FTEKFNet: Hybridizing Physical and Data-Driven Estimation Algorithms for Vehicle State
- 1 May 2024
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Intelligent Vehicles
- Vol. 9 (11) , 7301-7311
Abstract
Sideslip angle and vehicle velocity are crucial for both traditional and autonomous vehicles. They play essential roles in chassis stability control, as well as in tasks such as path planning and tracking control. However, these states cannot be directly measured by onboard sensors, therefore various vehicle state estimation algorithms have been developed. Most of these algorithms assume that the noise characteristics are known, ignoring the impact of missing measurement data, and cannot simultaneously handle the effects of colored noise and white noise. To address these issues, we propose a fault-tolerant extended Kalman filter network (FTEKFNet), which integrates both physics-based and data-driven methods for vehicle state estimation. Based on the Fault Tolerant Extended Kalman Filter (FTEKF) iterative framework, a pre-trained artificial neural network is utilized to directly predict the Kalman gain, and it is combined with FTEKF to form FTEKFNet. Experimental results under different conditions demonstrate that FTEKFNet can simultaneously deal with unknown noise and data loss problems and has good adaptability to color noise. The estimation performance of the proposed algorithm is better than the traditional FTEKF and EKF methods.Keywords
Funding Information
- Smart Traffic Fund (PSRI/47/2209/PR, PSRI/53/2210/PR)
- National Natural Science Foundation of China (52372379)
This publication has 40 references indexed in Scilit:
- Ensuring Secure Platooning of Constrained Intelligent and Connected Vehicles Against Byzantine Attacks: A Distributed MPC FrameworkEngineering, 2023
- RobustStateNet: Robust ego vehicle state estimation for Autonomous DrivingRobotics and Autonomous Systems, 2023
- Vehicle sideslip angle estimation using deep ensemble-based adaptive Kalman filterMechanical Systems and Signal Processing, 2020
- A robust observer based on filtering with parameter uncertainties combined with Neural Networks for estimation of vehicle roll angleMechanical Systems and Signal Processing, 2018
- Sideslip angle estimator based on ANFIS for vehicle handling and stabilityJournal of Mechanical Science and Technology, 2015
- A reduced-order nonlinear sliding mode observer for vehicle slip angle and tyre forcesVehicle System Dynamics, 2014
- A variable structure extended Kalman filter for vehicle sideslip angle estimation on a low friction roadVehicle System Dynamics, 2014
- Stochastically resilient extended Kalman filtering for discrete-time nonlinear systems with sensor failuresInternational Journal of Systems Science, 2014
- Design of a reduced‐order non‐linear observer for vehicle velocities estimationIET Control Theory & Applications, 2013
- On the vehicle sideslip angle estimation through neural networks: Numerical and experimental resultsMechanical Systems and Signal Processing, 2011