Multimodal Interaction in an Adaptive Dementia Exercise Robot

As a result of demographic change, the number of seniors in nursing homes will increase dramatically in the coming years. As the care of dementia patients is very time-consuming and staff is being cut wherever possible due to cost pressure in nursing homes, the nursing sector will face enormous challenges in the next years. However, the regular training of certain physical exercises adapted to the individual state of health is very important for the health promotion and resilience of dementia patients. Therefore, it is necessary to investigate a dementia robot that enables seniors to train these exercises independently and flexibly. As dementia exercises that can be transferred to a robot, ball throwing, high-five game and strength exercise have already been derived in preliminary work. Based on this, an adaptive, optimizing and real-time interaction system was now researched, which classifies the degree of dementia rule-based with fuzzy logic and adapts the exercise parameters to the health state by evolutionary algorithms. In the same way, the interaction system optimizes the motivation of the patient to perform an exercise. To develop the classifier, the expert knowledge of caregivers was collected using knowledge acquisition. The expert knowledge was formalized to a knowledge base with a hybrid inference mechanism. The results show that the degree of dementia can be correctly classified for the exercises ball throwing, high-five game and strength exercise and the exercise parameters can be optimized with respect to an individual therapy progress. The interaction system was first tested in a real-time robot simulation. After that, tests with a real two-armed robot were successfully carried out. Currently, the evaluation with trained personnel and the acceptance study in a nursing home are being prepared.