Open Muscle, an innovative project aimed at detecting finger movement and pressure using a forearm bracelet, has recently achieved a remarkable milestone. By using just 5 minutes of training data gathered from the Open Muscle 12 and the LASK4 device, the team was able to train an AI model that successfully detects finger movement and pressure. This exciting development showcases the potential of the Open Muscle suite to be used in prosthetics and finger movement detection, as well as the technology’s ability to produce near real-time results with low computational power requirements.
Open Muscle: A Brief Overview
Open Muscle leverages pressure sensors arranged radially around the forearm to detect changes in the topology of the forearm when muscles are contracted. These pressure sensors do not require a high sampling rate, reducing the need for computational power during both sensor readings and inputting the sensor values into the machine learning model for prediction.
The recent experiment proved that it is possible to achieve a prediction rate of 5-15 predictions per second, producing near real-time finger detection. While the accuracy of this 5-minute model was low, other tests have yielded much higher accuracy, providing a solid proof-of-concept for the Open Muscle suite as a whole. This type of device is often referred to as a somatosensory device.