Research

Tracking objects with an event-camera

Event-cameras can be used for high frequency tracking of objects without motion-blur problems.

A. Glover and C. Bartolozzi, “Robust visual tracking with a freely-moving event camera,” 2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2017, pp. 3769-3776, doi: 10.1109/IROS.2017.8206226. A. Glover and C. Bartolozzi, “Event-driven ball detection and gaze fixation in clutter,” 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2016, pp. 2203-2208, doi: 10.1109/IROS.2016.7759345.

High-frequency Human Pose Estimation

Lightweight neural networks can be trained with event data to perform complex tasks such as human pose estimation.

N. Carissimi, G. Goyal, F. D. Pietro, C. Bartolozzi and A. Glover, “[WIP] Unlocking Static Images for Training Event-driven Neural Networks,” 2022 8th International Conference on Event-Based Control, Communication, and Signal Processing (EBCCSP), 2022, pp. 1-4, doi: 10.1109/EBCCSP56922.2022.9845526.

Feature Points - Event-camera

Low-level processing of the event-stream (e.g. corner detection, and convolutions) can be achieved in real-time with asynchronous output.

A. Glover, A. Dinale, L. D. S. Rosa, S. Bamford and C. Bartolozzi, “luvHarris: A Practical Corner Detector for Event-Cameras,” in IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 44, no. 12, pp. 10087-10098, 1 Dec. 2022, doi: 10.1109/TPAMI.2021.3135635. V. Vasco, A. Glover and C. Bartolozzi, “Fast event-based Harris corner detection exploiting the advantages of event-driven cameras,” 2016 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2016, pp. 4144-4149, doi: 10.1109/IROS.2016.7759610. L. d. S. Rosa, A. Dinale, S. Bamford, C. Bartolozzi and A. Glover, “High-Throughput Asynchronous Convolutions for High-Resolution Event-Cameras,” 2022 8th International Conference on Event-Based Control, Communication, and Signal Processing (EBCCSP), 2022, pp. 1-8, doi: 10.1109/EBCCSP56922.2022.9845500.

Appearance-based SLAM

With the right algorithm autonomous vehicles can recognise where they are even in extreme weather conditions.

A. J. Glover, W. P. Maddern, M. J. Milford and G. F. Wyeth, “FAB-MAP + RatSLAM: Appearance-based SLAM for multiple times of day,” 2010 IEEE International Conference on Robotics and Automation, 2010, pp. 3507-3512, doi: 10.1109/ROBOT.2010.5509547. A. Glover, W. Maddern, M. Warren, S. Reid, M. Milford and G. Wyeth, “OpenFABMAP: An open source toolbox for appearance-based loop closure detection,” 2012 IEEE International Conference on Robotics and Automation, 2012, pp. 4730-4735, doi: 10.1109/ICRA.2012.6224843. M. Milford et al., “Condition-invariant, top-down visual place recognition,” 2014 IEEE International Conference on Robotics and Automation (ICRA), 2014, pp. 5571-5577, doi: 10.1109/ICRA.2014.6907678.

Simple Affordance Learning

Before Deep Neural Networks, learning could be achieved by on-line generation of models, such as Markov Decision Processes, allowing a robot to build their understanding of simple worlds and language.

A. J. Glover and G. F. Wyeth, “Toward Lifelong Affordance Learning Using a Distributed Markov Model,” in IEEE Transactions on Cognitive and Developmental Systems, vol. 10, no. 1, pp. 44-55, March 2018, doi: 10.1109/TCDS.2016.2612721. R. Schulz, A. Glover, M. J. Milford, G. Wyeth and J. Wiles, “Lingodroids: Studies in spatial cognition and language,” 2011 IEEE International Conference on Robotics and Automation, 2011, pp. 178-183, doi: 10.1109/ICRA.2011.5980476.