Search for: All records

In this paper, we present a compact calibrated radiometer for measuring internal body temperature. The radiometer architecture combines the benefits of a correlation and a Dicke radiometer. The sensitivity to load impedance variations is reduced through a balanced topology, while the pre-detection switch reduces sensitivity to gain fluctuations of all components in the receive chain that follows. The radiometer is designed to operate in the 1.4-1.427 GHz quiet band using off-the-shelf components on a 10cm×7.6cm printed circuit board. Two types of temperature estimation methods are compared and the errors analyzed using resistors at known controllable temperatures at both input ports. One of the ports is then connected to a near-field antenna probe matched to tissue layers of the cheek. When water of variable temperature is in the mouth, an independent thermocouple verifies the radiometric temperature of the water measured by the probe placed on the skin. 

This paper addresses microwave radiometry for passive non-invasive subcutaneous temperature measurements at a few centimeter depth in tissues. A correlation radiometer is designed in the quiet 1.4-GHz band and tested on aqueous phantoms. The radiometer is designed from off-the-shelf components and first tested with a matched load, and then with a near-field planar compact probe antenna, both with two temperature-controlled water phantoms of different volumes. The measurement resolution, sensitivity and long-term stability is quantified in terms of integration time for a simple three-point calibration. The lowest measured absolute error compared to a ground-truth thermocouple measurement is 0:25K over one hour of data collection with a single calibration. Measurements show that an integration time of > 1 s results in an absolute error limited by the radiometer gain fluctuations.