Current period tracking devices and methods pose several challenges. Firstly, existing apps and wearable devices are unreliable and inaccurate in predicting menstrual cycles, especially irregular cycle patterns. Many respondents in a 2022 survey reported that the apps they used were not sufficiently accurate to be helpful (Broad et al., 2022). This negates the purpose of the app, and results in the users being caught off-guard. Unreliable prediction of one’s period can hamper daily activities, medication, and workout schedules, pose barriers to accessing hygiene products, and disrupt the user from maintaining healthy habits (Chyu et al., 2022). Being able to track periods is important when scheduling one’s life, but for people with a cognitive disability, this can be extremely difficult to do on their own as apps and other tracking methods require a lot of user input. Physical disabilities will cause many irregularities in menstrual cycles, and most tracking methods do not apply to irregular cyhcles. A new, improved prediction method is necessary to meet user needs.

People with irregular periods do not have any accurate options for period tracking. They cannot use physical tracking methods, like calendars, because those methods rely on counting to 28 (or whatever regular cycle length someone has) on a loop. This is a problem with many apps as well. Irregular periods and variations in what are considered regular periods are very common and can be caused by a variety of different factors. Currently, people with irregular periods have very few options to track their periods, most of which are inaccurate.

Period tracking apps have become more controversial following the Supreme Court of the United States ruling to overturn Roe v. Wade, which had, before being overturned, granted the right to an abortion. Now that abortion rights are not protected at the federal level, many states are imposing harsh abortion bans, including jail time for people receiving or performing an abortion. Most abortions are performed within twelve weeks of the patient’s last period, which means that someone may miss three periods before getting an abortion. After getting an abortion, it typically takes four to eight weeks for someone’s period to return. From the last period before the abortion to the first period after, the app user has missed four to five periods (Planned Parenthood, 2020). An outside party buying the app’s data may be able to see that gap, which is not long enough to be a pregnancy, but likely too long to be natural, and determine that someone had an abortion. In a state that doesn’t allow abortion, the White House has warned that information from tracking apps can be used to prosecute someone on trial for abortion and recommended against using an online tracker (Vazquez, 2022).

The reasons laid out above were the basis for our motivation to pursue this project, and our desire to further the study of women’s medicine and give people with disabilities or irregular periods more control over their lives were what further influenced us to continue with this project.

When designing our initial prototype, we all had different roles that we fulfilled. Cecilia and I did the most work on the actual design, and Palak and Keira helped support us for both the testing of the different components, and with figuring out the stuff that Cecilia and I were not initially able to get.

For the design, we used an Arduino UNO that was connected to three vital sign sensors that measure skin conductivity, heart rate, and temperature. All of these sensors were attached to a silicon bracelet. The Arduino UNO has an SD card that can store data from these sensors, and the averages of the readings are sent to a Raspberry Pi via bidirectional serial communication. We have a machine learning model that is deployed to the Raspberry Pi, and it then executes the Decision Tree model that we developed to predict days until the next cycle start date, which is then sent back to the Arduino Uno and displayed on an LED screen.

I worked mainly on the Arduino UNO, the sensors, and the code for the Arduino, with some help from Cecilia. Cecilia worked on the Raspberry Pi, and then once we were both done with our sections, we combined the two to create our final device. Palak was the one who developed the machine learning model, and Keira and Palak together helped me test the sensors.

The final prototype consists of three vital sign sensors, a Raspberry Pi, an Arduino UNO, an LED screen, and bidirectional communication established between all these parts. This device can measure heart rate, wrist skin temperature, and skin conductivity with high sensitivity and specificity. The vital sign readings are recorded by the Arduino UNO setup and sent to the Raspberry Pi, which performs the algorithm and returns days until the next predicted cycle. This device aims to make menstrual health tracking easier, accessible, and safe for women of all backgrounds.