Blog: MIT’s Marko Is Like Small-Scale Radar That Can Track Individuals in Their Homes
Radar, or at least crude experimental versions of it, has existed since the beginning of the 20th century. Then, in the years leading up to World War II, modern radar systems were independently developed by multiple countries around the world. They, along with today’s systems, all work in the same way: radio or microwaves are pulsed by a transmitter, and the reflections of those pulses are picked up by a receiver. Those are then processed to identify objects that were reflecting the waves. Now, MIT researchers have developed a similar technology that track individuals as they move through their homes.
Marko works in much the same way as traditional radar, but at a smaller, higher-resolution scale using low-power radio-frequency wireless signals. As with conventional radar systems, those RF signals are sent from a transmitter and the reflections are picked up by a receiver and processed. In this case, however, the RF signals have a higher frequency, which means they have a higher resolution. Modern analysis techniques, specifically a convolutional neural network machine learning model, also make it possible for Marko to process those signals to do useful work.
When setup in a retirement home, for example, Marko can detect and track multiple individuals as they move about the building. It can also identify those individuals based on their physical attributes and activity. That could provide significant help in the care of people living in the home. In one test case, they used Marko to analyze a patient with dementia that would sometimes become agitated. They found that the patient would pace around when agitated, and through Marko’s tracking they were able to determine that the patient became agitated after family visits. That’s just one use case, but it demonstrates just how powerful Marko is.