A Milimeter Wave Radar-Driven Approach to Real-Time Motion Recognition in Exergames

The widespread adoption of sedentary lifestyles due to technological advancements has significantly reduced daily physical activity, contributing to both physical and cognitive health problems. Exergames-exercise-based video games-have emerged as a promising solution, particularly in the rehabilitation of conditions such as Mild Cognitive Impairment (MCI), by encouraging physical movement through interactive gameplay. However, traditional tracking methods that rely on depth cameras, inertial sensors, or webcam-based AI systems often suffer from limitations such as high cost, environmental sensitivity, calibration requirements, and poor performance under suboptimal conditions. This study enables the use of a new body-tracking system—equipped with precise, non-contact, and real-time motion detection capabilities based on millimeter-wave radar (MWR) technology—in exergames. The system, intended for use in exergames, utilizes rangedoppler data collected by the MWR sensor to train a 3D Convolutional Neural Network (3D CNN) for classifying three different types of movement: standing, walking, and bilateral arm lifts. In addition, the human motion detection system developed with mmWaveRadar sensor<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup> processes real-time radar data, classifies it via the model, and transmits it to a Unitybased game engine using TCP/IP communication. Within the scope of this study, the human motion detection system developed with mmWaveRadar sensor was integrated into two exergames. The first is the Balloon game, which responds to shoulder abduction movements by inflating and bursting a balloon to score points; the second is the Walk game, which detects walking movements to control the forward motion of a player avatar. These games demonstrated seamless, low-latency interaction between real-world physical actions and virtual gameplay elements. This study highlights the potential of MWR technology as a strong alternative to traditional motion capture systems in rehabilitation-focused exergames, offering a scalable, accurate, and user-friendly solution that can be easily adapted to both clinical and home-based therapy settings.<sup xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">1</sup>For more information, visit www.inosens.com.tr