[3S4-IS-4] Robots and real worlds

[3S4-IS-4-02] Counting and tracking people with mmWave sensor in the restaurant

〇Shenglei Li¹, Reiko Hishiyama¹ (1. Waseda University)

Regular

キーワード：millimeter wave radar, counting , tracking

Abstract—One key to creating the customized service in the future smart city is to let the smart space know where people are and how many they are. The visual method is a scheme to recognize people with very high accuracy, but concerns also arise regarding potential privacy leakage, environment limitations, user’s nonacceptance, etc. We propose a human tracking and counting system based on a millimeter wave radar (mmWave) alternative to conventional visual solutions. In this work, we collect the point cloud dataset and evaluate our system in the restaurant in Tokyo by a commercial low-cost, off-the-shelf mmWave FMCW radar. To better understand the performance of our mmWave system, we also apply the visual systems using body recognition and facial recognition, respectively. Our mmWave system performs a counting accuracy of over 83% with position tracking in the field test, showing a potential capacity to prevent crowds during the Covid-19.