Kinect Camera Based Gait Data Recording and Analysis for Assistive Robotics-An Alternative to Goniometer Based Measurement Technique

• Published in: Procedia computer science Vol. 133; pp. 763 - 771
• Main Authors: Roy, Ganesh, Bhuiya, Aritra, Mukherjee, Aditya, Bhaumik, Subhasis
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 2018
• Subjects: Gait; Goniometer; Kinect Camera; lower limb exoskeleton; Skeleton tracking; Gait; Kinect Camera; lower limb exoskeleton; Goniometer; Skeleton tracking
• ISSN: 1877-0509; 1877-0509
• DOI: 10.1016/j.procs.2018.07.121

With the increasing advancement of Gait analysis with marker-less skeletal tracking system, the Microsoft Kinect camera is becoming increasingly popular. This paper proposes a technique to record the joint coordinates of a lower extremity of a human being using a Microsoft Kinect camera which is quite reliable with respect to a wearable sensor like Goniometer. Using the raw data from Kinect camera a motion file is generated for simulating the musculoskeletal model of the OpenSim software. Further the moving average filter is applied to smooth the dataset to obtain a better motion. It provides a comparative study between the Kinect’s extracted gait data and a twin axis Goniometer data. Finally, the accuracy of the Kinect camera at different speed of range is estimated by making a study over a group of ten human subjects with normal gait pattern in a treadmill setup. The results obtained are verified by taking into account the angle data of five prominent lower limb sections during joint movement. Based on the experimental results the accuracy and correlation of various speed of motion for the Kinect v1 sensor is evaluated. One of the important application of this system in assistive robotics is to generate the trajectory of a lower limb exoskeleton device.