本文提出了一种基于自联想神经网络的传感器解析余度技术。在这种网络中，冗余传感器的信息被压缩、重组进入网络的第一部分，网络的第二部分将压缩信息恢复出来。基于数据融合原理，若一个传感器发生故障，其它传感器仍可提供足够的信息代替发生故障的传感器。理论分析和用于涡轴发动机的仿真结果表明，这种特殊结构的自联想网络具有良好的过滤噪声和故障信号的作用，特别适合于用作不易建模的复杂对象的传感器信号重构 This paper presents a sensor self-learning neural network redundancy algorithm. In this network, redundant sensor information is compressed, reorganized into the first part of the network, and the second part of the network recovers the compressed information. Based on the data fusion principle, if one sensor fails, other sensors can still provide enough information to replace the failed sensor. The theoretical analysis and the simulation results for the turboshaft engine show that the self-associative network with this special structure has a good function of filtering noise and fault signals and is particularly suitable for the reconstruction of sensor signals used as a complex object that is not easy to model