Research On Multi-sensor Network Information Fusion State Estimation Based On Data Scheduling

With the development of wireless communication technology and new sensors,multi-sensor information fusion estimation system has been widely applied in military,navigation,target tracking,medical and other fields.The introduction of wireless communication has saved the system cost,increased the flexibility of the system,and brought many benefits to the system.At the same time,new challenges are unavoidable.For example,communication bandwidth constraints,delay,packet loss,and so on.The problem of multi-sensor estimation under limited communication conditions has gradually become a hot spot of research.The main contents of this thesis are as follows:1.For the problem of multi-sensor information fusion estimation under communication constraints.A data transmission scheduling of the descending dimension is proposed to meet the limited communication constraints.Then,the kalman fusion estimation algorithm under the transmission strategy is designed.Finally,the effectiveness of the transmission strategy and the estimation algorithm is verified by the target tracking example of two sensors.2.With the introduction of a large number of sensors into the multi-sensor estimation system,the scale of the system becomes larger and larger.If user continue to use the first transmission scheduling strategy,the amount of information that needs to be transmitted is very large at the same time,more than the communication limits.So,a data scheduling strategy of multi-sensor grouping and periodic transmission is proposed.Under the scheduling strategy,multiple subsystems are formed by grouping,the information of the processing of subsystem is transmitted to satisfy communication restriction condition.Then the optimal estimation algorithm for this kind of data scheduling is given.Finally,a target tracking example of four sensors is used to verify the effectiveness of the transmission strategy and the estimation algorithm.3.In order to satisfy the actual application needs,the controller and actuator are added to the original multi-sensor estimation system so that it can respond to the change of state in time.The sensor samples the physical state during a sampling period,and sends the measured data to controller in a traditional way,the controller estimates the state based on this information,then it sends control command package to actuator,the actuator gets the command package and makes appropriate response.To reduce the deliveryperiod of controller and the energy consumption of system,a data scheduling strategy of communication and collaborative control is proposed.It aims to adjust the sending packet period of controller to actuator in real time,and an optimal solution algorithm of the controller sending package period based on particle swarm optimization(PSO)is designed.Finally,the effectiveness of the proposed data scheduling strategy is verified by an actual heating,ventilation and air conditioning system simulation.