Decentralized TOA-based localization in non-synchronized wireless networks with partial, asymmetric connectivity

Classical methods for cooperative time-of-arrival (TOA)-based self-localization in wireless sensors networks usually rely either on synchronization between sensors or on two-way TOA (TW-TOA) measurements. However, when the connectivity in the network is limited, accurate synchronization of the nodes may be too costly to attain, and full TW-TOA measurements may not be available either, especially when the connectivity matrix is asymmetric. Centralized processing for the location estimation is also rarely feasible under such conditions. In this paper we propose an algorithm (method and protocol) for decentralized, cooperative TOA-based self-localization, which does not require synchronization and does not involve any 'hand-shaking' procedures (of the kind required for TW-TOA measurements) between sensors. The procedure involves a single initial transmission by each sensor, followed by reports of the received TOAs by each sensor, followed by an iterative procedure in which each sensor estimates its own timing offset and position, as well as the timing-offsets and positions of relevant neighboring sensors, and announces its estimation results. We demonstrate convergence of the procedure to accurate estimates of the position of nearly all sensors, e.g., in scenarios involving 50 - 200 sensors in an area of 20 by 20 Km with an average 1/3 connectivity.