Rigorous Bayesian Computational Sensor Networks are developed to quantify uncertainty in 1 model-based state estimates incorporating sensor data, 2 model parameters, 3 sensor node model parameter values e.g., location, noise, and 4 input sources e.g., cracks holes. These decentralized methods have low computational complexity and perform Bayesian estimation in general distributed measurement systems i.e., sensor networks. A model of the dynamic behavior and distribution of the underlying physical phenomenon is used to obtain a continuous form from the discrete time and space samples provided by a sensor network. This approach was applied to the aircraft structural health monitoring problem. Structural health monitoring SHM deals with evaluating structures for changes in their characteristics, predicting useful lifetime without maintenance, and recommending maintenance strategies to increase lifetime and reduce downtime. Current aircraft construction often involves fiber-reinforced laminated composite materials which offer certain advantages, but can suffer internal damage with little external evidence. We developed specific Bayesian computational models of SHM transducers e.g., ultrasound acting in both undamaged and damages materials.