Localization Using Global Magnetic Positioning System for Automated Driving Bus and Intervals for Magnetic Markers

Ando, Takayuki; Mukumoto, Hirotaka; Aoki, Keiko M.; Okazaki, Shogo; Nagao, Tomohiko; Aoyama, H.; Yamamoto, Masahito; Nakano, Kimihiko

doi:10.1109/tiv.2022.3155324

Cited by 3 publications

(3 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Fig. 10b shows the vehicle's lateral and longitudinal acceleration, and it highlights how the driver is pushing the vehicle at the limit of handling in all the corners, see [1,7] s, [16,19] s and [23,29] s. The sideslip angle estimation performance of the four approaches is represented in Fig. 10a.…”

Section: A Full Dataset Resultsmentioning

confidence: 99%

See 1 more Smart Citation

An Unscented Kalman Filter-Informed Neural Network for Vehicle Sideslip Angle Estimation

Bertipaglia¹,

Alirezaei²,

Happee³

et al. 2023

Preprint

View full text Add to dashboard Cite

This paper proposes a novel vehicle sideslip angle estimator, which uses the physical knowledge from an Unscented Kalman Filter (UKF) based on a non-linear single-track vehicle model to enhance the estimation accuracy of a Convolutional Neural Network (CNN). The model-based and data-driven approaches interact mutually, and both use the standard inertial measurement unit and the tyre forces measured by load sensing technology. CNN benefits from the UKF the capacity to leverage the laws of physics. Concurrently, the UKF uses the CNN outputs as sideslip angle pseudo-measurement and adaptive process noise parameters. The back-propagation through time algorithm is applied end-to-end to the CNN and the UKF to employ the mutualistic property. Using a large-scale experimental dataset of 216 manoeuvres containing a great diversity of vehicle behaviours, we demonstrate a significant improvement in the accuracy of the proposed architecture over the current state-ofart hybrid approach combined with model-based and data-driven techniques. In the case that a limited dataset is provided for the training phase, the proposed hybrid approach still guarantees estimation robustness.

show abstract

Section: A Full Dataset Resultsmentioning

confidence: 99%

“…Kalman filter based on a kinematic model [12]- [14] Kalman filter based on IMU & GNSS measurements [15]- [22] EKF based on a dynamic model [23]- [29] UKF based on a dynamic model [30]- [35] Hybrid based on dynamic & kinematic models [3], [36],…”

Section: Model-basedmentioning

confidence: 99%

An Unscented Kalman Filter-Informed Neural Network for Vehicle Sideslip Angle Estimation

Bertipaglia¹,

Alirezaei²,

Happee³

et al. 2023

Preprint

View full text Add to dashboard Cite

show abstract

“…Kalman filter based on a kinematic model [17]- [19] Kalman filter based on IMU & GNSS [20]- [26] EKF based on a dynamic model [27]- [33] UKF based on a dynamic model [34]- [39] Sliding mode observer [40]- [42] H-infinity observer [43], [44] Luenberger observer [45] Hybrid -dynamic & kinematic models [3], [46]- [48] Online gradient descent [49] Modular estimation scheme [50], [51] Data-driven FFNN [52], [53] RNN [54]- [56] ANFIS [14] Kernel-based LPV [57] Hybrid FFNN, ANFIS & UKF [15], [58], [59] RNN (GRU) & UKF [16] Differentiable EKF [60] Kalman filter & FFNN [61] Piecewise Affine & Takagy-Sugeno [62] FFNN & Kalman in the back-propagation [63] KalmanNet [64] Kinematic model & RNN (GRU) [7] Deep Ensemble Network (LSTM) & EKF [11], UKF [12] in the proposed hybrid approach. Section IV describes how the experiments are conducted and evaluated.…”

Section: Model-basedmentioning

confidence: 99%

An Unscented Kalman Filter-Informed Neural Network for Vehicle Sideslip Angle Estimation

Bertipaglia,

Alirezaei,

Happee

et al. 2024

IEEE Trans. Veh. Technol.

View full text Add to dashboard Cite

This paper proposes a novel vehicle sideslip angle estimator, which uses the physical knowledge from an Unscented Kalman Filter (UKF) based on a non-linear single-track vehicle model to enhance the estimation accuracy of a Convolutional Neural Network (CNN). The model-based and data-driven approaches interact mutually, and both use the standard inertial measurement unit and the tyre forces measured by load sensing technology. CNN benefits from the UKF the capacity to leverage the laws of physics. Concurrently, the UKF uses the CNN outputs as sideslip angle pseudo-measurement and adaptive process noise parameters. The back-propagation through time algorithm is applied end-to-end to the CNN and the UKF to employ the mutualistic property. Using a large-scale experimental dataset of 216 manoeuvres containing a great diversity of vehicle behaviours, we demonstrate a significant improvement in the accuracy of the proposed architecture over the current state-of-art hybrid approach combined with model-based and data-driven techniques. In the case that a limited dataset is provided for the training phase, the proposed hybrid approach still guarantees estimation robustness.

show abstract

Mapping and Localization Method for Autonomous Vehicles on Roads Using Environmental Magnetic Field

Ishii,

Shimono,

Suda

et al. 2024

Lecture Notes in Mechanical Engineering

View full text Add to dashboard Cite

Vehicle localization is one of the key technical factors for autonomous vehicles. It requires high accuracy, precision, and robustness towards various road conditions. Popular localization methods include global navigation satellite system (GNSS) and visual methods, but their accuracy can degrade in some conditions. This work proposes to use the environmental magnetic field (EMF) for localization to complement the shortcomings of existing methods. EMF is a combination of the Earth’s geomagnetic field and magnetic field induced by man-made objects. It has local fluctuations that can be paired with coordinate positions and is time-invariant within a practical timescale. Past works considering the localization of road vehicles had few problems when applying them to the localization of autonomous vehicles. This work overcomes the problems in the existing method by creating a two-dimensional magnetic field map using Gaussian Process regression, using magnetic markers to enhance EMF fluctuations, and utilizing the Monte Carlo localization algorithm. The proposed method was validated through actual vehicle tests, and its robustness towards other vehicles was examined.

show abstract

Localization Using Global Magnetic Positioning System for Automated Driving Bus and Intervals for Magnetic Markers

Cited by 3 publications

References 25 publications

An Unscented Kalman Filter-Informed Neural Network for Vehicle Sideslip Angle Estimation

An Unscented Kalman Filter-Informed Neural Network for Vehicle Sideslip Angle Estimation

An Unscented Kalman Filter-Informed Neural Network for Vehicle Sideslip Angle Estimation

Mapping and Localization Method for Autonomous Vehicles on Roads Using Environmental Magnetic Field

Contact Info

Product

Resources

About