Data Fusion in Ubiquitous Sports Training: Methodology and Application

Abstract

We present a data fusion-based methodology for supporting the sports training. Training sessions are planned by coach on the basis of the analyzed data obtained during each training session. The data are usually acquired from various sensors attached to the athlete (e.g., accelerometers or gyroscopes). One of the techniques dedicated to processing the data originatnig from different sources is data fusion. The data fusion in sports training provides new procedures to acquire, to process, and to analyze the sports training related data. To verify the effectiveness of the data fusion methodology, we design a system to analyze training sessions of a tennis player. The main functionalities of the system are the tennis strokes detection and the classification based on data gathered from the wrist-worn sensor. The detection and the classification of tennis strokes can reduce the time a coach spends in analyzing the trainees’ data. Recreational players for self-learning may also use these functionalities. In the proposed approach, we used Mel-Frequency Cepstrum Coefficients, determined from the accelerometer data, to build the feature vector. The data are gathered from amateur and professional athletes. We tested the quality of the designed feature vector for two different classification methods, that is, k-Nearest Neighbors and Logistic Regression. We evaluate the classifiers by applying two tests: 10-fold cross-validation and leave-one-out techniques. Our results demonstrate that data fusion-based approach can be used effectively to analyze athlete’s activities during the training.

Remote Evaluation System of Tennis Batting Action Standard Based on Acceleration Sensor

“…Moreover, compared with previous methods, our method based on a bracelet embedded with an acceleration sensor is cost-effective, more energy-efficient, not influenced by the lighting environment, and easier to set up. Compared with five related studies that used systems such as IMU gyroscopes [15], Zepp Tennis Smart Sensor 2 [16], Brzostowski K.'s system [17], TenniVis [19], and Tac-Simur [20], our system performance is more stable and has stronger practical applicability. Moreover, our system strictly protects the user's privacy when uploading the user's data remotely.…”

“…We performed an analysis of evaluation results and system performance test that included signal smoothing performance and real-time response in Section 3. In the process of real-time response test of system performance, we compared with five related studies that used the system such as IMU gyroscopes [15], Zepp Tennis Smart Sensor 2 [16], Brzostowski K.'s system [17], TenniVis [19], and Tac-Simur [20]. e results demonstrated the superiority of our system.…”

“…A research used Zepp Tennis Smart Sensor 2 [16] to monitor technical-tactical actions and physical activity during a current tennis competition in the green stage; however, it also had a larger latency. A methodology based on data fusion for sports training [17] has been presented, and a system was developed to assess the performance of a tennis player so as to improve their strokes technique and overall performance (i.e., we call it Brzostowski K.'s system to compare with our system), but it required a large amount of data to train the model, and data collection was difficult. Meanwhile, the model was prone to overfitting resulting in inaccurate evaluation results.…”

Tennis Stroke Classification

“…In the tennis domain itself, there is different related work that classifies various tennis stroke types using IMU data and different sensor positions for the data collection. The majority of approaches use a sensor strapped on the wrist of the dominant arm [4,6,18,20,28,29]. Connaghan et al [10] use a similar position instead: on the middle of the forearm.…”

“…As sensors, both self-built prototypes [18,20,22,23] and commercial sensing systems [4,6,7,21,28,29] have been used. Most frequently, the IMUs used process acceleration within ±16д, and the gyroscopes used a range of ±2000 deg s [18,23,29,31].…”