Fabio Ramos has been awarded a grant from the Mid-Career Researcher Development Scheme for the following project:
Learning and predicting human behaviour
Understanding and predicting human behaviour, in particular motion, is paramount for the development of better cities, work place environments and surveillance systems. From shopping habits in supermarkets to the dynamics of people’s motion in train stations, there are many patterns of behaviour that, if captured and modelled accurately, can help companies and governments to design more efficient liveable environments. This project aims to devise methods to extract and model human behaviour from 3D laser sensor data, in both indoor and outdoor environments. We will devise statistical methods to model sequences of human actions utilising principles of inverse reinforcement learning and nonparametric Bayesian regression in Hilbert spaces.