Optimization of Sensor Data Processing Methods for Gait Tracking
No Thumbnail Available
URL
Journal Title
Journal ISSN
Volume Title
Sähkötekniikan korkeakoulu |
Master's thesis
Authors
Date
2023-08-21
Department
Major/Subject
Electronic and Digital Systems
Mcode
ELEC3060
Degree programme
AEE - Master’s Programme in Automation and Electrical Engineering (TS2013)
Language
en
Pages
57+1
Series
Abstract
Lower-limb prosthetic devices are designed to restore the mobility and functionality of amputees who have lost their lower limbs due to injury, disease, or congenital defects. Active prosthetic devices use actuators and sensors to adjust the device’s operation according to the user’s intention and environment. Inertial measurement units (IMUs) are commonly used sensors in prosthetic devices, as they can measure the motion of the device using gyroscopes and accelerometers. By integrating these measurements, the attitude and position of the device can be estimated. However, IMU-based methods suffer from drifts and noise due to integration and sensor errors, which affect the accuracy and reliability of the motion estimation. Therefore, accurate and robust estimation algorithms are needed to improve the functionality and performance of prosthetic devices. This thesis presents an optimized algorithm for estimating the attitude, velocity, and position of a lower limb prosthetic device based on an embedded IMU. The algorithm uses a Kalman filter to fuse the data from the gyroscope and accelerometer. Furthermore, it exploits the characteristics of the human gait cycle and applies a gait-based optimization method to improve the sensor data processing and reduce errors. The gait-based optimization method is based on the idea that during certain periods of the gait cycle, such as the stance phase, some specific states can be identified and used to correct the estimation errors. To evaluate the algorithm, the results were validated by an experiment using a vision-based motion capture system as a reference. The experiment involved a user wearing an active prosthetic knee with four markers on the device to track his motion using the motion capture system. The experiment included two test conditions: short-term ground walking with back-and-forth motions and long-term consecutive treadmill walking. The results showed that the algorithm achieved accurate and robust estimates of the attitude and position of the prosthetic device using an embedded IMU sensor with the aid of built-in force sensors. Meanwhile, it effectively reduced drifts during long-distance measurements. This enables the measurement of gait parameters, such as shank angles, walking speed, and stride length, using the IMU during movement and provides further valuable insights into an individual’s gait dynamics.Description
Supervisor
Särkkä, SimoThesis advisor
Sigurþórsson, StefánKeywords
lower-limb prosthesis, inertial measul, gait yccle, kalman filter, Attitude estimation, velocity estimation