Passive wireless sensor based on reflected electro-material signatures

The objective of the proposed research is to devise a methodology for sensing and tracking environmental variables using a passive wireless sensor based on reflected electro-material signatures. Viability of item level tracking demands the sensor to be extraordinary low cost, thus eliminating the use of any active sensor or memory circuitry. Recent developments of materials whose electrical properties can change significantly with the environmental conditions suggest the possibility of developing a passive sensor that can be interrogated remotely to extract data about the time tracked environmental changes at the sensor. A simple passive sensor, based on the concept of reflected electro-material signatures (REMS), consists of an antenna attached to a microstrip transmission line, which in turn is routed over one or more sections of variable permittivity material before being terminated in a load. The basic idea revolves around sensing the electrical properties of thermotropic liquid crystal (LCs) trapped in a polymer substrate to record the temperature data. As the temperature changes with time, the polymerization process through the material line records the historical temperature profile in the spatial distribution of the electrical properties, thus enabling the system to extract the historical profile of temperature without using any active memory circuitry. This concept can possibly be used to track a variety of variables of interest; however, the proposed research is focused on sensing and extracting the time profile of temperature. The problem of identifying medium properties from waves reflected from a device of this type is a form of the classical one dimensional inverse scattering problem. For profile inversion in a lossy inhomogeneous media, analytical techniques are difficult to implement in most practical situations. In the proposed research, neural networks with a back-propagation algorithm are used to reconstruct the historical temperature profile of the material by extracting the spatially distributed material properties of the electro-material line. After the initial proof of concept for a lossless medium, the methodology is extended to extract spatially distributed properties for a dissipative medium. Finally, for the implementation of REMS sensor concept, a neural network based methodology is developed to reconstruct the spatially distributed permittivity profile of a lossy electro-material line.