IR-UWB Radar-Based Situational Awareness System for Smartphone-Distracted Pedestrians

The increasing prevalence of smartphones has made pedestrian safety on roads a significant concern due to distractions caused by these devices. This paper introduces a unique and real-time assistance system, UWB-assisted Safe Walk (UASW), designed to detect obstacles and terrain features and warn users about ongoing situations. Our proposed technique harnesses the capabilities of Impulse Radio Ultra-Wideband (IR-UWB) radar integrated within the smartphone, which delivers exceptional range resolution and strong noise immunity employing short pulses. We developed UASW specifically for Samsung smartphones featuring IR-UWB connectivity with machine/deep learning (ML/DL) techniques. Our framework utilizes complex Channel Impulse Response (CIR) data to combine signal processing-based detection with artificial neural network (ANN)-driven classification for obstacles and terrains. The performance of our suggested UASW system is assessed using real-world collected data. Experimental outcomes indicate that the proposed UASW system achieves an obstacle detection accuracy rate of up to 97 % and obstacle and terrain classification accuracy of approximately 95%. With an inference latency of around 26 milliseconds, UASW effectively supports smartphone-distracted pedestrians and enhances their situational awareness.