Inertial Measurement Unit based Virtual Antenna Arrays - DoA Estimation and Positioning in Wireless Networks

Sammanfattning: Today we have different location based services available in a mobile phone or mobile station (MS). These services include: direction finding to nearby ATMs, locating favorite food restaurants, or finding any target destination. Similarly, we see different applications of the positioning and navigation systems in firefighting or other rescue operations. The common factor in almost all of the location based services is the system's ability to determine the user's current position, with reference to a floor plan or a navigation map. Current technologies are using sensor data measurements from one or more sensors, available to the positioning device, for positioning and navigation. Typical examples are radio based positioning such as global positioning system, inertial sensors based inertial navigation system, or camera based positioning systems. Different accuracy and availability conditions of the positioning and navigation solution can be obtained depending on the positioning algorithms and the available sensor information.Nowadays, the focus of research in positioning and navigation has been mostly on the use of existing hardware infrastructure and low-cost solutions, such that the proposed technique can be deployed with ease and without extra infrastructure requirements as well as without any expensive sensor equipment. In this work, we investigate a novel idea for positioning using existing wireless networks and low-cost inertial sensor measurements available at the MS. We propose to use received baseband radio signal along with inertial sensor data, such as accelerometer and rate gyroscope measurements, for direction of arrival (DoA) estimation and positioning. The DoA information from different base stations or access points can be used to estimate the MS position using triangulation technique. Furthermore, due to size and cost restrictions it is difficult to have real antenna arrays at the MS, the idea of DoA estimation and positioning is proposed to be used with single antenna devices by using the so-called virtual antenna arrays.We have presented our research results in three different papers. We provide measurement based results to perform a quantitative evaluation of DoA estimation using arbitrary virtual antenna arrays in 3-D; where a state-of-the-art high-resolution algorithm has been used for radio signal parameter estimation. Furthermore, we provide an extended Kalman filter framework to investigate the performance of unaided inertial navigation systems with 3-axis accelerometer and 3-axis rate gyroscope measurements, from a six-degrees-of-freedom inertial measurement unit. Using the extended Kalman filter framework, we provide results for position estimation error standard deviation with respect to integration time for an unaided inertial navigation system; where the effect of different stochastic errors noise sources in the inertial sensors measurements such as white Gaussian noise and bias instability noise is investigated. Also, we derive a closed form expression for Cramér-Rao lower bound to investigate DoA estimation accuracy for a far-field source using random antenna arrays in 3-D. The Cramér-Rao lower bound is obtained using known antenna coordinates as well as using estimated antenna coordinates, where the antenna coordinates are estimated with an uncertainty whose standard deviation is known. Furthermore, using Monte-Carlo simulations for random antenna arrays, we provide Cramér-Rao lower bound based performance evaluation of random 3-D antenna arrays for DoA estimation.