BEGIN:VCALENDAR VERSION:2.0 PRODID:-//wp-events-plugin.com//7.2.3.1//EN TZID:Asia/Kolkata X-WR-TIMEZONE:Asia/Kolkata BEGIN:VEVENT UID:155@cds.iisc.ac.in DTSTART;TZID=Asia/Kolkata:20251119T100000 DTEND;TZID=Asia/Kolkata:20251119T110000 DTSTAMP:20251104T085141Z URL:https://cds.iisc.ac.in/events/phd-thesis-defense-102-cds-seminar-hall- nov-19-wednesday-1000-am-scalable-platform-for-intelligent-orchestration-o f-autonomous-systems-across-edge-cloud-continuum/ SUMMARY:PhD Thesis Defense: #102: CDS Seminar Hall: Nov -19 (Wednesday) @ 1 0:00 AM: "Scalable Platform for Intelligent Orchestration of Autonomous Sy stems Across Edge-Cloud Continuum" DESCRIPTION:DEPARTMENT OF COMPUTATIONAL AND DATA SCIENCES\nPh.D. Thesis Def ense\n\n\n\nSpeaker : Ms. Suman Raj\nS.R. Number : 06-18-01-10-12-20-1-184 50\nTitle : "Scalable Platform for Intelligent Orchestration of Autonomous Systems Across Edge-Cloud Continuum"\nResearch Supervisors: Prof. Yogesh Simmhan\nThesis Examiner : Prof. Sandip Chakraborty\, Indian Institute of Technology\, Kharagpur\nDate &\; Time : November 19\, 2025 (Wednesday)\ , 10:00 AM\nVenue : # 102 CDS Seminar Hall\n\n\n\nABSTRACT\nThe benefits o f autonomous mobile platforms\, such as unmanned aerial vehicles (UAVs) eq uipped with onboard cameras\, are enhanced by compact edge accelerators th at are co-located\, such as the NVIDIA Jetson with 100s of CUDA cores. The y enable rapid inference of Deep Neural Network (DNN) models and computer vision algorithms to support real-time analytics workflows for diverse dom ains\, ranging from smart crop monitoring to assisting Visually Impaired P eople (VIPs)\, either individually or as part of a fleet.\n\nHowever\, pro gramming such drones and edge devices for efficient\, resilient and respon sive operations poses challenges. The limited compute capacity of edge dev ices needs to be intelligently complemented by cloud computing to offload compute-intensive analytics in a timely and cost-effective manner. Routing fleets of drones to accomplish complex tasks requires us to optimize for and adapt to network variability\, edge failures\, latency and energy cons traints\, and monetary costs. Further\, these need to be intuitively progr ammable to design practical applications across distributed autonomous pla tforms and edge resources. We address these challenges in this dissertatio n.\n\nFirst\, we design a task scheduling strategy\, GEMS\, that performs real-time decisions for DNN inference tasks generated by drones\, to execu te them either on local edge accelerators or remote cloud resources. The g oals are to maximize the Quality of Service (QoS) and Quality of Experienc e (QoE) for a VIP assistive application within the deadline constraints of each task. GEMS accounts for the task deadline\, cloud and edge pricing\, and dynamic network variability. Our realistic experiments using up to 84 emulated drones show up to 2.7x higher QoS utility\, up to 75% higher QoE utility\, within ±1m/s3 Jerk\, up to 42% lower yaw error\, and a task co mpletion rate of up to 88% compared to state-of-the-art baselines for dive rse computer vision workloads.\n\nNext\, we study co-scheduling of DNN inf erencing and routes for a fleet of drones used in collaborative applicatio ns such as smart agriculture. The drones need to visit a set of waypoints to collect data and perform analytics\, have access to onboard edge comput e\, stationary fogs at cell towers\, and mobile fogs on public buses. We d efine this as a Mission Scheduling Problem\, which is NP-complete\, and de sign MARC as a divide and assign heuristic to solve this optimization prob lem. Our simulation-based evaluation of MARC with fleets of up to 50 drone s achieves a 100% task completion rate and up to 31% higher average utilit y than contemporary baselines\, and is within 75% of the optimal solution solved using MILP\, which is tractable only for small inputs.\n\nFurther\, we explore resilient scheduling of autonomous systems to ensure continuit y of service despite drone and edge failures in the context of wildfire re sponse\, where a heterogeneous UAV fleet helps detect stranded individuals and generate evacuation routes to safety. We develop the AeroResQ platfor m with algorithms that dynamically adapt to failures by using heartbeats a cross drones and an onboard distributed datastore. Strategies like a load balancing algorithm to address the active requests being processed by fail ed drones and re-assignment of spatial regions across the available drones ensure uninterrupted detection\, route planning\, and monitoring. Our eva luations\, conducted using an emulated environment based on recent Souther n California wildfire data\, demonstrate the robustness of our platform un der failure scenarios and fleet configurations. The system achieves real-t ime performance with ≤1s end-to-end latency per evacuation request\, muc h below the 2s request interval\, while maintaining over 98% successful ta sk reassignment and completion.\n\nFinally\, we design the Ocularone platf orm as an integrated Drones-as-a-Service (DaaS) programming framework that enables rapid development of analytics-driven UAV applications across the edge-cloud continuum. It abstracts drone navigation\, control\, and sensi ng into intuitive\, composable Python-based interfaces and can embed the s cheduling strategies that we have developed\, such as GEMS. This is valida ted for an assistive application to help navigate visually impaired indivi duals using buddy drones. This is further enabled by accurate DNN-based di stance estimation algorithms we have developed to assess nearby obstacles\ , which are fine-tuned with curated datasets to improve real-world accurac y. The application can be implemented in under 40 lines of code using Ocul arone.\n\n\n\nALL ARE WELCOME CATEGORIES:Events,Thesis Defense END:VEVENT BEGIN:VTIMEZONE TZID:Asia/Kolkata X-LIC-LOCATION:Asia/Kolkata BEGIN:STANDARD DTSTART:20241119T100000 TZOFFSETFROM:+0530 TZOFFSETTO:+0530 TZNAME:IST END:STANDARD END:VTIMEZONE END:VCALENDAR