Multimodal active perception for safe autonomous vehicles

Robotic active perception is a key factor to develop autonomous systems capable of safely and intelligently interacting with humans and their surrounding environment. Active perception allows robots to not only make accurate decisions, but also to autonomously perform actions that will lead the robot to improve its own performance during a specific task, such as autonomous driving, wearable assistive robots and human-robot collaboration.

Even though state-of-the-art robots are equipped with multimodal sensors (e.g., vision, touch, audio, force, gyroscopes), the capability of accurately and reliably extracting and fusing this multimodal data remains a challenge. The lack of this capability in robotic systems limits robots’ potential for autonomous learning, control, decision-making and actions, which also reduces the capability of robots to behave safely and intelligently.

This project, which is aligned with the UKRI Centre for Doctoral Training in Accountable, Responsible and Transparent Artificial Intelligence (ART-AI), involves the following key aspects required for the development of safe autonomous vehicles:

• Research and development of cognitive architecture for robot sensing, decision-making and control composed of high-level, middle-level, and low-level processes.

• Research and development of advanced machine learning methods capable of exploiting the benefits of multimodal sensor data for reliable and responsible decision-making processes in autonomous vehicles.

• Development of human-machine interfaces that allow the designer and user to verify the decisions made by the autonomous vehicle. This interface provides a component of transparency for validation and accountability of the actions performed by the robotic system.

The methods developed in this project will be implemented and tested in ROSbot mobile robots, and with the racing car developed by the Team Bath Racing Electric – AI (TBRe-AI).

The research to be undertaken in this project has a strong multidisciplinary nature. Therefore, the student is expected to collaborate with students and researchers from the ART-AI CDT, computer science, electronic and electrical engineering, mechanical engineering, and psychology. Furthermore, the student is expected to attend multiple events such as conferences, workshops and publish the results from the research work in international conferences and journals.

Candidates are expected to have or near to complete an MSc or MEng in Robotics, Computer Science, Electronics, Mechanics, Mathematics, Physics or related areas.

Informal enquiries about the project should be directed to Dr Uriel Martinez Hernandez: umh21@bath.ac.uk.

Enquiries about the application process should be sent to art-ai-applications@bath.ac.uk.

Formal applications should be made via the University of Bath’s online application form: https://samis.bath.ac.uk/urd/sits.urd/run/siw_ipp_lgn.login?process=siw_ipp_app&code1=RDUCM-FP02&code2=0003

Start date: 4 October 2021.