| dc.description.abstract | The growing complexity of human-centric and ecological systems demands new sensing technologies capable of capturing holistic, contextual insights in real-world environments. However, a critical gap exists in the availability of integrated, intelligent platforms that can be adapted across these diverse domains. This dissertation addresses this challenge by designing, engineering, and validating a series of novel, multi-modal sensing platforms to provide a new and more insightful lens on a broad range of life.
The research spans two primary contexts. First, to explore the nuances of human well-being, the AirSpecs smart-eyeglass platform was developed and deployed. This system holistically measures an individual's proximate environment and physiological parameters, and was validated in a multi-site international study to investigate the dynamics of human comfort "in-the-wild". Second, to advance ecological monitoring, a progression of acoustic platforms was engineered. The SoundSHROOM system was created as a robust, multi-channel recorder for harsh environments and successfully deployed in the Arctic. Building on this, the BuzzCam system was developed for targeted pollinator monitoring, culminating in an end-to-end pipeline for on-device AI classification of endangered and invasive bee species in Patagonia. Finally, the CollarID platform was engineered and characterized as a versatile, low-power, multi-modal animal-borne sensor for wildlife tracking, integrating inertial, bioacoustic, and comprehensive environmental sensing to move beyond the limitations of location-only devices.
Key contributions of this work include the validated hardware platforms themselves; several unique, publicly available datasets from urban, Arctic, and Patagonian deployments; and a demonstrated methodology for implementing on-device AI to address the data-to-insight bottleneck in ecological monitoring. Collectively, this research provides the scientific community with a new suite of powerful research tools and demonstrates a cross-contextual design philosophy, leveraging engineering principles across disparate fields to enable a deeper understanding of organisms and their complex interactions with their environments. | |
