Pedestrian and Vehicle Area Positioning With Multiple UWB Signals

In industrial environments where pedestrians and vehicles closely interact, existing point positioning systems often yield inaccurate results by disregarding crucial vehicle information, such as shape and angles, thereby compromising effective early warning mechanisms. Moreover, existing positioning systems relying on visual or radar methodologies also have limitations due to environmental factors. To address these challenges, this article introduces a novel pedestrian and vehicle area positioning and warning system base on multiple ultrawideband (UWB) signals, offering detailed information about vehicle positioning, orientation, and shape. Employing multiple UWB tags alongside the two-way ranging (TWR) algorithm, this system models the vehicle's positioning area and triggers alerts when the pedestrian-to-vehicle distance falls below a specified threshold derived from the shape and orientation data of the positioning area. Additionally, this work proposes a priori data for coordinate calibration, integrating extended Kalman filtering (EKF) and a dynamic threshold algorithm to seamlessly recalibrate the vehicle's positioning in both stationary and mobile scenarios. Experimental results demonstrate that, compared with using original area data for positioning, the modified algorithm can reduce the positioning coordinate STD by 18.43%, and the vehicle body shake variance by 50.56%, while also achieving a 4.30% increase in success warning rate and a notable 33.35% decline in false warning rate.