mmBox: Harnessing Millimeter-Wave Signals for Reliable Vehicle and Pedestrians Detection

Abstract

Object detection plays a pivotal role in various fields, for example, smart traffic system relies on the detected results for decision-making. However, existing researches predominately utilize optical camera and LiDAR, which exhibit limitations in adverse outdoor environment, like fog weather. To address these challenges, millimeter-wave (mmWave) attracts researcher’s attention to detect objects in severe conditions since it can work effectively in low visibility condition and penetrate small obstacles. Yet, previous mmWave-based works have shown limited performance, such as no shape information of objects. Therefore, we design and implement a two stage system, mmBox , to accurately predict bounding box with depth for vehicles and pedestrian, which first generates heatmaps in different dimensions and then leverages a deep learning model to extract features for predictions. To evaluate the performance of mmBox , we collected real-world mmWave reflections from urban traffic intersections and dense foggy environment, and the extensive evaluation metrics show remarkable accuracy and low latency of our model.