Low-cost tags keep tabs on wearer's movements
Currently, if you're trying to digitally track someone's movements, a depth-sensing camera such as the Microsoft Kinect is one of the best ways to go. Researchers are developing a potentially better system, however, which involves attaching cheap sensors to the person's clothing.
Along with their expense, one of the problems with depth-sensing cameras is the fact that they can't track someone's movements when that person is out of the line of sight. One alternative is to attach inertial measurement units (accelerometer/gyroscope combos) to the person's body, although these can also be pricey, plus they require a power supply.
With that in mind, a Carnegie Mellon University team led by PhD student Haojian Jin is working with passive RFID (radio frequency identification) tags. Arrays of these battery-less paper-like devices are attached to a person's clothing on either side of joints in their body, such as the knee or elbow.
Upon being temporarily powered up by radio waves emitted from a handheld reader device, the RFID tags use a tiny integrated antenna to transmit a radio signal back to that device – that return signal is unique to each tag. By analyzing the tiny differences in the amount of time that it takes each of those signals to reach the reader's single antenna, it's possible to determine the bend/position of the joint to which they're all attached.
The system is known as RF-Wear, and according to Jin, it has been used "to demonstrate millimeter accuracy in skeletal tracking." Incorporated into washable items of clothing, it could potentially be utilized in applications such as controlling 3D video game avatars, creating computer-animated characters, or even for fitness tracking. A smartphone equipped with a special RFID-capable antenna could be used as the reader.
The researchers have also developed a system that detects changes in the shape of objects, known as WiSh (Wireless Shape-aware world). It likewise utilizes arrays of RFID tags, although in this case they're attached to non-wearable items that are squishable, stretchable or flexible.
Using a more sophisticated signal-processing algorithm than RF-Wear, it can determine when such an object has deviated from its previous shape, and in what way. Suggested uses include carpets that could detect the presence and location of people standing on them, stuffed toys that could react to squeezes and bends, or smart sleep-tracking pillows.
The technology has even already been used to measure the curvature of Pittsburgh's 10th Street Bridge, wherein a robot dragged a string of 50 RFID tags along the bridge's sidewalk.
Both RF-Wear and WiSh are demonstrated in the video below.
Source: Carnegie Mellon University