CC BY
19UDK 378 : 611.1
DEVELOPMENT OF THE AUSCULTATION SIMULATION MODEL
Zaidao Mei1'2 Aaron Stib1 Travis Scott1 Dudu Party1 1 - Department of Electrical and Computer Engineering, The Ohio Sate University,
Columbus, USA
2 - Wuhan University of Science & Technology, Wuhan, P. R. China
1. Introduction
For the Veterinary hospital and School of medicine, it is necessary for the trainee doctors , nurse and students to take auscultation practice. At present, real dogs are normally used for the auscultation practice, it is quite clear that using real dogs leads to high cost and management problems. It can be hard to find dogs with some of the more rare pathologies and it is hard to work with live animals when trying to create a productive learning environment for students. Live animals are unpredictable.
Therefore, some veterinary hospitals and colleges of veterinary medicine have demonstrated the need for a model that takes the place of real dogs for auscultation practice for veterinary trainee doctors and students. The model is needed to help increase exposure to rare canine pathologies and allow for greater control of the testing environment while reducing the use of animals in the clinical setting1.
The Purpose of this paper is to outline the design considerations, requirements, specifications, and testing and results for the cardiac auscultation simulation model. The cardiac auscultation simulation model developed will be used to simulate different canine cardiac pathologies that the veterinary students will be able to auscultate, observe, and diagnose.
With the completion of this project, more than 150 students per year will be able to practice these skills in an environment that is much more conducive to learning than previously.
The previous method for testing cardiac auscultation was to use a live dog with the certain pathology, however this comes with many complications and difficulties. This model will be able to simulate such rare pathologies consistently over time, removing the need for an animal each time [1-3] .
Environmental factors can also affect both an animal's behavior and condition and can create disruptive background noise. Using a model instead of a dog allows the user to completely control the testing environment, reducing stress and increasing accuracy of results. This model addresses both of these issues by allowing for any number of pathologies to be simulated and the removal of a live animal in exchange for a dog model.
1 URL: https://cdn-shop.adafruit.com/datasheets/MAX9744.pdf URL: https://cdn-shop.adafruit.com/datasheets/tca9548a.pdf URL: https://github.com/adafruit/Adafruit Python MAX9744
The process used to determine the system design involved studying multiple, volume control methods, microcomputer options, and python coding. The exact details of selection are outlined in the following pages, but decisions were made based on cost effectiveness, ease of use, and availability of supporting resources. The latter factors were the major drivers for the decision to select Raspberry Pi. Raspberry Pi is an open source tool that have large support networks and many example projects available online for our reference whenever we may need it. They are also quite flexible in terms of capability, meaning that we can solve problems using multiple approaches if there are any roadblocks.
Included below are the technical details for the design, as well as the managerial aspects of project execution. The project came in well under budget and should and delivered on time as well, which allowed us plenty of testing before the final delivery.
2. The principle and Design for the Auscultation Simulation Model
According to above analysis, the simulation model system must have requirement specifications as follows:
(1)A user menu interface that allows the following
a. Selection of specific sound files.
b. Random Selection of sound files from a group.
c. Selection of sound files (unknowns) by instructor for assessments.
(2) An administrator menu interface that allows the following:
a. Add new sound files.
b. Manipulate the volume levels (from 0 to 63).
(3)A sound playback and distribution system that allows for sound file delivery (from a single sound file) to six different, closely placed speakers within the canine model, with different volumes.
(4) Approach: give a single sound file (.mp3 or .wma), output the file to 6 outputs where the volume percentage of the original file output can be sent to one location at 100% output (point of maximal intensity) and at lower percentages to the other 3 speakers. The relative percentages will be determined by the clinician investigators after pilot studies are conducted.
