Track estimation using link line crossing information in wireless networks

Abstract

Abstract-Device-free or non-cooperative localization uses the changes in signal strength measured on links in a wireless network to estimate a person's position in the network area. Existing methods provide an instantaneous coordinate estimate via radio tomographic imaging or location fingerprinting. In this paper, we explore the problem of, after a person has exited the area of the network, how can we estimate their path through the area? We present two methods which use recent line crossings detected by the network's links to estimate the person's path through the area. We assume that the person took a linear path and estimate the path's parameters. One method formulates path estimation as a line stabbing problem, and another method is a linear regression formulation. Through simulation we show that the line stabbing approach is more robust to false detections, but in the absence of false detections, the linear regression method provides superior performance.

Device-Free Electromagnetic Passive Localization Using Link Line Information in Wireless Sensor Networks

“…The proposed method only exploits recent link lines detected by wireless sensor networks to estimate the target's location and thus reduces the algorithm's storage and computational requirements. Moreover, unlike geometric-based DFPL methods in [13][14][15], the proposed method makes use of prior information to improve the tracking results, making it more robust to noise.…”

“…Generally, after the affected link identification and outlier link rejection, the links in set S t can be considered as real links which have been sufficiently influenced by the presence of the target. In the LLC algorithm [15], the affected links in the network are represented as line segments, and then the LLC algorithm uses ordinary least squares (LS) to estimate the intersection points of these line segments. Different from applying LS method to estimate intersection points in [15], in this paper we seek to find a pivot point by forming virtual triangles consisting of links.…”

“…In the LLC algorithm [15], the affected links in the network are represented as line segments, and then the LLC algorithm uses ordinary least squares (LS) to estimate the intersection points of these line segments. Different from applying LS method to estimate intersection points in [15], in this paper we seek to find a pivot point by forming virtual triangles consisting of links. We choose any three links out of S t to form a virtual triangle.…”

“…This phenomenon is mainly due to the fact that when the target changes direction, some transient error will occur while the variations of wireless links traveling through the target are abrupt and discontinuous. Then, we compared the GDFPL method with two state-of-the-art DFPL schemes, i.e., the geometric-based LLC algorithm in [15] and the model-based RTI algorithm in [6] under the same conditions. In the RTI method, the width of the ellipse is set as 0.3 m, and grid size is 0.25 m × 0.25 m. Tables 1(a) and (b) summarize the statistical results of tracking errors in both scenes, respectively.…”

“…In addition, to generate a location estimate model-based DFPL methods need perform operations on large matrices for solving the underdetermined problem which results in high storage and computing requirements. Different from the model-based methods, Zhang et al [13,14] first proposed the geometric-based DFPL method and further explored by Hillyard et al [15]. Instead of exploiting the shadowing model, the geometric-based DFPL method detects the wireless link line information affected by targets to realize DFPL.…”