Using Artificial Neural Network For The Kick Techniques Classification - An Initial Study

In this initial study it is investigated the possibility of using simple artificial neural network for classification of kick techniques based on their specific force course profile. The aim is to investigate whether the neural networks could be a suitable tool for such task and can be possibly used in following research that will deal with classification of punch techniques and also the striker’s gender and level of training. INTRODUCTION The kick techniques are (apart from punching techniques) the most important and effective techniques in unarmed professional defense with significant force delivery. Various kick techniques are the subject of research investigation mostly for the needs of martial arts. (Liu et al. 2000, Pieter and Pieter 1995). This paper presents initial results of analysis of two different kick techniques: the direct kick and the round kick (Liu et al. 2000). The aim was to find out whether it is possible to distinguish these two techniques from a kick impact force profile. In this long-term research the participants were asked to perform a set of different punch and kick techniques on a measuring station. The impact force profiles were stored for further analysis. To uncover whether there are certain unique characteristics for the two kick techniques mentioned above the artificial neural network (ANN) was chosen as a suitable classifier. Firstly, kick techniques are explained. In the following paragraph, measuring devices, the method of data storage and experiment setup for measurement are described. Artificial neural network theory is depicted in the next section. Problem definition and consequent analysis are followed by result section. The conclusion summarizes the kick techniques classification. KICK TECHNIQUES In this study two different kick techniques are distinguished the direct kick (Fig. 1) and the round kick (Fig. 2). In professional defense, these kicks are used to stop and keep the attacker in distance where the attacker cannot touch us. The second way of use is destabilization of attacker. During the direct kick a sole or a heel are the hit areas. This kick is made directly and by the shortest way to the target. During the round kick an instep together with part of shank are hit areas. The direct kick is considered to be stronger than the round kick. Figure 1: Direct kick Figure 2: Round kick MEASURING DEVICES The strain gauge sensor L6E-C3-300kg (Fig. 3.) works as unilaterally cantilever bending beam. During force delivery the biggest deformation of sensor is in places with the thinnest walls – there are metal film strain gauges which change their electrical resistance depending on deformation. Strain gauges are plugged in Wheatstone bridge and this way is possible to convert difference of resistance to electrical signal which we can process. Figure 3: Strain gauge sensor L6E-C3-300kg Proceedings 28th European Conference on Modelling and Simulation ©ECMS Flaminio Squazzoni, Fabio Baronio, Claudia Archetti, Marco Castellani (Editors) ISBN: 978-0-9564944-8-1 / ISBN: 978-0-9564944-9-8 (CD) The sensor is connected to the computer, which is used for data storage, through the strain gauge. The strain gauge type TENZ2334 is an electronic appliance that converts the signals to data that is stored in memory. The core of the appliance is a single-chip microcomputer that controls all of the activities. The strain gauge sensor is connected to this appliance via four-pole connector XLR by four conductors. The number of values measured by the sensor averages around 600 measurements per second while the data is immediately stored in the memory of a device with a capacity of 512 kB (Lapkova et al., 2012). The mentioned above strain gauge sensor was placed on the measuring station according to the following schematic (Fig. 4): Figure 4: Measuring station schematic 1 – punching bag (made from hardened vinyl filled with foam) 2 –template 3 – strain gauge sensor L6E-C3-300kg 4 – board (200 x 200 x 5 mm) 5 – punching bag base