Kalman Filters With Adaptive Strategy For Flocking Of Multi-agents

In nature,collective characteristics is obviously observed when bees went to take honey or return back to honeycomb.The bees will modify their movement path according to the information of the adjacent bees or the information of the leading bees,therefore,the bees will reach one certain speed and direction before they land to the determine.The phenomenon is called flocking.Flocking of networked system have been proposed with this phenomenon.Complex network information interaction,topology and other aspects are studying in flocking of networked system,and this kind of system gains great enthusiasm.Deep research is moving forwad in optimization with flocking of networked system.Adaptive Kalman filter is studied in this paper,flocking of networked system has well response to changing surroundings,and lower reliability to global structure,higher adaptivity,and stability is realized.Firstly,general model of flocking of networked system is put forward,structure and topology of the model is characterized.Key parameters of the model are described.The connectivity is also mentioned.At second,an adaptive strategy is combined with general Kalman filter model,which can be applied in observation of sensor network,the combination consists of adaptive Kalman filter.One node in adaptive Kalman filter iterates the observation without adjacent nodes,relies less to global structure,comparation of these two Kalman filters shows more advantages to adaptive Kalman filter.The simulation confirms the validity of the results.Than,adaptive Kalman filter is applied to flocking of networked system.The information exchange between the sensor networks does not need more global information.A sensor only needs to obtain the adjacent sensors' information to modify its own movement path.The final network can be achieved in the same speed,the process of moving,and the network can maintain stability.The simulation confirms the validity of the results.Than,adaptive Kalman filter is also applied to flocking of networked system with pinning control strategy.Parts of sensors are informed in sensor networks,meanwhile,other sensors modify their movement path and position with information from adjacent sensors.The while network can reach into the same speed,and the process of moving the network can stay stability.The simulation confirms the validity of the results.Finally,this paper summarizes and prospects the application of adaptive Kalman filter into flocking of networked system.The study could be applied to many aspects with practicality.