Analysis of A Mahony Algorithm-Based System for Monitoring Swimming 3D Motion Data

• Published in: 2023 IEEE International Conference on Integrated Circuits and Communication Systems (ICICACS) pp. 1 - 5
• Main Author: Wang, Dongdong
• Format: Conference Proceeding
• Language: English
• Published: IEEE 24.02.2023
• Subjects: Filtering; Filtering algorithms; initial value calibration; Interference; Magnetic field measurement; Mahony complementary filtering; Micromechanical devices; motion sensors; swimming 3D movements; Three-dimensional displays; Trajectory
• DOI: 10.1109/ICICACS57338.2023.10100053

The monitoring of the movement state of an object has been generally done from the outside, i.e., the monitoring point is separated from the moving object. While this approach has the advantage of improving accuracy, swimming is an endurance competition that tests the overall strength of the athlete. The use of traditional external monitoring would therefore be costly. Furthermore, in this sport, as in many others, it is particularly important to track how swimming's 3D movements change over time. In addition, unlike other land sports, swimming needs to be monitored taking into account the effects of water on the monitoring equipment, particularly the absorption and reflection of signals by water, which can cause the monitoring data to lose its authenticity. Swimming 3D movement is an inertial stance and the measurement of inertial stance needs to ensure the accuracy of the initial data in order to calculate swimming 3D movement data. However, the swimmer's jump is of uncertain timing and the calibration of the initial data by the transmission technique requires a long preparation time to obtain highly accurate data. Therefore, to solve this problem. This paper proposes: to monitor swimming 3D movement data based on the Mahony algorithm, which is used to quickly align initial values and to collect 3D movement data from swimmers via the MEMS nine-axis inertial measurement unit. In addition, the athlete's stance information is collected through angular velocity, acceleration, and correction of the stance angle to harvest accurate stance angle information. In addition, due to acceleration and magnetic field interference, conventional filtering cannot solve the accuracy degradation caused by this interference. Therefore, there is a need to integrate multiple types of information using complementary filtering methods in the Mahony algorithm in order to quickly obtain initial measurements of inertial attitude. And the feasibility of the algorithm will be further demonstrated by experiments with real motion scenarios.