Lower Limb Kinetic Prediction While Walking Based on Machine Learning Algorithms Using IMU Sensor
Lower Limb Kinetic Prediction While Walking Based on Machine Learning Algorithms Using IMU Sensor
Iman Bagheri1 Mahdi Bagheri2 Ali Ahmadi3 Amirreza Rouhbakhsh4 Amirhossein Amadeh5 Somia Molaei6 Ahad Alvandi7
1) Electrical and Computer Engineering, Email:
2) Computer Engineering, Khavaran Institute of Higher Education,Email:
3) Civil Engineering, Iran University of Science and Tech, Email:
4) Electronic Engineering, Northwestern Polytechnic University, Email:
5) Independent Researcher, Email:
6) Software Engineering, Iran University of Industries & Mines, Email:
7) Electrical and Computer Engineering, Kharazmi University, Email:
Publication :
7th International Conference on Applied Researches in Science & Engineering (7carse.com)
Abstract :
The use of artificial neural network (ANN) approaches on data from inertial measurement units (IMUs) for prediction has been reported in recent publications. These techniques could be used as quantitative markers of athletic performance or rehabilitation. The quantity and composition of IMUs. The selection of these parameter values is often made heuristically, and the justification for this is not discussed. We suggest employing an ANN to forecast the dynamic data of the lower limbs using a single measurement point based on the dynamic link between the center of mass and joint kinetics. From a single IMU worn close to the sacrum, data from seven subjects walking on a treadmill at various speeds were gathered. The data was divided into steps and given a numerical treatment for integration. From the kinematics of the measurement from a single IMU sensor, segment angles of the stance and swing leg and joint torques were estimated with fair accuracy. These findings highlight the significance of dynamic multi-segment kinetics during walking. A machine-learning approach based on the dynamic features of human walking can be used to resolve the tradeoff between data volume and wearable convenience.
Keywords :
Wearable Devices
IMU Sensors
Gait
Machine Learning
Walking