Physical human-environment interaction via wearable sensors : motion tracking and localization

• Main Author: Yuan, Qilong
• Other Authors: Chen I-Ming
• Format: Theses and Dissertations
• Language: English
• Published: 2014
• Subjects: DRNTU::Engineering::Industrial engineering::Automation; DRNTU::Engineering::Electrical and electronic engineering::Industrial electronics
• Online Access: https://hdl.handle.net/10356/61853
• Institution: Nanyang Technological University

Human-environment interaction (HEI) is how humans affect and are affected by the surroundings. From the kinematic and kinetic points of view, the motion interaction between the surrounding and the human body can be called physical human-environment interaction (pHEI). Studying pHEI is very useful in understanding human-centered activities in daily live. The objective of this dissertation is to study the tracking of pHEI in daily applications via wearable inertial and contact sensor systems. In body motion tracking, multiple wearable inertial sensors are used to capture the human body kinematics. Meanwhile, contact sensors worn on the body are used to detect the contact interactions with the environment. To determine the precise human kinematic model of a human subject, a quick template-based calibration method is proposed. Subsequently, a three dimensional human motion tracking and localization method termed as simultaneous localization and capture (SLAC) based on the continuous contacts with the environment and the human body kinematics is introduced to record human motions and positions. For tracking of human motions with phases where there is no contact with the environment such as jumping and running, the velocity based SLAC (V-SLAC) and acceleration based SLAC (A-SLAC) are introduced. The proposed SLAC, V-SLAC and A-SLAC methods are able to simultaneously record body motion and track body location over a large space regardless of whether it is indoor and outdoor, or having or not having contacts with the environment. With the proposed quick template-based calibration method, precise human kinematic model can be achieved without using additional external devices. Experimental results and benchmark study with the optical-based Motion Analysis® system showed that the proposed SLAC can control the localization errors within 1% to 2% of the total distance travelled. The velocity errors of the V-SLAC can be controlled within 2% of the moving velocity for daily activities such as walking, jumping and jogging. Experiments on human subjects confirmed that full-body motion and the contact interaction during the pHEI can be properly captured in real-time. With the development of more advanced integrated MEMS inertial sensing technology, SLAC, V-SLAC and A-SLAC can offer great advantage in many daily applications requiring human motion tracking and localization.