A Dynamic Constrained Cooperative Localization Algorithm

Distributed Spatio-Temporal Information Based Cooperative 3D Positioning in GNSS-Denied Environments

“…In Fig. 2 we compare the positioning performance of our EKF-STDF against the particle-based SPAWN [1], STOC-EKF [10], SPA-EKF [11] and NEBP [12], by considering a single agent of interest that may have insufficient spatial ranging measurements from time to time. We set the number of agents to 40 and l max = 30.…”

“…Since we consider a fully distributed network, it is sufficient to analyse the computational complexity imposed on a single agent during one time slot. Specifically, the proposed EKF-STDF has a complexity of O (N rel l max + log 2 (N rel l max )), while each of the particle-based schemes, e.g., the SPA-EKF [11], the particle-based SPAWN [1] and the NEBP [12], has a complexity of O (N rel N s l max + log 2 (N rel N s l max )). Here N rel is the number of neighbors of the agent considered, while N s denotes the number of particles required.…”

“…However, it approximates the nonlinear system model as a linear model around selected linearization points, which inevitably degrades the positioning accuracy. In [11], the authors proposed an SPA-aided CP scheme, dubbed SPA-EKF, which utilized EKF to estimate the velocities of agents as the a priori information. However, it requires a large number of samples to approximate the nonlinear terms in the ranging measurements, hence imposing a high computational complexity.…”

“…ii) We develop a joint prediction and refinement framework based on the integration of EKF and the second-order TP based parametric STDF to estimate the mobility state information of nodes, which is beneficial not only for compensating the lack of spatial ranging measurements caused by sparse distribution of nodes, but also for improving the accuracy of the prior knowledge used by STDF at each time slot. Thus, the proposed EKF-STDF alleviates performance degradation caused by the EKF-based model linearization and is more suitable to distributed CP than the data fusion module of [11] that relies on the original FG-free BP method. iii) Simulation results demonstrate that the proposed EKF-STDF achieves a significantly higher positioning accuracy than the schemes of both [10] and [11] at an even lower computational complexity.…”

“…Thus, the proposed EKF-STDF alleviates performance degradation caused by the EKF-based model linearization and is more suitable to distributed CP than the data fusion module of [11] that relies on the original FG-free BP method. iii) Simulation results demonstrate that the proposed EKF-STDF achieves a significantly higher positioning accuracy than the schemes of both [10] and [11] at an even lower computational complexity.…”