This technology enables wearable devices to achieve precise gesture recognition and robotic arm control in complex environments such as intense physical exercise and underwater conditions, opening new doors for fields like virtual reality, rehabilitation medicine, and industrial rescue.
Targeting the Pain Point: Motion Interference is the "Enemy" of Gesture Recognition
Smartwatches, fitness trackers, and other wearable devices have long become an integral part of daily life. However, traditional inertial measurement units (IMUs) consistently face a core challenge—weak anti-interference capability.
When you run, the natural swinging of your forearms generates motion artifacts; when riding in a vehicle, environmental vibrations interfere with the signal; even simple changes in posture can cause device misalignment due to shifts in the gravitational vector.
These interference signals either have frequencies close to those of gesture signals or are of greater amplitude, easily "drowning out" genuine commands. Moreover, significant differences in individuals' movement habits further reduce recognition accuracy, rendering human-machine interaction in motion scenarios merely a "showpiece.".
In specialized scenarios such as industrial operations and rehabilitation training, this issue becomes even more critical. Imagine rescue personnel attempting to control a drone to survey hazards while running, only to suffer frequent misoperations due to signal interference; or rehabilitation patients trying to operate assistive devices with hand gestures, only to have their commands disrupted by body tremors—these limitations of traditional equipment have consistently constrained the practicality of human-machine interaction.
Dual Innovation: Hardware + AI to Create "Anti-Interference"
To tackle this challenge, the research team proposed a dual solution combining "hardware innovation and algorithm optimization," enabling wearable sensors to possess both "flexibility" and an "intelligent brain.".
Miniature stretchable sensor
Thin as paper, comfortable to wear, it integrates a six-channel IMU, electromyography (EMG) module, Bluetooth microcontroller unit, and stretchable battery, enabling wireless capture and transmission of gesture signals.
This sensor measures only 1.8×4.5 square centimeters in size, with a thickness of 2 millimeters and a stretchability exceeding 20%. When attached to the forearm, it does not impede movement. It features a four-layer, meticulously designed structure, with each layer serving a specific function: the first layer is the battery layer, which remains stable after 60 charge-discharge cycles, achieving a coulombic efficiency close to 100%; the second layer includes an antenna matching unit and an EMG acquisition module, utilizing a three-layer electrode structure to ensure stable signal collection; the third layer integrates a six-channel IMU and a Bluetooth microcontroller unit for signal processing and wireless transmission. In open environments, the Bluetooth signal remains stable within a 20-meter range, with a maximum temperature of just 27.7°C after continuous operation for 30 minutes. Worn for one hour, the skin temperature remains at 34.5°C, ensuring safety and comfort without any burden.