2010 International Conference on Indoor Positioning and Indoor Navigation (IPIN), 15-17 September 2010, Zürich, Switzerland 978-1-4244-5864-6/10/$26.00 ©2010 IEEE Linear Antenna Array, Ranging and Accelerometer for 3D GPS-Less Localization of Wireless Sensors Patryk Mazurkiewicz, Athanasios Gkelias and Kin K. Leung Imperial College London, Exhibition Road, SW7 2AZ, London, UK [patryk.mazurkiewicz06, a.gkelias, kkleung]@imperial.ac.uk Abstract—Localization algorithms for wireless sensors can use diverse measurement hardware as a source of geographical information in order to provide better location estimates. A relatively extensive set of measurement hardware comprises of machinery capable of measuring range, angle-of-arrival and earth gravity direction. Literature covers distance vector (DV) exchange algorithms [1] for the measurements which provide the complete distance vector. In this work we develop a method for computing the complete DV in a distributed way, using linear antenna array and ranging to calculate a complete DV. We develop a localization algorithm that uses this concept. We also determine pathological situations of flip- error [2] occurrence, but we incorporate this knowledge into the algorithm so that we prevent the flip-errors from occurring. Keywords — 3-D localization, 3-D positioning, Angle-of- arrival, Linear antenna array, Hybrid localization, Multi-hop network, Wireless sensor; I. INTRODUCTION Localization capability is required in many applications of wireless sensor networks (WSN). For example, monitoring WSN of any type requires node location awareness in order to stamp the measurements with the location (and also with timestamp), otherwise the measurement is of no meaning for the infrastructure owner [3]. A particularly interesting indoor localization scenario is node positioning for automated building monitoring. The goal there is to greatly reduce the human-generated mistakes and errors, and the overhead of manual work during the deployment by automating the localization process. The 3-D localization of sensors in WSNs used in construction measurements, such as buildings, tunnels, bridges etc., is a very challenging task. This task becomes even more challenging in multi-hop WSN scenarios, where network connectivity (i.e. the number of neighbors) locally may become relatively low, e.g. in narrow passages. For those reasons we focus on an algorithm which makes use of diverse sources of geo-information, namely: ranging, angle-of-arrival (AOA) and accelerometer for earth gravity direction awareness. There are two main families of localization algorithms in the literature. The first family comprises of algorithms based on ranging techniques, i.e., Cricket [4]. Those algorithms require that the node undergoing localization has at least 3 location-aware neighbors in 2-D and 4 of them in 3-D [5]. To the second family there belong localization algorithms that are based on the measurement of the angle-of-arrival of signals from neighbor nodes [6]. Hybrid schemes also exist in the literature that obtain more robust results by combining ranging and AOA techniques, such as [1], [7], [8]. DV algorithms inherently make use of both angle and range measurements, since rich information is needed in order to determine the DV. Motivation and contributions : Our focus in this work is the sensor localization in three-dimensional (3-D) WSNs where sensors employed are equipped with linear antenna arrays. For a complete AOA direction estimation (comprising of a unique azimuth and elevation), a 3-D antenna array is required. On the other hand, planar antenna arrays produce two indistinguishable estimates of AOA, while linear arrays give only the elevation-type measurement and therefore provide a class of candidate directions instead of a particular direction or directions. Unfortunately, 3-D or planar antenna arrays are difficult to be deployed on wireless sensors, mainly due to higher cost and size limitations. Therefore, our motivation behind this work is to provide a novel and robust distributed algorithm for localization in 3-D WSNs employed with very basic linear antenna arrays (of only 2 elements), by allowing sensors to exchange their measurements. Our main contributions are threefold and can be summarized in the following: 1. We develop a method for computing the complete DV in a distributed way when only linear antenna arrays are available. 2. based on the proposed method we develop a novel distributed localization algorithm for 3-D WSNs. 3. We research the nature of flip-error of the proposed algorithms and we propose ways of avoiding it. The remainder of this paper is organized as follows. In section II the terminology is introduced. Section III focuses on explaining the principia of the concept of using linear antenna array, whereas section IV includes more detailed description of the concept. The reader can find there the discussion on the “three possible orientations” that matter for the performance of the algorithm. Section V describes the basic variant of the localization algorithm which uses our novel idea. Finally, there is a summary and a broad overview of further and ongoing research in section VI. II. OVERVIEW AND TERMINOLOGY Lets assume a WSN where all sensors are equipped with the ranging device, accelerometer and 2-element antenna array. Using information obtained by this hardware the sensors can calculate circumferences of candidate locations of the target.