Undergrad research projects of Shovito
Academic projects formulated, developed, and demonstrated by Shovito during his undergrad and graduate studies.
Designing active learning strategies that collect labeled sensor data in uncontrolled environments is challenging because (i) without taking into account user's cognitive factors, active learning places much burden on the user and lowers adoption of the technology; and (ii) the labels expressed by end-users exhibit significant amounts of temporal and spacial variations leading to poor performance of the learned models. In this project, we design, develop, and validate human-in-the-loop technologies that collect sensor data in-the-wild and develop algorithms and tools (i) to maximize the active learning performance taking informativeness of sensor data, burden of data labeling on user, and reliability of prospective labels into account; and (ii) for understanding and inferring complex health events using wearable and mobile sensors.
We develop novel approaches for reliable gait monitoring and investigate applications of wearable-base monitoring in various populations. The utility of wearable sensors for continuous gait monitoring has grown substantially, enabling novel applications on mobility assessment in healthcare. Existing approaches for gait monitoring rely on predefined or experimentally tuned platform parameters and are often platform-specific, parameter-sensitive, and unreliable in noisy environments. To address these challenges, we investigate platform-agnostic and reconfigurable computational approaches to gait monitoring, step counting, mobility assessment, and related problems.