An EKF based overlapping coalition formation game for cooperative wireless network navigation

The problem of efficient link selections for cooperative navigation is investigated in dynamic wireless networks, using an extended Kalman filter based overlapping coalition formation (OCF) game approach. Due to the densely deployment of nodes without proper link selections, the conventional cooperative navigation approaches usually yield high time-frequency resources as the existing resource management methods only work properly in static localisation systems. A new link selection mechanism which is reliable with limited resources in dynamic positioning scenarios is proposed. The varying topological structure and states of every node will be taken into consideration while each node abandons a considerable number of unnecessary links. Firstly, the state of every mobile node will be estimated in real-time. Then a selection criterion is applied to determine whether nodes are well navigated. If the criterion is satisfied, an OCF game will be employed to construct the effective link selection. Otherwise, a re-link mechanism will be used to obtain precise state information of each node. Numerical results show that the cooperative navigation algorithm can significantly improve the positioning accuracy and the proposed OCF game approach can build a cooperative network with limited number of links at the cost of appropriate performance degradation.