A Parametric TDoA Technique in the IoT Localization Context

Internet of Things (IoT) has been scaling up over the last few years in multiple applications and due to the need for geolocation and tracking capabilities, the usage of traditional Time Difference of Arrival (TDOA) arises. In this paper, a novel methodology for localizing using TDoA is presented, after the detailed and complete description of the TDoA has been provided. This proposed method depends on the hyperbolic functions to localize the node on a hyperbola, rather than locating it in a free position in the space potentially suffering from the influence of the timestamp imperfections. Thus, the proposed approach is finding this location on a hyperbola at a point which has the minimum Euclidean distance to all the other hyperbolas. A comparison is performed investigating the attainable accuracies for localizing based on this parametric TDoA and the classical TDoA method, on a well-defined simulation environment. The simulator is based on a Poisson distribution approach for defining the gateways and the node topology, as well as a noise model for emulating the oscillator drift at the gateways. In the given results, the feasibility of the proposed technique is asserted by a drastic improvement over a wide range of drift variances and the number of gateways. This manifests the robustness of the contributed method to the outlier timestamps and its optimum rendering, especially when the number of gateways is expected to be increased in the future.