(5) Appropriate sized audio speakers for the canine model.
a. The audio speakers should be small enough to fit into the room between two adjacent ribs.
b. The audio speakers need to be covered by durable material like silicon, so that it will not affected by stethoscope pressure or by movement of the model. Based on these specifications , as shown in Fig.1, the function block has been
designed, which includes human-computer interaction, signal processor, multiplexer, audio splitter, auto amplifier & speaker, power supply, etc.
Fig. 2 shows the Software block or flow diagrams, which includes: select file or random play, set volume (large dog, medium dog, or small dog), reveal last file played, pause, etc.
Figure 1 - Functional block diagram
Small Dog [set volume high)
Figure 2 - Software block diagram
3. Experimental
The experimental prototype of auscultation simulation system has been made in this paper, which includes hardware and the software.
3.1 The hardware. The hardware part basically includes three parts: the workstation, the 3D printed cardiac rib cage model, and a stuffed dog.
Fig. 7 shows the overview of our prototype.
Fig. 8 shows the first layer of our workstation design. The red Raspberry Pi pin breakout cable connects the Raspberry Pi which have been placed in the second layer to with operation circuit in the first layer. The I2C multiplexer connects to six different audio amplifiers with the same address so that each of them can be controlled separately. Each of the audio amplifier receives sound files from a 1-to-6 audio splitter. The voltage input of these 6 audio amplifiers are in series and
connected to a 12V voltage source. Each of the audio amplifier is connected to a button speaker as its output.
Fig. 9 shows the second layer of the workstation. The 1-to-6 audio splitter is fixed on the board using a plastic zip tie. The Raspberry Pi is fixed on the PCB board using screws and nuts. The layout makes it easy for the user to plug in their keyboards, mouses, monitors and also to change the SD card if any failure occurs.
Figure 7 - Prototype Overview
Figure 8 - First layer of the workstation
Figure 9 - Second layer of the workstation
Fig. 10 shows the button speaker that we are using for the design.
Apart from its audio output performance, the main reason that we are using it is because of its size. The diameter of this button speaker is just 14mm, which is small enough to fit between two adjacent ribs.
Fig. 11 shows the 3D printed rib cage model which is provided by the veterinary school.
Fig. 12 shows how the speaker will be placed in two adjacent ribs.
Figure 10 - Button Speaker
Figure 11 - 3D printed rib cage
Figure 12 - Speaker between two adjacent ribs
Наука и образование Большого Алтая» выпуск 2 , 2017, страница 51 из
In our prototype a sound filter material is used to cover the rib cage with button speaker embedded. The reasons that we are using this material is as follows:
1) filters some level of the noises;
2) spreads the sound from the button speaker
3) provides user with similar touch feelings to a real dog tissue
3.2 Development of Cardiac simulator software ccording to the software block diagram shown in the Fir.2, we developed the cardiac simulator software system , Fig.15 and Fig.16 show the software interfaces of selecting pathology and random test mode respectively. The random test mode will randomly play one of the pathologies from the library. User can click on "Reveal Test" button to check their judgements' correctness. User can select their audio files through the menu drop down. Also, it is easy to import or remove the audio files through the excel sheet. Customer can remove or import their files by pasting and cutting the name of their audio files into the excel sheet. Below the reveal test button are three buttons that designed for different body types of dogs. The factor of volumes are redesigned for these dogs. The general buttons are on the left, which indicate pause and resume the audio file.
Cardiac Simulator
Figure 15 - Selecting pathology
Figure.16 - Random test mode
4. Di scussion: Challenges And Solution
The research of this paper was met with two main types of challenges: technical and logistical. Technical challenges arose in the amplifier selection stage. The first choice of amplifier presented a challenge because there was no library available to control the output with. This presented the group with an important choice between continuing using those amplifiers or to explore other options. Ultimately, the group decided that the larger amplifiers that did have control libraries available would be a better choice due to the increased reliability and lower chance of bugs in the code compared to what we would have been able to provide by porting C code to python. Despite having to wait for those amplifiers to be delivered, the group was able to press forward on the GUI since we already had the premade control libraries for the new amplifiers. This choice of higher power amplifiers resulted in a more reliable product and a more efficient use of time by the group, since we were
able to start the GUI instead of redoing work that was already available to us. The small financial burden was well worth moving to higher power amplifiers. Other small technical challenges arose throughout project execution that were dealt with using a cost-benefit analysis, keeping customer needs in mind. For our group, this meant prioritizing ease of use and reliability over strictly financial cost.
