A collaborative navigation method based on a delay-aware factor graph

Abstract Aiming at the problem of degradation in navigation accuracy caused by communication delay during cooperative interaction of unmanned aerial vehicle (UAV) formations, a distributed cooperative navigation algorithm based on a delay-aware factor graph is proposed. First, a factor graph model is constructed based on the measurement equations of each sensor. Subsequently, considering the mechanism of ultra-wideband (UWB) transmission delay, a delay factor is formulated and introduced into the factor graph to enable joint estimation with the UAV state. During the initialisation of the factor graph, a prior delay estimate is introduced as an initialisation factor. In order to reduce computational redundancy, a sliding window strategy combined with marginalisation is employed. Furthermore, an adaptive momentum accelerated gradient descent (AMAGD) algorithm is designed to optimise the step size for delay updates. Finally, the global state optimal solution of the delay factor graph is obtained through a message iteration algorithm. The simulation results show that the proposed method can effectively compensate for the loss of accuracy caused by communication delay and improve the positioning accuracy.