Learning Uncertainties in Wheel Odometry for Vehicular Localisation in GNSS Deprived Environments

Uche Abiola Onyekpe; Vasile Palade; Stratis Kanarachos; Stavros Christopoulos

doi:10.1109/ICMLA51294.2020.00121

Learning Uncertainties in Wheel Odometry for Vehicular Localisation in GNSS Deprived Environments

Uche Abiola Onyekpe, Vasile Palade, Stratis Kanarachos, Stavros Christopoulos

Research output: Chapter in Book/Report/Conference proceeding › Conference proceeding › peer-review

11 Citations (Scopus)

Abstract

Inertial Navigation Systems (INS) are commonly used to localise vehicles in the absence of Global Navigation Satellite Systems (GNSS) signals. However, they are plagued by noises, which grow exponentially over time during the triple integration computation, leading to a poor navigation solution. We explore the wheel encoder as an alternative to the accelerometer of the INS for positional tracking, and for the first time investigate the capability of deep learning using the Long Short-Term Memory (LSTM) neural network to learn the uncertainty inherent in the wheel speed measurements. These uncertainties could be manifested as changes in the tyre size or pressure, or wheel slips as a result of worn out tyres or wet/muddy road drive. The proposed solution has less integration steps in its computation, therefore providing the potential for a more accurate positioning estimation. Through a performance evaluation on several challenging scenarios for vehicular driving, such as hard braking, quick changes in vehicular acceleration, and wet/muddy road driving, we show that the wheel speed-based positioning approach is able to achieve up to 81.46 % improvement compared to the INS accelerometer approach.

Original language	English
Title of host publication	2020 19th IEEE International Conference on Machine Learning and Applications (ICMLA)
Editors	M. Arif Wani, Feng Luo, Xiaolin Li, Dejing Dou, Francesco Bonchi
Publisher	IEEE
Pages	741-746
Number of pages	6
ISBN (Electronic)	978-1-7281-8470-8
ISBN (Print)	978-1-7281-8471-5
DOIs	https://doi.org/10.1109/ICMLA51294.2020.00121
Publication status	Published - Feb 2021
Event	IEEE 19th International Conference on Machine Learning Applications - Miami, United States Duration: 14 Dec 2020 → 16 Dec 2020 Conference number: 19

Publication series

Name	Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020

Conference

Conference	IEEE 19th International Conference on Machine Learning Applications
Abbreviated title	ICMLA
Country/Territory	United States
City	Miami
Period	14/12/20 → 16/12/20

Keywords

Autonomous Vehicle Navigation
Deep learning
GPS outage
Inertial Navigation System (INS)
LSTM
Wheel Speed

ASJC Scopus subject areas

Artificial Intelligence
Computer Science Applications
Computer Vision and Pattern Recognition
Hardware and Architecture

Access to Document

10.1109/ICMLA51294.2020.00121Licence: Unspecified

Cite this

Onyekpe, U. A., Palade, V., Kanarachos, S., & Christopoulos, S. (2021). Learning Uncertainties in Wheel Odometry for Vehicular Localisation in GNSS Deprived Environments. In M. A. Wani, F. Luo, X. Li, D. Dou, & F. Bonchi (Eds.), 2020 19th IEEE International Conference on Machine Learning and Applications (ICMLA) (pp. 741-746). Article 9356239 (Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020). IEEE. https://doi.org/10.1109/ICMLA51294.2020.00121

Learning Uncertainties in Wheel Odometry for Vehicular Localisation in GNSS Deprived Environments. / Onyekpe, Uche Abiola; Palade, Vasile; Kanarachos, Stratis et al.
2020 19th IEEE International Conference on Machine Learning and Applications (ICMLA). ed. / M. Arif Wani; Feng Luo; Xiaolin Li; Dejing Dou; Francesco Bonchi. IEEE, 2021. p. 741-746 9356239 (Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020).

Research output: Chapter in Book/Report/Conference proceeding › Conference proceeding › peer-review

Onyekpe, UA, Palade, V, Kanarachos, S & Christopoulos, S 2021, Learning Uncertainties in Wheel Odometry for Vehicular Localisation in GNSS Deprived Environments. in MA Wani, F Luo, X Li, D Dou & F Bonchi (eds), 2020 19th IEEE International Conference on Machine Learning and Applications (ICMLA)., 9356239, Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020, IEEE, pp. 741-746, IEEE 19th International Conference on Machine Learning Applications , Miami, Florida, United States, 14/12/20. https://doi.org/10.1109/ICMLA51294.2020.00121

Onyekpe UA, Palade V, Kanarachos S, Christopoulos S. Learning Uncertainties in Wheel Odometry for Vehicular Localisation in GNSS Deprived Environments. In Wani MA, Luo F, Li X, Dou D, Bonchi F, editors, 2020 19th IEEE International Conference on Machine Learning and Applications (ICMLA). IEEE. 2021. p. 741-746. 9356239. (Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020). doi: 10.1109/ICMLA51294.2020.00121

Onyekpe, Uche Abiola ; Palade, Vasile ; Kanarachos, Stratis et al. / Learning Uncertainties in Wheel Odometry for Vehicular Localisation in GNSS Deprived Environments. 2020 19th IEEE International Conference on Machine Learning and Applications (ICMLA). editor / M. Arif Wani ; Feng Luo ; Xiaolin Li ; Dejing Dou ; Francesco Bonchi. IEEE, 2021. pp. 741-746 (Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020).

