A Smart Carpet: Technology for Persons with Alzheimer’s Disease

Prinicpal Investigator:

Co-Investigators: ,


Project Summary

Unobtrusive technological monitoring of seniors at risk for falls, wandering and dangerous behavior is a difficult problem. Here at the University of Missouri, we are monitoring many residents of an aging care facility focused towards their increased independence. A particularly difficult problem is that of sensing the proximity or location of the resident, detecting possible falls and doing so as unobtrusively as allowed by the resident. While there are many types of sensors to do so, all carry substantial drawbacks, for example video cameras require recognition software, but typically are considered very obtrusive by the residents.

We propose a carpet with pressure sensors distributed throughout the floor with an average of 10 sensors per square foot sheet; these will be incorporated onto flooring material such as carpeting, flexible tiles or linoleum. For appropriate carpet sizes we will provide the necessary support logic to provide signal information several times per second. We will be able to “see” the person’s footsteps, assess their gait, and identify their location. We will detect any falls, detect that the individual has wondered off, and follow them to high risk areas such as their kitchen. This smart carpet will help us identify changes in gait which may be indicative of worsening conditions and tell us that they have left the kitchen with the stove on at 3:00 in the morning. We also expect this to be a very low cost solution.

Our goal is to produce a prototype carpet or flooring material that can be tested in the third year. We expect to construct an intelligent carpet in the residence of 2 frail but normal elderly and 2 Alzheimer’s patients in the third year.

We have the following schedule. While we list the senior investigators, the work will be carried out in part by graduate students and technical staff.

  • Year 1: We will produce initially sheets with one sensor. We will test arrays these with prototypical logic and microprocessor systems. And at the end of the year we will produce sheets with 24 sensors. Dr. Tyrer will work to simulate the data acquisition and organization system to test systems with sensors. Dr. Aud and Dr. Tyrer will work to design the flooring material or carpet.
  • Year 2: We will work to improve the design of the sensor system. Dr. Tyrer will develop the data acquisition and organization system. With Dr. Aud we will develop the final design of the flooring and develop the testing protocols with older adults with Alzheimer’s disease and related dementia.
  • Year 3: In TigerPlace with whom we have a research relationship we will initially test the carpet with frail but elders with no appearance of dementia. At The Bluffs (another long-term care facility with a research relationship with MU) we will find the Alzheimer’s patients and we will test their use of the carpet. Dr. Aud, who has conducted two research projects at The Bluffs and who is a co investigator in the current research projects at TigerPlace, will be essential for these studies.

The testing will really occur throughout the lifetime of the grant. We will determine body weight discrimination, detection of falls, detection of gait, real time response (response delay), and ability to use wireless transmission. More technically we will need to determine the yield of the printing process. Can we do this on cloth or more suitable materials? How robust is the sensor system? How many sensors can fit on each sheet? For the resident and patient information not only do we need to determine the technical features but other issues comer into play. How can we make the design safe? That is design the carpet so that it will not cause the walker to misstep. Will the sensors tolerate repeated walking?

This development will solve a very important problem of monitoring elderly in their homes and rapidly detect falls. The prototype proposed here can be tested for further development. While we have looked at the direct data coming from the carpet system, modeling and further evaluating the data and the use of intelligent techniques will give further information such as changes in gait that presage changes in patient’s condition. Using computational intelligence techniques we can provide distinction between caregivers and patients in the carpeted room. We can follow the daily activities of the patient, and we can determine the location of the patient at any time. Finally, the electronic nature will substantially help care providers since the data from the carpet system can be made available over the web (with appropriate security measures), and ensure that the active care givers can be enlarged by including family at a distance.