Especially in developed countries, unhealthy lifestyles are having a serious impact on individual quality-of-life, but also on society as a whole. Lifestyle diseases like smoking, overconsumption of alcohol, unhealthy diets and an inactive life are ever increasing. Unlike traditional diseases which are caused by an external trigger (like a virus or bacteria), these kinds of illnesses are caused by the unhealthy behavior of people. Hence, unlike traditional diseases, treating lifestyle diseases requires a complete new treatment paradigm, namely behavioral chance. Unfortunately human beings are creatures of habit and comfort, and hence inducing behavioral change is easier said than done.
In recent years, a number of efforts have been surfacing to try to address the issue. The widespread use of smart phones, make them a logical platform to implement behavioral change apps. The best of these show great promise, though usually only for people that are already motivated to change their lifestyle (cfr. exercise tracking apps). However these apps fail to reach the true problematic group of patients, those that lack the self-motivation. Similarly generalized advice (and particularly repetitive advice) does not work - we all know we should be active, eat healthy, not smoke or drink, and repeating the message is counterproductive. Furthermore, providing advice when one can't act upon it is equally unproductive.
This doctoral research will research a personalized virtual coach. Based on various contextual and physiological information, triggers for unhealthy behavior will be identified. This will require continuous and pervasive data collection. Through sensor fusion and machine learning, the virtual coach must learn the specific triggers for specific behaviors that would need to change. Once the correct triggers have been found and identified, the virtual coach must learn to make the most optimal suggestions at the most optimal time, again based on contextual awareness and through learning the behavioral patterns of the patient.
Electrical or biological engineer with a strong affinity for algorithm development and electronic systems.
Type of work:
55% algorithm development, 15% test system design, 30% field studies and validation.
Supervisor: Chris Van Hoof and Rudy Lauwereins
Daily advisor: Chris Van Hoof and Giuseppina Schiavone
When you apply for this PhD project, mention the following reference code in the imec application form: ref. SE 1704-09.