Highly Accurate Ultrasonic Positioning and Tracking Systems

Download files
Access & Terms of Use
open access
Embargoed until 2017-03-23
Copyright: Khyam, Mohammad
Altmetric
Abstract
Localization is the process of determining the current location of a target(s) within given coordinates using a location system. The localization process has two main phases: firstly, the measurement phase, that establishes a relationship in terms of the distance and/or angle between the targets(s) to be localized and the system infrastructure; and secondly, the positioning phase, that exploits the measured information to calculate the absolute or relative location coordinates of the target(s). So far the most widely used positioning system is the Global Positioning System (GPS). However, in a GPS system, the receiver requires line-of-sight (LOS) reception from different satellites, which is usually difficult to obtain indoors. Therefore, as there is a need for alternate location systems in these GPS-obstructed environments, indoor positioning has drawn considerable attention from both academia and industry. Indoor environments, which are characterized by obstacles such as walls, floors, ceilings, and furniture, provide countless opportunities for a wide range of applications requiring different levels of accuracy. High-accuracy applications, such as gait analysis, usually use optical motion capture systems (MCSs), which are cost-prohibitive for many users, and also require complex arrangements of expensive equipment. In addition, they are responsive to changes in lighting and shadow. Therefore, with the aim of overcoming these limitations, ultrasonic positioning systems (UPSs) have drawn considerable attention. Usually, a narrowband UPS uses either a single tone or narrowband chirp signal in the measurement phase in which accurate estimations of distance, through time-of-flight (TOF) techniques, are fundamental. Generally, cross-correlation, which produces a peak at the time delay between a transmitted and received signal, is considered the optimal TOF estimation technique. Since their accuracy depends on the width of the correlation peak, which is inversely proportional to the signal’s bandwidth, these systems can only be said to be highly accurate if the reflected or multipath signal at the receiver is separated in time by more than the width of the correlation peak; otherwise, errors are introduced into the system. To improve the accuracy of such systems, the bandwidth of the signal must be increased, which increases the cost of the system. In the first part of this thesis, a new phase-correlation-based TOF estimation technique, that uses a narrowband chirp signal working in a closely spaced multipath environment, is proposed. In this system, the correlation peak becomes narrower by virtually, rather than physically, increasing the signal’s bandwidth, which reduces system cost. The performance of the proposed method is evaluated experimentally. As the correlation technique finds matches between transmitted and received signals, both signals need to be stored at the receiver, which increases hardware cost and computational complexity. In the second part of this thesis, firstly, to solve the limitations of the correlation technique, a narrowband orthogonal frequency division multiplexing (OFDM)-based TOF technique is introduced in which the TOF is estimated based on a pre-defined threshold. The proposed technique has advantages in terms of computational complexity and hardware cost when compared to the correlation technique, as in this approach TOF estimation decisions are made using threshold detection and only the received signal needs to be stored. An additional feature of this technique is that it has good noise cancelation properties. In the positioning phase, existing UPSs generally use the lateration algorithm to obtain a target’s location information. For the accurate positioning of reference points in an indoor environment, it is logistically simpler if they are installed in a fixed plane. This means that the distances between the reference points will usually be smaller than the distances between the reference points and the target(s). In this configuration, when lateration is used for positioning target(s), the surfaces of the spheres centered at the reference points will be almost parallel. This results in larger errors in the positions of the intersecting points of the spheres for directions tangential to the surfaces of the spheres than for those normal to these surfaces. This phenomenon is known as dilution of precision (DOP). To overcome this DOP problem, a steepest descent optimization algorithm is proposed. The proposed algorithm places the reference points at positions which best correspond to their measured distances from the target utilizing a three-dimensional (3D) rigid-body transformation. An additional feature of this algorithm is that it allows errors to be ignored in the distance measurements of the receivers corresponding to one complete cycle of the transmitted signal. Experimental results demonstrate the improvement obtained by the proposed methods over the traditional lateration based positioning systems that use cross-correlation techniques with a transmitted chirp signal. Finally, when UPSs use single tone or narrowband chirp signals, they cannot simultaneously localize multiple targets due to signal interference. This is generally overcome by either using the time division multiplexing (TDM) technique, which reduces the positioning update rate, or introducing a broadband transducer, which increases system cost. In the last part of this thesis, the proposed OFDM-based steepest descent optimization algorithm is first extended to handle multiple target positioning utilizing the orthogonal property of the OFDM signal. Since TDM techniques and broadband transducers are not required, this system can maintain the single target system update rate without increasing system cost. For the positioning of a moving target(s), most of the existing localization systems use the matched filtering technique, where a bank of correlators is used to estimate the Doppler shift associated with the target’s movement. This requirement increases computational complexity and system cost. The proposed OFDM-based steepest decent optimization algorithm is further extended for tracking a moving target. The Doppler shift is estimate by introducing a pilot carrier to the transmitted OFDM signal. As it does not require matched filters to estimate the Doppler shift, the proposed system does not require extra computational complexity or system cost. The accuracy of the proposed system is compared with an optical motion capture system, Vicon, and it is shown to have the same order of precision while incurring less cost and complexity.
Persistent link to this record
Link to Publisher Version
Link to Open Access Version
Additional Link
Author(s)
Khyam, Mohammad
Supervisor(s)
Pickering, Mark
Lambert, Andrew
Creator(s)
Editor(s)
Translator(s)
Curator(s)
Designer(s)
Arranger(s)
Composer(s)
Recordist(s)
Conference Proceedings Editor(s)
Other Contributor(s)
Corporate/Industry Contributor(s)
Publication Year
2015
Resource Type
Thesis
Degree Type
PhD Doctorate
UNSW Faculty
Files
download public version.pdf 3.58 MB Adobe Portable Document Format
Related dataset(s)