Grant: Alliance Grants
University of Windsor
$25,000.00
Uncertainty-aware Pedestrian Behavior Prediction for Autonomous Vehicles
EF 2023-24
Résumé
Predicting pedestrian behavior is crucial in Advanced Driver Assistance Systems (ADAS) due to their vulnerability on the roads. By anticipating the pedestrians' behavior, preventive measures can be implemented to create safer environments for them. However, pedestrian behavior prediction is a challenging task due to dynamic and interactive urban environments. The main objective of this research proposal is to estimate and deploy uncertainties for a more accurate pedestrian behavior prediction. Deep Neural Networks (DNNs) play a crucial role in autonomous vehicles by enabling advanced perception, decision-making, and control systems. Generally, DNN uncertainty can be decomposed into aleatoric and epistemic uncertainties. While aleatoric uncertainty arises from the defect in the collected data, epistemic uncertainty reflects the DNN's lack of knowledge about the data it has previously not observed. In this study we aim to estimate uncertainties in DNNs by utilizing Bayesian neural networks, Monte Carlo Dropout, and variational inference approaches. This international research grant will have several benefits. First, it contributes to enhanced safety and improved decision-making. By accurately anticipating how pedestrians are likely to move, drivers can proactively adjust their trajectories and actions to avoid potential collisions. Moreover, it improvs the overall flow of traffic and enhances comfort and experience for drivers. Finally, through this international research collaboration, the University of Windsor in Canada, the Norwegian University of Science and Technology, and Leeds University in the UK can share their infrastructure, experiences, data, and methods in the field of automated vehicles. This would create a unique opportunity to train and exchange HQPs in the emerging fields of artificial intelligence and machine learning, human-machine interaction, automotive safety regulations, etc.
Programme
Alliance Grants
Bénéficiaire
Emplacement
Ontario, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
NSERC's Awards Data
Tel que publié
Application id
590682-2023
Name nom
Alirezaee, Shahpour DR
Department département
Mechanical, Automotive and Materials Engineering
Organization id
37
Institution établissement
University of Windsor
Province en
Ontario
Province fr
Ontario
Country en
CANADA
Country fr
CANADA
Fiscal year exercice financier
2023
Competition year année de concours
2023
Award amount
25000
Program id
ALLRP
Program name en
Alliance Grants
Program name fr
Subventions Alliance
Group en
RESEARCH PARTNERSHIPS
Group fr
PARTENARIATS DE RECHERCHE
Committee code
96
Committee name en
RPP Internal Decision Cttee
Committee name fr
Cté de décision interne - PPR
Area of application code
1206
Area of application group en
Advancement of knowledge
Area of application group fr
Avancement des connaissances
Area of application en
Engineering
Area of application fr
Génie
Research subject code
2800
Research subject group en
Artificial intelligence
Research subject group fr
Artificial intelligence (computer vision, use2603)
Research subject en
Intelligence artificielle
Research subject fr
Intelligence artificielle (vision artificielle, utiliser 2603)
Application title
Uncertainty-aware Pedestrian Behavior Prediction for Autonomous Vehicles
Keyword
Automated vehicles; Behavior prediction; Deep learning; Long short-term memory; Pedestrian behavior model; Pedestrian intention; Road safety application; Trajectory prediction; Vulnerable road user
Application summary
Predicting pedestrian behavior is crucial in Advanced Driver Assistance Systems (ADAS) due to their vulnerability on the roads. By anticipating the pedestrians' behavior, preventive measures can be implemented to create safer environments for them. However, pedestrian behavior prediction is a challenging task due to dynamic and interactive urban environments. The main objective of this research proposal is to estimate and deploy uncertainties for a more accurate pedestrian behavior prediction. Deep Neural Networks (DNNs) play a crucial role in autonomous vehicles by enabling advanced perception, decision-making, and control systems. Generally, DNN uncertainty can be decomposed into aleatoric and epistemic uncertainties. While aleatoric uncertainty arises from the defect in the collected data, epistemic uncertainty reflects the DNN's lack of knowledge about the data it has previously not observed. In this study we aim to estimate uncertainties in DNNs by utilizing Bayesian neural networks, Monte Carlo Dropout, and variational inference approaches. This international research grant will have several benefits. First, it contributes to enhanced safety and improved decision-making. By accurately anticipating how pedestrians are likely to move, drivers can proactively adjust their trajectories and actions to avoid potential collisions. Moreover, it improvs the overall flow of traffic and enhances comfort and experience for drivers. Finally, through this international research collaboration, the University of Windsor in Canada, the Norwegian University of Science and Technology, and Leeds University in the UK can share their infrastructure, experiences, data, and methods in the field of automated vehicles. This would create a unique opportunity to train and exchange HQPs in the emerging fields of artificial intelligence and machine learning, human-machine interaction, automotive safety regulations, etc.