A force-feedback upper-limb exoskeleton designed for post-stroke rehabilitation. Uses real-time biomechanical modeling and adaptive resistance to guide patients through recovery exercises.
Real-time force feedback exoskeleton for upper-limb motor rehabilitation
Custom 4-channel brushless motor driver board assembled. All channels tested at rated load. DRV8301 current sensing calibrated — reading ±2% of actual.
MPU6050 was showing 4°/min drift in yaw. Implemented complementary filter first, then upgraded to a full Kalman filter — now stable at <0.3°/min.
3rd iteration of forearm brace printed in PETG. Revised cable routing channels, added ergonomic padding cutouts. Fit-testing with 5 subjects showed improved comfort over v2.
6x FlexiForce sensors calibrated across full torque range. Linearity within 3% across 0–50 Nm. Multiplexer integration and I2C pipeline working at 200Hz.
5-segment upper limb model with inverse kinematics validated against published clinical datasets. Joint torque predictions within 8% of measured values.