As is required for technological systems, e.g. airplanes, the service life should be maximised to fulfil the criteria for high efficiency in terms of costs and sustainability. One option to achieve this goal is to monitor the current health state of the system and derive measures to maintain the reliability of the system. The state-of-the-art for thin-walled structures is a structural health monitoring system in which actuators induce a guided ultrasonic wave or wave-field measured by a sensor or a network of sensors. The sensor measurements provide information that can be used to determine the respective system’s health state. Actuators and sensors may be embedded into the system for a maintenance-friendly application. However, embedding the actuator and sensors in most cases leads to a distortion of the wave-field. The distortion, or rather the reflections of the wave-field, can lead to over- or underestimation of the quantity, location, and severity of the damage. These false-detections may lead to both an unnecessary high effort for retrofitting and a reduction of the technical reliability of the structure. To avoid these false-detections, the focus of the present contribution is on reducing the distortion/reflections of the wave-field caused by the sensors. The reduced distortion of the wave-field is achieved using a functionally graded material and an acoustic impedance matching based on a mechanical model. Therefore, tungsten particles were added to the uncured epoxy resin to adjust and control the acoustic impedance. This novel approach allows for the design of an interphase between the sensor or its glass housing and the surrounding structure. The interphase properties are controlled by the tungsten particle content which varies with the radial distance to the sensor. In this work, several models regarding the radial distribution of the tungsten particle content, the excitation frequency and the number of interphase-steps are investigated. The approach results in (i) reduced reflections from the sensor (less distortion) and, for specific configurations, (ii) amplified measuring signals of the sensor. A number of numerical examples show the applicability of the presented method.