Figure 1: Team
Figure 1.1: Opening screen for the app
Objective: Our team, Thurs-48, aimed to design SPECTRA, a hands-free assistive device to enhance the independence of our client, a 59-year-old woman with Usher’s Syndrome (90% hearing loss and progressive vision impairment due to retinitis pigmentosa), by providing real-time environmental awareness and audio feedback compatible with her Signia hearing aids, within a $100 budget.
Project Journey: We started by researching our clients needs and existing assistive technologies, identifying gaps in hands-free functionality. Our goal was to achieve >95% object detection accuracy and a latency of <0.8s to ensure real-time usability. Early on, we explored multiple concepts, such as a Smart Assistive Vest and a Kitchen Scale with Haptic Feedback, but faced challenges with feasibility and scope. After feedback to focus on specific problems, we pivoted to an iPhone-based solution with a neck strap, leveraging our client’s existing device. Throughout March to April 2025, we iterated through prototyping and testing, addressing issues like poor low-light performance (4.5% error rate in dim settings) by proposing flashlight activation and upgrading our AI model (YOLOv8 to Depth-Anything-V2-Small-hf), which reduced distance errors from 50cm to 5-10cm. The final design featured a silicone neck strap (230g), a voice-controlled iOS app (Spectra), and achieved 95.5% accuracy, though latency (0.85s) slightly missed the target. SPECTRA cost $40, enhancing our client’s ability to navigate and perform tasks like shopping and cooking safely.
Team Contributions: We worked together to design SPECTRA, integrating AI-driven object detection, haptic feedback, and audio output. Our iterative process involved functional analysis, prototyping, and user testing, ensuring the system met our client’s needs while staying within budget.
Technical Skills
Soft Skills
We have worked as a team in designing an app that would be ideal for our targeted client. To thoroughly perform well in this project we did the following:
My primary responsibilities and contributions included:
This project was a pivotal experience that deepened my understanding of human-centered design and the complexities of assistive technology, shaping my growth as an engineer. Developing the AI model for SPECTRA was a huge learning moment, I initially assumed YOLOv8 would perform flawlessly, but testing revealed a 4.5% error rate in low-light and cluttered environments. Through iterative debugging, I adjusted datasets and implemented real-time bounding box adjustments, achieving 95.5% accuracy, which taught me the value of resilience and adaptability in engineering. This process also highlighted the importance of balancing technical precision with user needs, as our latency of 0.85s slightly missed the <0.8s target but remained functional for our client’s daily tasks.
Collaboration was another key lesson. When we pivoted from a chest strap to a neck strap after a comfort rating of 8.7/10 (target ≥9/10), I worked with Dhruv to refine the design, ensuring it met our client’s needs for stability and jacket compatibility. However, I realized I could have contributed more to early-stage discussions, such as in the functional analysis. I focused heavily on AI development, which limited my input on hardware decisions, like the initial waist-mounted belt concept. Reflecting on this, I see the need to balance technical depth with broader project involvement. In future projects, I’ll ask questions like “How can my work better integrate with other components?” to ensure a more integrated contribution from the start.
SPECTRA also opened my eyes to the societal impact of engineering. Designing for someone with dual sensory impairments made me realize how technology can empower independence, directly affecting our client’s quality of life, whether navigating a grocery store or cooking safely. This experience sparked my interest in the ethical dimensions of assistive technology, such as ensuring privacy with iPhone camera usage, an area we identified but didn’t fully address. It also taught me the importance of scalability in design, a principle I’ll apply as I pursue a career in AI and accessibility. For example, I now understand that assistive devices must adapt to diverse environments and user needs, a lesson I’ll carry into future projects. Looking back, I could have managed my time better to engage more with the team, especially during user testing, where my input could have helped us reach the ≥9/10 user experience target. If I could do this project again, I’d schedule regular check-ins to align my AI work with the team’s broader goals. This experience has redefined what it means to be an engineer for me, it’s about technical innovation, empathy, and teamwork, lessons that will guide my future work in creating inclusive technologies.