基于LoRa信号的行走距离和方向估计

行走追踪在真实场景中具有广泛应用,可以用于安防监控、老人看护、室内导航等场景.近年来,基于无线射频信号的非接触式行走追踪受到了研究人员的广泛关注,包括基于Wi-Fi信号、RFID信号等的行走追踪系统.然而,现有的行走追踪系统依然面临感知范围小、感知设备部署稠密等问题.在本文中,我们首次将用于物联网低功耗、远距离通信的LoRa技术应用到非接触式的大范围行走追踪中,显著地增加了行走追踪系统的感知距离.特别地,通过利用LoRa网关上配置的多天线,利用两根天线上接收信号的比,可以有效地消除噪声干扰以及收发不同步带来的误差,从而进一步提升了感知范围,然后利用计算切线向量相位变化的方法准确计算原始信号中动态分量的相位变化来在估计行走距离和方向.基于此,本文提出基于LoRa的非接触感知系统,可以在一段自然连续的行走活动中准确地判断人的动静状态并切割出行走片段,进而计算出行走的距离和方向,实现人的行走追踪.实验验证了系统计算行走距离和方向的准确性和实时性,其中距离计算的平均误差为3.8%,准确判断行走方向所需时间为0.7 s.

Walking Distance and Direction Estimation Using LoRa Signal

Human walking tracking is widely used in many real applications,such as security monitoring,elderly care and indoor navigation.In recent years,RF based contactless human walking tracking has attracted much attention from researchers,including Wi-Fi based or RFID based human walking tracking system.However,existing sensing systems face some serious problem,such as short sensing range,dense device deployment and weak through-wall sensing capability.In this paper,we are the first to explore LoRa,originally used for low-power and long-distance communication in Internet of Things(IoT),to track human walking at a long distance in a contactless way.This new sensing method can significantly increase the sensing range of the system.Particularly,we exploit the multiple antennas equipped on LoRa gateway,and propose to utilize ratio of the received signals from multiple antennas.This proposed division operation can effectively remove the signal noise and cancel out the random phase offsets,and further increase the sensing range.Then,the phase change of the dynamic component in the original signal is calculated accurately by calculating the phase change of the tangent vector to estimate the walking distance and direction.Thus,we propose the LoRa based contactless sensing system,which can accurately determine the motion status of the human target and segment the walking activity in the continuous signal.Then,the distance and direction of walking can be accurately calculated to achieve the human walking tracking.Experimental results demonstrate the accuracy and robustness of the system.Our system achieves an average error of 3.8%in walking distance estimation and can reliably detect walking direction with 100%accuracy for a latency of 0.7 s.