Analysing human behaviour and health by means of a large-area capacitive sensor floor

People are in almost constant contact with the floor throughout the day. With high-resolution sensors discreetly installed under the floor covering, data on the number, location, direction of movement, and speed of individuals can be recorded continuously. In addition to detecting emergencies such as falls, immobility, nightly wandering etc., the data sets can also be used to draw conclusions about behaviour, daily routines, activity levels, sleep quality, social isolation and even health status. The latter is particularly possible through the analysis of gait patterns, which can be used to detect physical and neurological diseases at an early stage.

We show examples from nursing homes that are fully equipped with sensor flooring, as well as from hospitals that use the sensor technology in geriatric care and casualty units for gait analysis and initial assessment. We present data from several studies in which the indications of the test subjects were known and assigned to gait patterns.

While even a layperson can recognize deviations from normal gait patterns in many cases, experts can detect the onset of pathological changes by analysing the gait parameters automatically extracted by the system. By tracking these parameters over a long period of time across multiple data recordings, the development of impairments and the success of rehabilitation measures can be quantitatively recorded. We also address the fact that the invisibility of the sensor technology under normal flooring allows for natural behaviour without people feeling like they are in a test situation.

In addition to a brief description of the sensor technology, we also discuss the AI methods that are used to extract those behaviour-relevant features from the data sets that are not visible to the naked eye. Previous scientific work that we have carried out in cooperation with the universities of Lübeck, Munich, and Ostrava suggests that it may even be possible to estimate gender and age from gait patterns and, with sufficiently large data sets, even to identify individuals from a defined group based on their gait.

We will discuss these current research topics at the end of our presentation. In addition, we hope that intensive discussions with the medical experts present will open up new questions that we could address in the future using the sensor system described.

Dr. Steinhage is co-founder, managing director, and CTO of Future Shape, a company that has been developing, manufacturing, and marketing large-area floor sensor technology for over 20 years. The sensor floor is used worldwide by nursing homes and hospitals for AI-supported behavior and emergency detection as well as for gait analysis. In 1998, Mr. Steinhage received a physics-Ph.D. from the Institute for Neuroinformatics at the University of Bochum, Germany, specializing in neuroinformatics and anthropomorphic robotics. He participated in more than 20 national and European research projects in the fields of sensor technology, AI, robotics, and ambient assisted living in a leading role.