An anomalous situation detection system for cognitive impaired people

Cognitive diseases such as Alzheimer's, Parkinson's, Autism, etc. affect millions of people around the world and reduce the quality of life for the patient and their relatives. An impaired patient may show irrationally behaviors which could lead him/her to perform abnormal and/or dangerous actions for his/her safety. This paper presents an approach for modeling, detecting and predicting anomalous and dangerous situations. The proposed method is based on two technologies: it adopts the Situation-awareness paradigm for the detection of anomalous situations in a living environment and the LZ78 loss-less algorithm for the definition of a probabilistic model to predict the next action. Test performed in the laboratory and theoretical results show the validity of the approach.In our future work we will develop a smart kitchen able to detect risks for the patient.