Research On Sensor Management System Based On Efficiency Evaluation

Multi-sensor management is to allocate sensor resources reasonably,which according to different optimization criteria to meet the sensing requirements of detection.In sensor detection system,real-time scheduling of sensor resources are often limited,at the same time,due to the sensor detection ability,different limit risk and task requirements,make for different scenarios of sensor management technology research has the vital significance,has become one of the problem to be solved in the field of scientific research workers.In this paper,based on multi-sensor collaborative tracking target as the research background,based on the theory of sensor management,the Bayesian theory,filtering estimation,weighted fusion,threat assessment and so on are used as the technical means.The efficiency evaluation functions under the criteria of tracking accuracy,risk control and information incremental optimization are designed based on task requirements,and improve the data fusion method under different scenarios,objective risk assessment criteria and sensor control strategy.In order to realize the reasonable distribution of sensor limited resources,improve the sensor to the target tracking performance.The main research contents of this paper are as follows:Aiming at the problem of target tracking discontinuity caused by sensor switching in the target detection mission,this paper proposes a sensor management algorithm based on prediction criterion.Firstly,an efficiency function based on target tracking accuracy is constructed.Secondly,combined with the constructed efficiency function and the target state prediction variance,the prediction criterion for determining whether the heterogeneous sensors are switched is given.Finally,based on the prediction results,a changed weighted extended Kalman filter(CWEKF)algorithm is proposed to improve the measurement fusion accuracy of heterogeneous sensors in the switching phase.Simulation and analysis show that the proposed algorithm is feasible and effective,which can reduce the measurement jump error caused by the switch of heterogeneous sensors effectively,and provide more reliable measurement information for the subsequent estimation of target state.Aiming at the problem of unreasonable sensor allocation due to the uncertainty of risk assessment in the target detection task,which causes a large potential risk to the party concerned,a sensor management algorithm for risk control is proposed.Firstly,by improving the threat assessment algorithm,the dynamic fuzzy Bayesian network(DFBN)method is proposed to evaluate the target threat risk.Secondly,the target prediction covariance is used to evaluate the tracking risk of the sensor.Finally,the efficiency evaluation function is constructed by combining the tracking risk and threat risk,and the sensor resources are allocated reasonably with the minimum risk as the optimization criterion.Simulation and analysis show that the proposed algorithm reduces the potential risk of sensor control strategy to target measurement effectively,and improves the target tracking performance.Aiming at the problem of the sensor measurement estimation error is large due to the change of the number of multiple targets in target detection task,a sensor management algorithm based on the Rényi information increment is proposed.Firstly,the multi-objective state process and measurement process are modeled according to the Random Finite Set(RFS).Secondly,according to the sequential Monte Carlo recursive process of multi-target multi-Bernoulli filtering,the target probability density function is solved.Finally,the efficiency function was constructed according to the Rényi information increment,and based on the optimization criterion of maximizing the increment of information,the efficiency function was solved and the optimal control sequence of the sensor was calculated.Simulation and analysis show that the proposed algorithm improves the rationality of multi-sensor allocation for multi-target detection,and improves the accuracy of multi-target quantity and state estimation.Aiming at the problem of sensor management performance verification,a set of sensor management system simulation platform is designed in this paper.Firstly,the simulation requirement analysis of sensor management system is given.Secondly,several functional modules including data preprocessing,data fusion,data analysis and data display are designed according to specific requirements.Finally,Visual C++ language is used to realize the User Interface(UI)of the system simulation platform,and the functions of different modules on the simulation platform interface are tested and analyzed.The system simulation platform provides effective verification and evaluation support for the engineering application of sensor management.