@inproceedings{f5f6551c4ed54bf9b8877d8c406ded4b,

title = "Learning Uncertainties in Wheel Odometry for Vehicular Localisation in GNSS Deprived Environments",

abstract = "Inertial Navigation Systems (INS) are commonly used to localise vehicles in the absence of Global Navigation Satellite Systems (GNSS) signals. However, they are plagued by noises, which grow exponentially over time during the triple integration computation, leading to a poor navigation solution. We explore the wheel encoder as an alternative to the accelerometer of the INS for positional tracking, and for the first time investigate the capability of deep learning using the Long Short-Term Memory (LSTM) neural network to learn the uncertainty inherent in the wheel speed measurements. These uncertainties could be manifested as changes in the tyre size or pressure, or wheel slips as a result of worn out tyres or wet/muddy road drive. The proposed solution has less integration steps in its computation, therefore providing the potential for a more accurate positioning estimation. Through a performance evaluation on several challenging scenarios for vehicular driving, such as hard braking, quick changes in vehicular acceleration, and wet/muddy road driving, we show that the wheel speed-based positioning approach is able to achieve up to 81.46 % improvement compared to the INS accelerometer approach.",

keywords = "Autonomous Vehicle Navigation, Deep learning, GPS outage, Inertial Navigation System (INS), LSTM, Wheel Speed",

author = "Onyekpe, {Uche Abiola} and Vasile Palade and Stratis Kanarachos and Stavros Christopoulos",

year = "2021",

month = feb,

doi = "10.1109/ICMLA51294.2020.00121",

language = "English",

isbn = "978-1-7281-8471-5",

series = "Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020",

publisher = "IEEE",

pages = "741--746",

editor = "Wani, {M. Arif} and Feng Luo and Xiaolin Li and Dejing Dou and Francesco Bonchi",

booktitle = "2020 19th IEEE International Conference on Machine Learning and Applications (ICMLA)",

address = "United States",

note = "IEEE 19th International Conference on Machine Learning Applications , ICMLA ; Conference date: 14-12-2020 Through 16-12-2020",

}

TY - GEN

T1 - Learning Uncertainties in Wheel Odometry for Vehicular Localisation in GNSS Deprived Environments

AU - Onyekpe, Uche Abiola

AU - Palade, Vasile

AU - Kanarachos, Stratis

AU - Christopoulos, Stavros

N1 - Conference code: 19

PY - 2021/2

Y1 - 2021/2

N2 - Inertial Navigation Systems (INS) are commonly used to localise vehicles in the absence of Global Navigation Satellite Systems (GNSS) signals. However, they are plagued by noises, which grow exponentially over time during the triple integration computation, leading to a poor navigation solution. We explore the wheel encoder as an alternative to the accelerometer of the INS for positional tracking, and for the first time investigate the capability of deep learning using the Long Short-Term Memory (LSTM) neural network to learn the uncertainty inherent in the wheel speed measurements. These uncertainties could be manifested as changes in the tyre size or pressure, or wheel slips as a result of worn out tyres or wet/muddy road drive. The proposed solution has less integration steps in its computation, therefore providing the potential for a more accurate positioning estimation. Through a performance evaluation on several challenging scenarios for vehicular driving, such as hard braking, quick changes in vehicular acceleration, and wet/muddy road driving, we show that the wheel speed-based positioning approach is able to achieve up to 81.46 % improvement compared to the INS accelerometer approach.

AB - Inertial Navigation Systems (INS) are commonly used to localise vehicles in the absence of Global Navigation Satellite Systems (GNSS) signals. However, they are plagued by noises, which grow exponentially over time during the triple integration computation, leading to a poor navigation solution. We explore the wheel encoder as an alternative to the accelerometer of the INS for positional tracking, and for the first time investigate the capability of deep learning using the Long Short-Term Memory (LSTM) neural network to learn the uncertainty inherent in the wheel speed measurements. These uncertainties could be manifested as changes in the tyre size or pressure, or wheel slips as a result of worn out tyres or wet/muddy road drive. The proposed solution has less integration steps in its computation, therefore providing the potential for a more accurate positioning estimation. Through a performance evaluation on several challenging scenarios for vehicular driving, such as hard braking, quick changes in vehicular acceleration, and wet/muddy road driving, we show that the wheel speed-based positioning approach is able to achieve up to 81.46 % improvement compared to the INS accelerometer approach.

KW - Autonomous Vehicle Navigation

KW - Deep learning

KW - GPS outage

KW - Inertial Navigation System (INS)

KW - LSTM

KW - Wheel Speed

UR - http://www.scopus.com/inward/record.url?scp=85102506142&partnerID=8YFLogxK

U2 - 10.1109/ICMLA51294.2020.00121

DO - 10.1109/ICMLA51294.2020.00121

M3 - Conference proceeding

SN - 978-1-7281-8471-5

T3 - Proceedings - 19th IEEE International Conference on Machine Learning and Applications, ICMLA 2020

SP - 741

EP - 746

BT - 2020 19th IEEE International Conference on Machine Learning and Applications (ICMLA)

A2 - Wani, M. Arif

A2 - Luo, Feng

A2 - Li, Xiaolin

A2 - Dou, Dejing

A2 - Bonchi, Francesco

PB - IEEE

T2 - IEEE 19th International Conference on Machine Learning Applications

Y2 - 14 December 2020 through 16 December 2020

ER -

Learning Uncertainties in Wheel Odometry for Vehicular Localisation in GNSS Deprived Environments

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this