Researchers in fields ranging from robotics to artificial intelligence, healthcare, and virtual reality are scrambling to close the gap between the physical and digital worlds. A crucial aspect of this process involves feeding digital systems with information about the real world and our interactions with it, which can then be leveraged to create more intelligent and responsive digital models of reality. All manner of sensing systems have been developed in recent years to collect the raw data needed to support these emerging applications.
Resistive tactile sensors are a very popular choice for a wide range of use cases due to their high performance, low power consumption, and simple design. However, making these sensors in complex shapes is a challenging undertaking. The fabrication process is frequently very labor-intensive, and it is common that expensive equipment is needed along the way. As such, access to this technology is limited. There are many applications that could benefit from resistive tactile sensors, but where the labor and other costs associated with producing them are too prohibitive.
Each glove contains 167 sensing pixels (đź“·: D. Murphy et al.)
A multidisciplinary team at MIT and the University of Washington has come up with a way to fabricate resistive tactile sensing gloves through a simple and inexpensive process. Their method uses flexible printed circuit board (FPCB) technology to automate the creation of personalized glove designs from just a photograph of a user’s hand on a standard 8.5-inch x 11-inch sheet of paper. From that image, an algorithm detects key hand landmarks using MediaPipe and defines fifteen sensing regions — three for each finger, two for the thumb, and one for the palm. Within each region, the algorithm generates horizontal and vertical copper traces that intersect to form “taxels,” or tactile sensing pixels. For each glove, the layout includes a total of 167 taxels.
Next, the algorithm creates a binary hand mask to identify the hand’s contour and generates routing paths to connect the electrodes to a readout circuit. These paths are carefully traced to avoid crossing the hand’s interior regions, and the thickness and positioning of these traces are adjusted to meet standard manufacturing tolerances.
The glove structure consists of five layers: two FPCBs (top and bottom) with orthogonal copper traces, a middle layer made of piezoresistive Velostat material, and two outer layers of laser-cut silicone rubber. These are precisely aligned using a 3D-printed model and assembled using pressure-sensitive adhesive. A custom-sewn fabric glove wraps around the sensor to provide comfort and wearability.
The materials used to make the gloves (đź“·: D. Murphy et al.)
The total cost is less than $130 per glove, with an assembly time of under 15 minutes. But the low price tag does not mean they are of low quality. The sensors were tested using a mechanical press and repeated force cycles, showing strong sensitivity and repeatability comparable to commercial tactile sensors.
This approach not only reduces fabrication costs and complexity but also opens up tactile sensing gloves to a broader community of researchers and developers. With further improvements, these gloves could be a major boon to applications in robotic manipulation, medical diagnostics, immersive virtual environments, and beyond.