Novelty detection in user behavioural models within ambient assisted living applications: An experimental evaluation

Current approaches to networked robot systems (or ecology of robots and sensors) in ambient assisted living applications (AAL) rely on pre-programmed models of the environment and do not evolve to address novel states of the environment. Envisaged as part of a robotic ecology in an AAL environment to provide different services based on the events and user activities, a Markov based approach to establishing a user behavioural model through the use of a cognitive memory module is presented in this paper. Upon detecting changes in the normal user behavioural pattern, the ecology tries to adapt its response to these changes in an intelligent manner. The approach is evaluated with physical robots and an experimental evaluation is presented in this paper. A major challenge associated with data storage in a sensor rich environment is the expanding memory requirements. In order to address this, a bio-inspired data retention strategy is also proposed. These contributions can enable a robotic ecology to adapt to evolving environmental states while efficiently managing the memory footprint.