Research On Cooperative Navigation And Location Of Mobile INS/GNSS/UWB Platform Based On Distributed Dynamic Aggregation Network

With the development of economy and society,navigation and positioning services have gradually penetrated into all aspects of people's lives.In particular,the development of “smart transportation and smart city” has further promoted new types of vehicles(such as unmanned vehicles,small drones,intelligent robots,etc.)and miniaturized distributed intelligent devices(such as wearable devices,VR,etc.).development of.These new carrier and instrument technologies put forward higher requirements for low-cost,high-precision navigation and positioning,and the integration of multi-sensor technology and network navigation nodes has significantly improved the navigation positioning accuracy and system anti-interference ability.The INS/UWB/GNSS navigation and positioning platform has three positioning technologies: INS/UWB/GNSS and can build communication and ranging networks based on ultra-wideband communication technology.In the positioning based on the platform,the navigation system runs simultaneously in the combination of multi-sensors and network nodes in two modes.However,the dual-mode navigation system first faces the problem of establishing a filter model that takes into account both navigation modes.Secondly,due to the large number of measurements,the update and dispersion of the navigation and positioning system,the traditional Kalman filter operation time concentration,and the network topology.The sharp increase in cooperative filtering operation leads to problems in node platform computing and communication difficulties.In the distributed filtering algorithm,the federated filtering is an algorithm suitable for the fusion of multi-sensor technology,which is equivalent to the traditional centralized Kalman filter,and the federated filtering decentralizes the traditional Kalman filter operation to the spatially distributed multi-node and each sensor.At the time distribution point of the data update,in the case of the positioning accuracy equivalent to the centralized Kalman filter,the aggregation of the filter update operation is significantly weakened,and the requirement of the navigation operation on the data processing hardware index is reduced.Collaborative Kalman filtering refers to the distributed filtering estimation of the relative and absolute information of the combined neighboring nodes in order to improve the positioning accuracy of the network node platform.In view of the above INS/UWB/GNSS navigation platform,in order to improve navigation and positioning accuracy,and to consider the use of network coordination and multi-sensor fusion,the filtering model design problem,at the same time,the filtering model is complex,the observation dimension is high,and the computational complexity is increased and the time is agglomerated.A distributed cooperative federated filtering model is designed.The distributed cooperative federation filtering of the INS/UWB/GNSS navigation platform in distributed dynamic aggregation is based on ins as the navigation reference,establishes multi-sensor technology combined navigation sub-filter,and uses UWB network coordination as a network collaboration of federated filtering.The observation of the sub-filters,thus establishing a network cooperative federated filtering model,estimating the state quantity of the output platform,and giving the navigation parameters of each platform.The system not only improves the positioning accuracy of single platform INS/UWB/GNSS,but also reduces the aggregation of filtering update calculations,and does not affect the stability of the navigation system due to the central processor anomaly,affecting the entire network.After simulation and actual measurement,the cooperative federated filtering can significantly improve the platform positioning accuracy,and the distributed filtering model can also effectively distribute the filtering update operation time aggregation.In order to improve the navigation and positioning accuracy of the INS/UWB/GNSS navigation platform,considering the use of network collaboration and multi-sensor fusion,the design of filtering model is a problem.The complex filtering model and the high dimension of observation result in the increase of computation and time aggregation.A distributed collaborative federated filtering model is designed.Therefore,distributed collaborative federated filtering in distributed dynamic aggregation INS/UWB/GNSS navigation platform is based on INS as the main navigation reference.A multi-sensor integrated navigation sub-filter is established,and the UWB network collaboration is used as the observation of a network collaborative sub-filter in the federated filtering.Thus,a network collaborative federated filtering model is established to estimate the state of the output platform.It not only improves the positioning accuracy of INS/UWB/GNSS,but also reduces the aggregation of filtering update calculation,and does not affect the whole network due to the abnormal central processor,thus increasing the stability of navigation system.The simulation and experimental results show that the cooperative federated filtering can significantly improve the positioning accuracy of the platform,and the distributed filtering model can also effectively decentralize the filtering update operation time aggregation.