There were also some minor logistical problems related to procuring items. These came due to unanticipated problems that caused the first Raspberry Pi to overheat, damaging the Pi and memory card beyond repair or use. Additionally, switching to higher power amplifiers caused some delays in our hardware construction. Overall, the group was able to focus on software elements of the project which reduced the overall impact of getting additional materials. In fact, even with changing the amplifiers and getting a new Pi, the project was still completely finished on time and below budget. This was achieved through proper planning to minimize downtime.
5. Conclusion
The project of this paper was highly successful based on the requirements specified by the group as well as the Veterinary doctors and student we worked with. The project was completed within the scheduled time and well below the allotted budget. The success was gauged on the criteria of cost, sound output, and usability. The cost was measured against the $500 benchmark set in the beginning of the course, which the project was well under. Sound output quality was judged by a panel of veterinary doctors and students and was given unanimously positive reviews from them when using the chosen button speakers. The doctors were pleased with the realistic sound observed through the silicone material which will cover the model while in use. Lastly, the students tested the GUI to ensure usability and reliability of the features included. The initial design was agreed upon with the doctors and executed to specification by the team. All of the features work together to create a user friendly experience so that students can focus on auscultation rather than figuring out how to operate the model.
This model will help more than 150 students each year by allowing them to practice auscultation without animals. This model makes it easier to practice testing on rarer pathologies as well, since a file can be created once when an animal comes to the vet school and then reused for testing for years to come. It will also reduce animal stress and exposure in the veterinary school.
The model will be handed over with a parts list and instructional manual for everyday use, troubleshooting, and even construction. This will be useful to the doctors because they wish to create more of these models as well as respiratory models. The proliferation of this project will be made easier by providing printed circuit boards for them to use in the future, as well as detailed plans on how to construct the model. Future cardiac models could also include small servo motors powered by the GPIO pins on the pi to simulate the vibrations that sometimes accompany advanced conditions. This could be implemented by assigning a flag within the excel file that, when present, turns on the servos. Then, if the flag is not
present, there will be no vibration. This idea was beyond the scope of our project but could potentially be implemented in a future model.
The only remaining step in the project is handing the model over to the school, which will take place when winter break is over, as some group members will still be on campus. The handover will consist of the completed model, instructional documentation, group member contact information, and remaining spares that we still have.
Completing this project was rewarding to the team as well as the veterinary doctors and students that were involved in the brainstorming, designing, and implementation of this project. The group learned about hardware selection and configuration, software development, and working in interdisciplinary experience that would never occur inside a typical classroom setting. It was also rewarding to work on a project that will be used for practical purposes here on campus for a long time to come. Overall, this project met and surpassed all set objectives and will serve as a model for future auscultation models built by the doctors of each respective department at the college of veterinary medicine.
References
[1] Robert Hasty. Apps for improving cardiac auscultation. Osteopathic Family Physician, Volume 5, Issue 5, September-October 2013, Pages 208-209
[2] Robyn M.Maude, Joan P. Skinner, Maralyn J.Foureur. Putting intelligent structured intermittent auscultation (ISIA) into practice. Women and Birth, Volume 29, Issue 3, June 2016, Pages 285-292
[3] Yutaka Kagaya, Masao Tabata, Yutaro Arata, Junichi Kameoka, Seiichi Ishii. Variation in effectiveness of a cardiac ausculation training class with a cardiology patient simulator among heart sounds and murmurs. Journal of Cardiology, Volume 70, Issue 2, August 2017, Pages 192-198
