RESEARCH OF KINEMATICS OF KARATE PUNCHES BASED ON AN ARTIFICIAL NEURAL NETWORK

The aim of the study was to investigate the automation of measuring the speed characteristics of karate athletes based on an artificial neural network with the use of Inertial Measurement Units (IMU).

Methods and organization of the research. A model based on an artificial neural network in the form of a multilayered perceptron was developed to recognize karate athletes' punches. The input parameters of the artificial neural network were absolute linear and angular velocities. The measurement was carried out with the use of Inertial Measurement Units (IMU) fixed on the athletes' hands. Along with the IMU in the case, a microcontroller for process control and a wireless transmitter were installed. The transmitter transferred data to the computer via Bluetooth for analysis using an artificial neural network. Weight of the package with IMU, microcontroller, and wireless transmitter was 35 grams. Three groups of athletes with different levels of training participated in the experiment. These groups were divided into subgroups – one for collecting kinematic data, the second was a control group and was used to assess the accuracy of the model.

Research results and discussion. As a result, the accuracy of the model was: the 1st group with a maximum of 1 year of training experience – 91.89 ± 3.45%, the 2nd group (2-3 years of training experience) – 95.33 ± 2.51%, the 3rd group (over 5 years of training experience) – 92.93 ± 4.33%.

Conclusion. Experiments revealed that the application of neural networks greatly facilitates the collection of data on the kinetic characteristics of athletes' punches, and allows automating the process. Such parameters of punch as velocity and acceleration can be determined during the fight with shadow, while identifying the type of punch and its parameters.

• Emily E Cust, Alice J Sweeting, Kevin Ball & Sam Robertson (2019) Machine and deep learning for sport-specific movement recognition: a systematic review of model development and performance, Journal of Sports Sciences, 37:5, 568-600, DOI: 10.1080/02640414.2018.1521769
• Polak, Ewa & Kulasa, Jerzy & VencesBrito, Antonio & Castro, Maria António & Fernandes, Orlando. (2016). Motion analysis systems as optimization training tools in combat sports and martial arts. Revista de Artes Marciales Asiáticas. 10. 105 – 123. 10.18002/rama.v10i2.1687.
• Malawski, Filip & Kwolek, Bogdan. (2018). Recognition of Action Dynamics in Fencing Using Multimodal Cues. Image and Vision Computing. 75. 10.1016/j.imavis.2018.04.005.
• Schöllhorn, Wolfgang & Jäger, Jörg & Janssen, D. (2008). Artificial neural network models of sports motions. Routledge Handbook of Biomechanics and Human Movement Science. 50-64.
• Robson Bonidia, Luiz A. L. Rodrigues, Anderson P. Avila-Santos, Danilo Sipoli Sanches, Jacques Brancher: Computational Intelligence in Sports: A Systematic Literature Review. Adv. Hum. Comput. Interact. 2018: 3426178:1-3426178:13 (2018).
• Rajšp, A.; Fister, I., Jr. A Systematic Literature Review of Intelligent Data Analysis Methods for Smart Sport Training. Appl. Sci. 2020, 10, 3013.
• Zhang, X.; Shan, G.; Wang, Y.; Wan, B.; Li, H. Wearables, Biomechanical Feedback, and Human Motor-Skills’ Learning & Optimization. Appl. Sci. 2019, 9, 226.
• Worsey MT, Espinosa HG, Shepherd JB, Thiel DV. Inertial Sensors for Performance Analysis in Combat Sports: A Systematic Review. Sports (Basel). 2019; 7(1):28. Published 2019 Jan 21. doi:10.3390/sports7010028
• Kasiri-Bidhendi, S., Fookes, C., Morgan, S., Martin, D. T., & Sridharan, S. (2015). Combat sports analytics: Boxing punch classification using overhead depth imagery. In 2015 IEEE International Conference on Image Processing (ICIP) (pp. 4545-4549). Quebec City, Canada: IEEE. https://doi.org/10.1109/ICIP.2015.7351667
• Kasiri, S., Fookes, C., Sridharan, S., & Morgan, S. (2017). Fine-grained action recognition of boxing punches from depth imagery. Computer Vision and Image Understanding, 159, 143-153. https://doi.org/10.1016/j.cviu.2017.04.007
• Hachaj, T., Ogiela, M. R., & Koptyra, K. (2015). Application of assistive computer vision methods to Oyama karate techniques recognition. Symmetry, 7(4), 1670-1698. https://doi.org/10.3390/sym7041670
• Shen, Yijun & Wang, He & Ho, Edmond & Yang, Longzhi & Shum, Hubert. (2017). Posture-based and Action-based Graphs for Boxing Skill Visualization. Computers & Graphics. 69. 10.1016/j.cag.2017.09.007.