Fusion in Integrated Navigation

The extended Kalman filter has been widely used in sensor fusion to achieve integrated navigation and localization. Efficiently integrating multiple sensors requires prior knowledge about their errors for setting the filter. The recently emerged KalmanNet managed to use recurrent neural networks to learn prior knowledge from data and carry out state estimation for problems under non-linear dynamics with partial information. In this letter, the KalmanNet is implemented for integrated navigation using data from GPS/Wheels and the Inertial Measurement Unit. Therein, a practical strategy for the training algorithm of truncated backpropagation through time is presented by taking advantage of the first-order Markov property of the system state of the Kalman filter, which improves the training robustness and performance of the existing KalmanNet. Experimental results on the Michigan NCLT dataset show that our fusion KalmanNet significantly outperforms the conventional EKF-based fusion algorithm with an improvement of 20% $\sim$ 40% in average RMSE.

I. Introduction

Navigation and localization systems are widely used in various fields, such as aerospace, smart agricultural, and vehicle navigation, and have attracted intensive research in recent years [1], [2], [3], [4], [5], [6]. For vehicle navigation, various measurement systems, e.g., Inertial Measurement Unit (IMU), Wheel Encoders (Wheels), and GPS are often utilized to achieve integrated navigation with enhanced performance. Each kind of sensor has its distinctive characteristics. GPS generally has an accurate performance but can degrade in circumstances with mask effects and multipath [7]; IMU and Wheels are a class of dead reckoning sensors that can provide relative position of a vehicle, but their positioning errors drift with time [8].