Autonomous vehicles in highly crowded pedestrian environments

With Connected and Automated Vehicles (CAVs) becoming increasingly prevalent on our roads, developing autonomous vehicles that interact well with pedestrians is of crucial importance for both pedestrian and passenger safety, as well as the efficient and seamless operations of the vehicle and network.

This project will address the development of algorithms to estimate the position and intention of pedestrians (including prediction and path of movement) essential to implement the safe interaction between CAVs and pedestrians.

It will also investigate the Human Machine Interface between CAVs and road users (how vehicles communicate to humans) and incorporate infrastructure information to improve the safety of such interaction.

There will also be a practical focus to develop solutions, observations or recommendations for areas related to the interaction of vehicles and pedestrians which could be implemented to improve road safety outcomes.

Participants

• Transport for NSW
• Australian Centre for Field Robotics at the University of Sydney
  

Project background

Transport for NSW (TfNSW) has set an ambitious vision for a technology- enabled future through Future Transport 2056. A stronger understanding of how automated vehicles will operate in pedestrian environments through the development of algorithms, sensors, and communications will greatly improve outcomes for our transport customers and society.

TfNSW and the University of Sydney are working together to establish the University of Sydney’s Camperdown campus and suitable surrounding areas as test beds for the university’s automated vehicles. This will enable the university to safely develop, test, and prove the technology in real-world environments.

This PhD project is linked to the Safely introducing CAVs into integrated transport networks project.

Project objectives

The project will develop the fundamental science to enable the implementation and understanding of the interaction between CAVs and vulnerable road users (VRU), in this case mainly pedestrians.

The research will develop intelligent algorithms to infer pedestrian intentions and positions. It will use multiple sources of information provided by different sensing modalities installed in the vehicle and intelligent infrastructure.

It also aims to develop effective Human Machine Interfaces to communicate vehicle intentions to VRU.

This project is expected to make scientific contributions to the state-of-the-art of CAV / Human Machine interactions. It will also provide TfNSW with valuable experimentation results to enable it to evaluate the potential implication and benefits of the introduction of this new technology, and the importance of intelligent infrastructure to improve safety outcomes for all road users.

Learnings from this work will also improve the safe rollout and use of automated or autonomous vehicles and associated technologies in NSW and Australia more broadly, including supporting the development of a set of safe operating conditions or principles for the use and deployment of automated or autonomous vehicles.

December 2022: Project completed

This project has been completed.