An Innovative Multi-Sensor Fusion Algorithm to Enhance Positioning Accuracy of an Instrumented Bicycle

Cycling is an increasingly popular mode of travel in cities, but its poor safety record currently acts as a hurdle to its wider adoption as a real alternative to the private car. A particular source of hazard appears to originate from the interaction of cyclists with motorized traffic at low speeds...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:IEEE transactions on intelligent transportation systems Jg. 21; H. 3; S. 1145 - 1153
Hauptverfasser: Miah, Shahjahan, Milonidis, Efstathios, Kaparias, Ioannis, Karcanias, Nicholas
Format: Journal Article
Sprache:Englisch
Veröffentlicht: New York IEEE 01.03.2020
The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
Schlagworte:
Accurate localisation
Algorithms
Angles (geometry)
bicycle geometry
bicycle kinematics
Bicycles
Bicycling
Collision avoidance
data acquisition
instrumented bicycle
Kalman filters
Kinematics
Localization
MEMS sensors
Multisensor fusion
Position measurement
Rolling motion
safety
sensor fusion
Sensors
Steering
survey data
Traffic speed
Trajectory
Turning
Urban areas
Urban environments
Wheels
Yaw
ISSN:1524-9050, 1558-0016
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Cycling is an increasingly popular mode of travel in cities, but its poor safety record currently acts as a hurdle to its wider adoption as a real alternative to the private car. A particular source of hazard appears to originate from the interaction of cyclists with motorized traffic at low speeds in urban areas. But while technological advances in recent years have resulted in numerous attempts at systems for preventing cyclist-vehicle collisions, these have generally encountered the challenge of accurate cyclist localization. This paper addresses this challenge by introducing an innovative bicycle localization algorithm, which is derived from the geometrical relationships and kinematics of bicycles. The algorithm relies on the measurement of a set of kinematic variables (such as yaw, roll, and steering angles) through low-cost on-board sensors. It then employs a set of Kalman filters to predict-correct the direction and position of the bicycle and fuse the measurements in order to improve positioning accuracy. The capabilities of the algorithm are then demonstrated through a real-world field experiment using an instrumented bicycle, called "iBike", in an urban environment. The results show that the proposed fusion achieves considerably lower positioning errors than that would be achieved based on dead-reckoning alone, which makes the algorithm a credible basis for the development of future collision warning and avoidance systems.
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1524-9050
1558-0016
DOI:10.1109/TITS.2019.2902797