Giuseppe Destino

On the Maximum Likelihood Approach for Source and Network Localization

We consider the source and network localization problems, seeking to strengthen the relationship between the Weighted-Least-Square (WLS) and the Maximum-Likelihood (ML) solutions of these problems. To this end, we design an optimization algorithm for source and network localization under the principle that: a) the WLS and the ML objectives should be the same; and b) the solution of the ML-WLS objective does not depend on any information besides the set of given distance measurements (observations). The proposed Range-Global Distance Continuation (R-GDC) algorithm solves the localization problems via iterative minimizations of smoothed variations of the WLS objective, each obtained by convolution with a Gaussian kernel of progressively smaller variances. Since the last (not smoothed) WLS objective derives directly from the ML formulation of the localization problem, and the R-GDC requires no initial estimate to minimize it, final result is maximum-likelihood approach to source and network localization problems. The performance of the R-GDC method is compared to that of state-of-the-art techniques such as semidefinite programming (SDP), nonlinear Newton least squares (NLS), and the Stress-of-a-MAjorizing-Complex-Objective-Function (SMACOF) algorithms, as well as to the Cramér-Rao Lower Bound (CRLB). The comparison reveals that the solutions obtained with the R-GDC algorithm is insensitive to initial estimates and provides a localization error that closely approaches that of the corresponding fundamental bounds. The R-GDC is also found to achieve a complexity order comparable to that of the SMACOF, which is known for its efficiency.

Weighing strategy for network localization under scarce ranging information

We propose a robust non-parametric strategy to weight scarce and imperfect ranging information, which is shown to significantly improve the accuracy of distance-based network localization algorithms. The proposed weights have a dispersion component, which captures the effect of noise under the assumption of bias-free samples, and a penalty component, which quantifies the risk of the latter assumption and penalizes it proportionally. The dispersion weights result from the application of small-scale statistics with confidence bounds optimized under a maximum entropy criterion that mathematizes the empirical concept of reliability commonly found in related literature. In turn, the penalty weights are derived from the relationship between the risk incurred by the bias-free assumption and the geometry of 3-node cliques, established by statistical-geometry. The performance of the distance-based network localization algorithm employing the proposed dispersion-penalty weights is compared against the Cramer-Rao lower bound (CRLB) and to equivalent algorithms employing alternative weights. The comparison reveals that, amongst the alternatives, the network localization algorithm with the proposed weights performs best and closest to an unbiased estimator.

MAC Performances for Localization and Tracking in Wireless Sensor Networks

Time delay rather then throughput, is a constraint of greater importance in tracking systems. In particular, the maximum accces delay permissible by the application ins strongly related to the dynamics of theracked objects. The purpose of this article is to study the performance of a new media access control (MAC) technology specifically suited for LDR UWB systems [3] under the point of view of a tracking application. Specifically, the time delay necessary to collect the ranging information in both, star and meshed topology networks, have been studied as function of the number of mobiles in the network. More importantly, we propose two new ranging packet to be used inside the aforementioned MAC, in order to achieve in both the network topologies, a clear advantage compared to the current solution.

5G Position and Orientation Estimation through Millimeter Wave MIMO

Millimeter wave and massive MIMO are considered enabling technologies for future 5G networks. While their benefits for achieving high-data rate communications are well- known, their potential advantages for accurate positioning are largely undiscovered. We derive sufficient conditions under which transmission from a single mm-wave base station leads to a non- singular Fisher information matrix associated with the position and orientation of a user terminal equipped with multiple antennas, which is in turn a prerequisite for joint estimation of the position and orientation.

Indoor positioning: A key enabling technology for IoT applications

Motivated by the recent advances on internet of things (IoT) and the importance that location information has on many application scenarios, this article offers references to theoretical and localization-algorithmic tools that can be utilised in connection with IoT. We develop this discussion from basic to sophisticated localization techniques covering also some less-intuitive notions of localization, e.g. semantic positioning, for which we provide a novel solution which overcome the problem of privacy. We analyze the localization problem from a mathematical perspective; reviewing the most common and best-performing class of localization methods based on optimization and algebraic approaches and we discuss benefits of location information in a wireless system. In this regard we discuss few concrete applications scenario currently under investigation in the largest EU project on IoT, namely the FP-7 Butler project, how location information is one of the key enabling technology in the IoT. In addition to the theoretical aspect, this article provides references to the pervasive localization system architecture using the smart sensors developed within the Butler project.

