2021-: System Support to Enable AI at the Edge (Shiv Mishra)

Edge computing introduces middle-tier compute servers much closer to the sensors and end users to build context aware IoT and mobile applications, thereby minimizing latency and preserving privacy. Applications include recognizing and monitoring people in work and public spaces, smart management of city operations including water and energy distribution, smart agriculture, managing traffic and transportation systems, law enforcement, and augmented reality applications. These applications require distributed execution of complex, compute-intensive algorithms such as deep learning, gesture recognition, acoustic detection and classification, and anomaly detection. The main goal of our research project is to develop core system-level services to enable a distributed, microservice-based edge system architecture that facilitates building such complex applications. The key features of this system include incorporating humans in the loop, optimized placement of compute and data elements in a dynamically changing environment, and computing over a diverse set of processing elements including traditional CPUs and ASICs such as GPUs and FPGAs. The proposed system aims to integrate and augment elements of current edge solutions such as EdgeX, Kura, OpenEdge and Azure IoT Edge.