Fetal Heart Rate Monitoring Implemented by Dynamic Adaptation of Transmission Power of a Flexible Ultrasound Transducer Array

Figures

• 

  Figure 1. (a) Process of making a flexible transducer array. (1) Polymethylmethacrylat (PMMA) mold with plate for inserting (2) the piezotransducers. (3) Casting Polydimethylsiloxan (PDMS) on the backside of the transducers. (4) Flipping the first cast to mold a second layer of PDMS on the front side of the transducers (5) to create an air-free layer above the transducer elements. The processing steps are described in detail in Section 2.1; (b) Photograph of the flexible transducer array. For reference, the element number is shown on top of each transducer element.

• 

  Figure 2. (a) Doppler signal acquired with a single element of the flexible transducer array, measured on an in-vitro beating fetal heart setup. In (b), the autocorrelation function is depicted. The maximum found at τ = 0.5 s corresponds to a heart rate of 120 bpm.

• 

  Figure 3. Block diagram of the dynamic apodization adaptation (DAA1, DAA2 and DAA3) process. An initial apodization is used in a first transmit-receive sequence. From the resulting Doppler signals, the power and fHR are estimated and used to update the apodization function, which modulates the transmission pulse of the next transmit-receive sequence.

• 

  Figure 4. Illustration of the working principle of the dynamic apodization adaptation: (a) Received Doppler power measured in the individual transducer elements of the flexible transducer array. (b) For the DAA1-method, the apodization of a single element, which received the highest Doppler power, is set to 1. For the window-based apodization methods DAA2 and DAA3, the power center C defines the center of the window-based methods and the apodization of the elements is set accordingly.

• 

  Figure 5. (a) Influence of the parameter σ (DAA2) on the total transmission power. The dark- and light-colored lines indicate the total transmission power depending on whether the power center C coincides (Con) with an element position or if it lies in between elements (Coff). As reference, the dashed line at 0 dB indicates the total transmission power of a seven-element transducer used in clinical practice and the solid line is the transmission power when only a single element is used; (b) Influence of the parameter D (DAA3), for three selected decay parameters λ; (c) Cumulative apodization, the added apodization values set in all transducer elements, for selected parameter settings. Each block in the stacked horizontal bar plots corresponds to the apodization value of a single transducer element. The light and dark colored blocks correspond to the set apodization depending on whether the power center C coincides (Con) with an element position or if it lies in between elements (Coff); (d) 2-dimensional representation of set apodization function with specific parameters (DAA3, D = 17.7 mm, λ = 0.05 mm−1).

• 

  Figure 6. (a) Schematic of the fetal heart in-vitro setup. A function generator drives an actuator that is connected via a string to a chicken heart. In this way, the chicken heart is moved along the z-direction in a beat-like fashion. By translating the flexible transducer array through the water tank, a changing fetal heart location can be mimicked. (b) Illustration of the transducer array geometry and heart position during experiments.

• 

  Figure 7. Definition of active and passive phase of the Doppler signal for the calculation of the SNR. Each colored line (37 in total) represents the Doppler signal of a single transducer element. The two events visible in the Doppler signal during the active phase correspond to forward and backward motion of the heart relative to the transducer. In the inset of the figure, strong phase variations of the different Doppler signals are visible.

• 

  Figure 8. Effect of the curvature of the flexible transducer array on the received Doppler power in the individual elements. The number next to the transducer element indicates the element index i. The measurements were performed using an in-vitro beating heart setup. The heart was centered in front of the center of the array. (a) Received Doppler power when transmitting with all elements using a flat array; (b) Received Doppler power when transmitting with all elements using a curved array; (c) Received Doppler power when transmitting with element i = 9 only, using a flat array; (d) Received Doppler power when transmitting with element i = 9 only, using a curved array.

• 

  Figure 9. SNR¯ measured during the static (a) and dynamic experiments (b–d), with velocities vm = 2 mm/s, vm = 5 mm/s and vm = 10 mm/s respectively. Both the results using a flat (black) and a curved (white) array are depicted. On the horizontal axis, the different apodization function and, if applicable, parameter settings are indicated.

• 

  Figure 10. Cost function (Equation (7)) relating the measured SNR¯ to the total transmission power. The black and white color corresponds to the flat and curved array, respectively, and the different symbols correspond to the different velocities with which the experiments were performed. For each experiment, the method which has the lowest cost function is highlighted in red.

• 

  Figure 11. Dynamic apodization adaptation and received Doppler power during an in-vitro measurement. The flexible transducer array was curved during the experiment. After an initial time of 6 s, the DAA3 method with width D = 17.7 mm and λ = 0.2 mm−1 was used.

• 

  Figure 12. Illustration of a special behavior of the DAA1-method. In the experiment, the chicken heart is centrally located in front of a curved transducer array and beats at fHR = 120 bpm. During the experiments, the dynamic apodization adaptation method using a single element (DAA1) was used. The top graph shows the set apodization function over time and the bottom graph shows the Doppler signal received with element i = 14 and i = 19. For visualization, the time delay caused when updating the apodization function was removed from the apodization and Doppler signal graph.