LimbMotion: Decimeter-level Limb Tracking for Wearable-based Human-Computer Interaction
summary
A real-time system fuses smartwatch IMU data with fast acoustic ranging to recover full 3D limb posture, implemented on a Ticwatch2 with a Huawei Honor 7 anchor and validated against Kinect v2 ground truth with eight volunteers. Key components are a two-step filter (complementary + linear Kalman) and a point-cloud posture lookup aided by fast acoustic ranging. Automatic arm-length initialization uses swinging motions. System achieves median error 7.5–8.9 cm, reducing prior error by about 32% and runs in real time.
AI-written summary · about