CAJ | 학술논문

为了降低水下潜器组合导航模型中传感器信息的不确定性，提出了一种基于证据理论的多传感器信息融合方法。在该算法中，将导航传感器信息数据输入到ELMAN网络来获取D-S证据理论中各传感器故障的基本概率赋值，再利用改进后的证据理论对导航系统的信息进行决策级融合。仿真结果表明，导航传感器信息的不确定性降低到0.001，可以精确检测到故障传感器，有效提高了系统信息的可靠性和对传感器故障的识别能力，为水下潜器精确导航提供了保障。
위료강저수하잠기조합도항모형중전감기신식적불학정성，제출료일충기우증거이론적다전감기신식융합방법。재해산법중，장도항전감기신식수거수입도ELMAN망락래획취D-S증거이론중각전감기고장적기본개솔부치，재이용개진후적증거이론대도항계통적신식진행결책급융합。방진결과표명，도항전감기신식적불학정성강저도0.001，가이정학검측도고장전감기，유효제고료계통신식적가고성화대전감기고장적식별능력，위수하잠기정학도항제공료보장。
In order to reduce information uncertainty in integrated navigation model for underwater vehicle, a multi-sensor information fusion algorithm based on evidence theory is presented. In the algorithm, the navigation sensor measured data were input into ELMAN network and used to obtain basic probability assignment for sensor fault in D-S evidence theory. And then the improved evidence theory was used to execute the decision and fusion for navigation information. The simulation results prove that the uncertainty of navigation sensor information is reduced to 0.001, and the fault sensor can be detected. The algorithm can effectively improve the reliability of navigation system information and the recognition of sensor fault, which provides the guarantee for the precise navigation of underwater vehicle.