Scottish Sensor Systems Centre

University: Professor John Soraghan and Dr Lyourgis Petropoulakis, Centre for Excellence in Signal and Image Processing (CeSIP), University of Strathclyde; Dr Stephen Butler, School of Psychological Sciences and Health, University of Strathclyde; and Dr Simon Kyle, Glasgow Sleep Centre, University of Glasgow

Collaborative Partner: Brookes Bell Ltd.

Project Type: Prototype delivery (Grant £20K)

Project Duration: 12 months (start date March 2012)

Project Abstract

Employee fatigue has serious consequences for driver safety, aviation, and maritime control. Facial analytic based fatigue detection systems are generally based eye characteristics, including some recent developments in fatigue detection using eye-tracking. Many of these use disparate measures and, to date, there are no fatigue monitoring systems based on naturalistic fatigue indicators.

This project will develop new signal processing methods that aim to non-invasively measure the level of fatigue an individual is experiencing, using facial analytics based on static and dynamic images, as well as eye-movements of sleep-deprived individuals. A prototype continuous fatigue monitoring technology will be designed that will add an important human (operator) specific measurement dimension to the normal mechanistic inputs (such as doors open/closed, pressure levels, sea states etc.) in existing decision support systems used in commercial shipping.

Novel facial analytic methods have been developed through a collaborative research between CeSIP, the School of Psychological Sciences and Health, and the Glasgow Sleep Centre. The team is currently working on a pilot study to develop a fatigue database that can be utilised to further test the novel methods. This work will provide a base measurement of fatigue level from which the performance of new algorithms can be quantifiably evaluated. The monitoring is carried out through non-invasive facial analytics and is based on initial work that was carried out on clinical facial monitoring and spot lie detection.

Project Impact

Shipping plays a fundamental role in the global economy, with 90% of world trade carried by ship. Irrespective of how efficient crew performance is, infrequent mistakes can result in catastrophic outcomes: it is estimated that around 80% of serious accidents at sea can be attributed to the human factor.

Fatigue in ship operators is a concern, where shift patterns, ship movements, vibration, noise or lighting can disrupt quantity or quality of sleep. The project outcomes will improve safety in the sector: employed at work and control stations, the system will be designed with predictive capabilities based on serially progressive measures of fatigue, enabling the efficient co-ordination of work and rest schedules. For individuals, the system will provide an ‘early warning’ indication for operators of tiredness. A second warning system for more imminent fatigue threats based on eye closure, blinks and head movements will also be implemented as a further precaution.

The system will have clear applications in road, air and rail traffic. Furthermore, from the standpoint of accident analysis, it could be employed by accident investigators and insurance companies, which will ultimately improve subsequent safety procedures and recommendations.