Chipping Away at Challenges: Drafting a Roadmap

As part of C2S2's mission, we aim to design a proof-of-concept chip to address a current challenge. Since last spring, C2S2 has been partnered with a Ph.D. candidate at the Cornell Lab of Ornithology. From meeting with our Campus Partner, we learned more about his main challenges while using tags for his research. The primary concern has been the battery life of his current tags, as seen below, which last roughly twenty-four hours and decays after each deployment. Due to short battery life, he frequently has to recapture the Scrub Jays to collect data stored on the tag and to retag birds to collect more data. This retagging process may introduce biases to data collected as the luring process and human interaction may alter Scrub Jaysl natural behavior. C2S2 is focused on designing a low-power IoT chip that can extend the deployment period of a tag. By extending the deployment period of a tag, the team will aid in mitigating human interaction with birds during research studies, thus reducing human influence on data collected and reducing stress on birds being tagged.

To begin the development of a proof-of-concept chip, we've been working with other members of Project Management and members of System Architecture to draft a roadmap to keep the team informed about tasks and address any questions they might have regarding our campus partner's main challenges. The primary focus for our subteams revolves around two main projects: creating a more battery-efficient design and incorporating audio sample collection alongside acceleration data. To begin the road mapping process, the Project Management subteam first needed to develop a stronger understanding of the technical work of subteams. Although intended to inform alumni and sponsors of the team's progress, we started to build our understanding and documented what we learned of more technical aspects of projects for roadmapping by working on the team's newsletter.

While working on the newsletter, members of Project Management frequently sat in work sessions and meetings of other subteams and spent more time reaching out to more technical subteams to learn more about what they previously were able to accomplish as well as their current progress and goals for the semester. In our newsletter, we began to gather information from each subteam about their projects from last year. Based on our understanding of what the subteam had worked on as well as official goals to accomplish for the campus partner, we helped create a flow chart, as seen below, to determine how each subteam should connect within the design of the chip.

To create the flow chart, we started by initially thinking of how to address the main challenge and need of our campus partner to collect audio recordings and acceleration data. These prompted the first inputs to be decided as both a microphone to collect audio samples and an inertial measurement unit (IMU) to measure acceleration. The data collected from the IMU would be input into the Software subteam's position estimator, an algorithm designed during the spring semester, to determine the relative position of the tag. This relative position data would need to be offloaded so that a researcher would be able to relocate the tag, however, the final project of what offloading was left open-ended until the Project Management subteam could learn more about feasible and low power methods that could be used.

Following the input of the microphone, we planned for the Analog subteam to work on an amplifier to amplify the sound signal as last semester they taped out a two-stage operational amplifier with Miller compensation. This semester the Analog subteam has been working on an Analog-to-Digitial Converter (ADC), which we included in our roadmap after the amplifier. As the name suggests, this ADC would convert the analog signal to a digital signal, this digital signal would then be used in the Digital subteam's Fast Fourier Transform (FFT) to convert sound signals from the time domain to the frequency domain, thus allowing for spectral analysis of different frequency components. From the FFT, we move towards a filter, also to be designed by the Digital Subteam, which would be used to isolate specific frequency bands or remove unwanted noise. This filtered sound data, would then be stored in Flash Memory on the tag for a researcher to save and use for additional data processing.

To address one of the main challenges of extending the deployment period of the tag, we also included CLK, an internal clock signal, as an input. This internal clock signal, with software developed by the Software subteam, would aid in providing synchronization of sound and acceleration data, as it would allow the researcher to relate both the sound and acceleration to a certain time. Time indication would also be used as an input into software, which would be developed by the software subteam, for scheduling when to collect samples. This scheduling would aid in extending deployment periods of tags, as currently once the tag turns on, it remains on and continuously collects data until it runs out of battery. From our Campus Partner, we learned that it would be ideal for the tag to be scheduled to collect samples approximately four hours after sunrise, as this is when Scrub Jays are most active. We're also considering a solution by introducing a "sleep mode" for the chip during the birds' inactive periods, aiming to extend battery life and prolong the time before tag replacement is necessary. Additionally, our partner has requested the tag to capture audio samples from Scrub Jays, requiring the integration of a microphone and accelerometer to record both sound and bird accelerations over time. The time when scheduled data was collected would similarly need to be stored in flash memory to relate sound and acceleration data collected to time and offloaded to assist the research in relocating the tag.

After forming this general draft of the roadmap, Abigail was able to help present the roadmap with members of System Architecture during the Mid-Semester Design Reviews to get feedback from the team. Following our recent Mid-Semester Design Reviews, we discovered misconceptions about the challenges of our Campus Partner. To address these misconceptions and ensure that everyone on the team had access to research collected by the Project Management subteam from conversations with our Campus Partner, Abigail crafted a document to tackle common questions and provide a more comprehensive understanding of our campus partner's needs. We're introducing a dedicated Slack channel for roadmapping discussions, to not only facilitate streamlined communication but also so that other subteams and members can have access to information shared by Project Management to answer a question that others may also have.

Similarly, from feedback from the team and learning more recent updates on projects from other subteams, we are currently working on updating the roadmap. Some updates we are working on updating our roadmap to reflect are alterations in subteam tasks, such as including Software's work on spectrography and Digital's work on creating a classifier, and including a Particle Photon board for offloading data. Additionally, we will continue to expand on this current draft of the roadmap to account for more complexity of how some projects may have the same inputs or rely on a different project's output by adding in levels to organize the roadmap into phases. Furthermore, we will be creating edits as we continue to meet both with subteams to learn more about their progress and our Campus Partner to learn more specific details regarding Scrub Jays' behavior and their vocal calls. Although our roadmap is still a work in progress, our roadmap update plans aim to enhance clarity, promote efficient communication, and ensure everyone is aligned as we move forward.