Application of distributed motion estimation for swarm MAVs in a GPS-restricted environment based on a wireless sensor network

To meet the requirements for motion estimation of a swarm of micro-aerial vehicles in a GPS-restricted environment, a distributed motion estimation algorithm is proposed that combines an inertial measurement unit, magnetometer, wireless ranging network and altimeter data. Based on the information sharing and mutual positioning ability of the wireless sensor network, the long-term stability of motion estimation in a GPS-restricted environment is improved on the basis of information from neighboring vehicles. A centralized information filter for large-scale swarm micro-aerial vehicles is proposed that uses the information distribution principle. To address the difficulty of high-dimensional state inversion in centralized information filtering, the theory of distributed information filters for large-scale systems is adopted to decompose the high-dimensional optimal centralized information filter into distributed sub-filters. Based on the logged data, the simulation results show that the accuracy of motion estimation using the distributed filtering method is not significantly lower than with centralized filtering, and is more suitable for distributed deployment on micro-aerial vehicles, enabling stable motion estimation to be maintained in a GPS-restricted environment.