Understanding and Solving Flip-Ambiguity in Network Localization via Semidefinite Programming

We employ the semidefine programming (SDP) framework to first analyze, and then solve, the problem of flip-ambiguity afflicting range-based network localization algorithms with incomplete ranging information. First, we study the occurrence of flip-ambiguous nodes and errors due to flip ambiguity by considering random network topologies with successively smaller connectivity ranges RMax >RMax - ΔR >. . . >RU >RL, and employing an SDP-based unique localizability test to detect the limiting connectivity ranges RU and RL that are respectively sufficient and un-sufficient to ensure unique localizability. Then, we utilize this information to construct an SDP formulation of the localization problem with Genie-aided constraints, which is shown to resolve flip-ambiguities. Finally, we derive a flip-ambiguity-robust network localization algorithm by relaxing the Genie-aided constraints onto feasible alternatives. Finally, the performance of the so-obtained localization algorithm is studied by Monte-Carlo simulations, which reveal a substantial improvement over the conventional SDP-based algorithm.

Efficient and accurate localization in multihop networks

We present evidence that multihop node-to-anchor distance information is sufficient to allow accurate self-localization in multihop wireless networks (such as ad hoc and sensor networks, as well as future cellular systems based on LTE). To this purpose we have implemented two new distance-based source localization algorithms, which prove highly robust to inaccurate range information characterized by distance estimates exceeding the correct ones. Our contribution is a contrasting alternative to current distributed self-localization algorithms, which are founded on the idea of “diffusing” the known location of a few nodes (anchor) to the entire the network via a typically large number of message exchanges amongst neighbors, resulting in high communications costs, low robustness to mobility, and little (location) privacy to end users. To the best of our knowledge, this work is the first example that the aforementioned disadvantages are not an unavoidable price to be payed for accurate location information in multihop networks.

Super MDS: Source Location from Distance and Angle Information

We consider the simultaneous localization of multiple sources from distance and angle information. An extension of the multidimensional scaling (MDS) technique is given, which allows for both distance and angle information to be processed algebraically (without iteration) and simultaneously. Simulations demonstrate the superiority the super MDS algorithm compared to conventional metric MDS, which relies only on Euclidean distances, and illustrate the impact that angle information may have on the accuracy of source localization. An advantage of the method is that localization under an absolute coordinate system is achievable with knowledge of the coordinates of a single node.

Localization and Tracking for LDR-UWB Systems

Localization and tracking (LT) algorithms for low data rate (LDR) ultra wideband (UWB) systems developed within the Integrated Project PULSERS Phase II are reviewed and compared. In particular, two localization algorithms, designed for static networks with mesh topologies, and one Tracking Algorithm, designed for dynamic network with star topologies are described and/or compared. Each of the localization algorithms adopts a different approach, namely, a centralized non-parametric weighted least squares approach (WLS), and a distributed Bayesian approach that relies on the cooperative maximization of the log-likelihood of range measurements (DMLL). The performance of these two alternatives are compared in a 3D indoor scenario under realistic ranging errors. The tracking algorithm is a fast non-parametric technique based on multidimensional scaling (MDS) and its performance is tested in a dynamic scenario. The proposed algorithms are practical and robust solutions addressing distinct network topologies and/or service requirements related to LDR-LT applications.

WSN06-2: Sensor Localization from WLS Optimization with Closed-form Gradient and Hessian

A non-parametric, low-complexity algorithm for accurate and simultaneous localization of multiple sensors from scarce and imperfect ranging information is proposed. The technique is based on a weighted least-squares (WLS) optimization, where the gradient and Hessian of the quadratic objective are given in closed-form. The performance of the proposed technique is studied through extensive computer simulations, with the intra-node distances randomly generated in accordance to a statistical model constructed from the results of a measurement campaign conducted with a pair of impulsive ultra-wideband (UWB) radios in an indoor scenario. The simulation results reveal that the proposed algorithm, despite its low complexity, is nearly as accurate as the known alternative of best performance, which is based on semi-definite programming and demands significantly more computational power.