Model-based Reconstruction and Classification Based on Near-Field Microwave Measurements

Research Project, 2011 – 2014

The rapid development of wireless communication technology implies the availability of cost-efficient microwave components and processors. This has already resulted in safer cars that employ microwave radar sensors to detect obstacles and other vehicles. Our vision is that the same technology can be used for cost-efficient diagnosis in medical and process industry applications. The goal of the proposed multi-project is to pave the ground for such a development. From other on-going and planned projects, we have experience from detection of cancer and stroke, analysis and diagnosis of a pharmaceutical manufacturing process, and wood quality assessment. The proposed research program will support these application-oriented projects with theory and methods from more general viewpoints. There are several challenges when using microwave sensing (radar) in medical and process engineering applications. The subject is situated in the near-field of the sensor, and there is a need for resolution in both time and space. In many cases, very high bandwidths are required to provide the necessary resolution and penetration. Thus, design of Ultra Wideband (UWB) antennas is a crucial part. Further, the complicated electromagnetic environment calls for computationally demanding reconstruction algorithms, as well as advanced estimation and classification techniques. The proposed multi-project is an inter-disciplinary effort to address all of these issues in a coherent research program.