Robotic Waste Collection in Train Carriages

Waste collection in train carriages (coffee cups, drink tins, newspapers, packages, etc.) is a tiring, tedious, and dull task currently performed by human train service personnel either in the nights or when the train is specifically put in idle for cleaning purpose. A mobile robotic manipulator can be designed and used to collect such waste in train carriages. The challenge in designing such robotic solution is that the space in the carriages is confined and the waste is usually in places difficult to reach (under the seats) with conventional robot arms. We have been working on this topic (with support of and collaboration with Railway Safety and Standards Board – RSSB) and up to now have developed a Deep Learned image processing algorithm to detect waste objects and an investigation of different conceptual designs for the needs of the task (see the references). This project aims at:
1) Developing the conceptual design of a collection mechanism from the confined places (under the seats),
2) Implementation of the collection mechanism with an available wheeled mobile robot,
3) Integration of the object detection and mobile collection mechanisms as a wholistic robotic system,
4) Testing and verification of the overall solution in the lab environment and in actual train carriages.

References:
[1] Western, N., Kong, X., and Erden, M. S., 2 Jun 2021, “Design of a Train Cleaning Robot for the Train Carriage Interior”, Twente, Netherlands, 19-21 May, In: Procedia CIRP: 100: 804-809.
[2] Western, N., Kong, X., and Erden, M. S., “Training Convolutional Neural Networks to Detect Waste in Train Carriages”, IEEE International Conference on Signal and Image Processing Applications (ICSIPA), Malaysia, 13-15 September, pp. 24-29.