Human Arm Motion Tracking by Orientation-Based Fusion of Inertial Sensors and Kinect Using Unscented Kalman Filter.

Area of Science:

• Biomechanics
• Robotics
• Sensor Fusion

Background:

• Low-cost human motion capture is crucial for applications like home-based rehabilitation.
• Existing methods using MEMS inertial sensors or Kinect have limitations in non-laboratory environments.
• MEMS inertial sensors suffer from drift, while Kinect can experience joint occlusion.

Purpose of the Study:

• To develop a low-cost, accurate human arm motion tracking system for non-laboratory environments.
• To fuse data from MEMS inertial sensors and Kinect to overcome individual device limitations.
• To improve the robustness of motion capture during dynamic movements and potential occlusions.

Main Methods:

• Developed an unscented Kalman filter (UKF) approach to fuse orientation data from MEMS inertial sensors and Kinect.
• Created a new measurement model within the UKF algorithm.
• The fusion algorithm compensates for inertial sensor drift and Kinect joint occlusion.

Main Results:

• The proposed fusion algorithm significantly improved human arm motion tracking accuracy.
• Errors were reduced by approximately 50% compared to using either MEMS inertial sensors or Kinect alone.
• The system demonstrated effectiveness during both stationary joint positions and dynamic body movements.

Conclusions:

• The UKF-based fusion of MEMS inertial sensors and Kinect offers a robust and accurate solution for low-cost human motion capture.
• This method enhances reliability in real-world, non-laboratory settings, particularly for rehabilitation.
• The developed approach effectively mitigates common issues like sensor drift and occlusion.