Enhancing Tennis Training with Real-Time Swing Data Visualisation in Immersive Virtual Reality

TL;DR

The paper presents a VR tennis training system that delivers real-time swing metrics via a wearable IMU and evaluates its impact on novice players. By overlaying swing speed and swing power within a PC-VR environment, the study shows increased higher-end swing activity and selective accuracy improvements, with participants reporting strong engagement but mixed perceptions of the overlay’s integration. The findings suggest real-time data visualization in VR can enhance motor learning and decision-making for beginners, with implications for broader adoption of immersive, data-driven training tools. Overall, this work demonstrates the feasibility and potential of immersive visual analytics to democratize advanced sports training.

Abstract

Recent advances in immersive technology have opened new possibilities in sports training, especially for activities requiring precise motor skills, such as tennis. In this paper, we present a virtual reality (VR) tennis training system integrating real-time performance feedback through a wearable sensor device. Ten participants wore the sensor on their dominant hand to capture motion data, including swing speed and swing power, while engaging in a VR tennis environment. Initially, participants performed baseline tests without access to performance metrics. In subsequent tests, participants executed similar routines with their swing data displayed in real-time via a VR overlay. Qualitative and quantitative results indicated that real-time visual feedback led to improved performance behaviors and enhanced situational awareness. Some participants exhibited increased swing consistency and strategic decision-making, though improvements in accuracy varied individually. Additionally, subjective feedback highlighted that the immersive experience, combined with instantaneous performance metrics, enhanced player engagement and motivation. These findings illustrate the effectiveness of VR-based data visualisation in sports training, suggesting broader applicability in performance enhancement.

Figures (15)

• Figure 1: Component connection diagram - 3.7 V LiPo battery connects to the PowerBoost Charger using a JST connector. The PowerBoost Charger provides a stable 5 V to the Arduino Nano ESP32 using a USB-A to USB-C cable. Highlighted in yellow is the area for the soldered USB-A port. I2C communication occurs between pins A4 and SDA (Yellow) and A5 and SCL (Orange). 3.3 V pin supplies 3.3 V to the BNO055 module (red), and ground connection is made between GND pins (black).
• Figure 2: BNO055 Calibration (Left) and Arduino Cloud Connection (Right) Flowcharts. Calibration is achieved by moving the BNO055 module in specific axis patterns. The program checks for the calibration status before initialising Arduino Cloud connection. Cloud credentials are initialised before attempting a connection. The process finishes once the Cloud connection is achieved.
• Figure 3: Complete software flowchart - Data acquisition begins after Wi-Fi connection, calibration, and cloud connection are achieved. The program enters the main loop if a valid swing has been detected. Sensor data is sampled to obtain swing speed and then interpreted to calculate normalised swing power. After each swing, peak values are logged and the Cloud is updated. The process repeats until ended by the user.
• Figure 4: Testing Workflow - A preliminary practice session was followed by the baseline test. After an informal verbal discussion, the participants had a 5-10 minute break while the data visualisation overlay was set up. The visualisation test concluded the physical portion of the study and participants were given a questionnaire after testing.
• Figure 5: Player with VR headset and glove - Glove is only used on the right hand.