Using Wearable Sensors for Real-Time Recognition Tasks in Games of Martial Arts - An Initial Experiment

Abstract

Beside their stunning graphics, modern entertain-compensate for various inaccuracies inherent in inexpensive ment systems feature ever-higher levels of immersive user-wearable sensor system. The inaccuracies result from the interaction. Today, this is mostly achieved by virtual (VR) and limited resolution and sampling rate of the sensors, variations augmented reality (AR) setups. On top of these, we envision to add ambient intelligence and context awareness to gaming in sensor placement, dynamic sensor displacement during applications in general and games of martial arts in particular. user motion, and other sources of noise (e.g., environmental To this end, we conducted an initial experiment with inexpensive magnetic fields or temperature-related sensor drift). The body-worn gyroscopes and acceleration sensors for the Chum challenge in algorithmic design is to find features that are Kiu motion sequence in Wing Tsun (a popular form of Kung sensitive to the relevant motion characteristics and at the Fu). The resulting data confirm the feasibility of our vision.. .s Fine-tuned adaptations of various thresholding and pattern-same time insensitive to the inaccuracies mentioned before. matching techniques known from the fields of computational As decribed in Section III, we conducted an initial experintelligence and signal processing should suffice to automate the iment capturing Wing Tsun movements with wearable gyroanalysis and recognition of important Wing Tsun movements scopes and acceleration sensors to test the feasibility of our in real time. Moreover, the data also seem to allow for the vision. Section IV discusses and analyzes the experimental possibility of automatically distinguishing between certain levels vision. Section IV discussesand ana es the erieon of expertise and quality in executing the movements.results, provilng to be very promsilng indeed. Based thereon, Keywords: Body-worn Sensors, Experiment, Games of it certainly seems worthwhile to continue in this direction Martial Arts, Kung Fu, Motion Analysis, Movement Recog-and try to automate at least some parts (if not all) of an nition, Wearable Computing, Wing Tsun expert analysis for many important Wing Tsun movements. Beside in sports and game play, martial arts from the I. INTRODUCTION Far East gain ever more popularity and importance in many Video analysis and motion capturing are standard tools in other areas as well. Tai Chi, for instance, is of special professional sports to monitor and improve athletic perfor-interest because clinical studies show that it helps to reduce mance by recognizing and fine-tuning the quality of move-the probability of falling, especially for the elderly [l] and ment. Cutting-edge systems with high-quality sensors hardly patients with chronic conditions [2]. suffice to fulfill these professionals' needs. Quite often, trainers and other experts still process the recorded data by II. RELATED WORK hand. The whole setup and procedure are not only expensive By now, many independent researchers have demonstrated and...

Power-Aware Activity Monitoring Using Distributed Wearable Sensors

“…The accuracy of such a classifier depends on a variety of parameters including the classification algorithm, sensor node placement, types of features extracted from the signal, and number and type of actions/activities to be recognized. While many algorithms such as k-nearest neighbor (k-NN) [7], hidden Markov models (HMM) [8], Naive Bayes [9], and support vector machines [10] have been investigated, the k-NN and HMM are more common in action recognition in the wireless healthcare domain when motion sensors are generally used for information inference.…”

Postural activity monitoring for increasing safety in bomb disposal missions

“…The variety of systems and applications in the literature, similar to those described above, show that posture tracking is a relatively well-covered research subject with a number of branches and applications: from activity detection [20,21] to position recognition [6,8,9], to real-time movement recognition tasks for martial arts [22] and manufacturing environments [23], added to gait measurement [24]. The systems reported, although by and large application specific, often share a common sensor placement on the body in order to accurately detect the subject's movement and limb positions [25][26][27] but require different degrees of movement sensing accuracy to fulfil the specific application.…”

Human Activity Recognition Using Accelerometer and Gyroscope Sensors

“…It is discovered that exercises can be recognized by utilizing moderately few elements [4], [5], [9]. The studies have also been performed to see if using more components will produce better results [10], [11]. Table I display a comparative analysis of the earlier works performed, for detecting activities using phone sensors.…”