Career: reliable and communication-aware ad-hoc teaming and collaborative computing for cloudless AI at the edge

The National Science Foundation Division of Information and Intelligent Systems awarded the University of Houston $293,442 on October 1, 2026, under the Computer and Information Science and Engineering program (CFDA 47.070) to develop methods enabling distributed edge devices to form ad-hoc computing teams and solve complex tasks without cloud infrastructure dependency.

The research addresses reliability of smart devices such as sensors, phones, and small robots when cloud connections fail due to outages or poor network coverage in rural areas. The project investigates how nearby devices can dynamically coordinate computing resources and communication under resource constraints and intermittent connectivity to support critical applications including disaster response and remote healthcare. Research thrusts include decentralized decision-making mechanisms for devices to form cooperative groups, adaptive approaches for partitioning and scheduling artificial intelligence workloads across heterogeneous devices in real time, learning-based reliability modeling and distributed validation mechanisms to identify faulty or adversarial behavior, and communication-efficient machine learning techniques that reduce overhead in constrained environments.

The award runs through September 30, 2031, with work performed in Houston, Texas.

Abstract

Many smart devices such as sensors, phones, and small robots often depend on distant cloud servers to process data and run complex tasks such as artificial intelligence. When these connections fail, such as during outages or in rural areas with weak networks, many systems stop working. This project develops a new way for nearby devices to form teams on the fly, share their computing power, and solve complex problems together without relying on the cloud. The goal is to make technology more reliable, especially in places where internet access is limited or unreliable, and to support critical applications like disaster response and remote healthcare. The research investigates how distributed edge devices can dynamically coordinate computing and communication under resource constraints and intermittent connectivity. The work is organized into several interconnected research thrusts. One thrust studies decentralized decision-making mechanisms that enable devices to form cooperative groups while balancing efficiency, incentives, and resource availability. IT also explores adaptive approaches for partitioning and scheduling large-scale artificial intelligence workloads across heterogeneous devices in real time. Additional efforts focus on improving robustness and trustworthiness through learning-based reliability modeling and distributed validation mechanisms that can identify faulty or adversarial behavior. A final thrust examines communication-efficient machine learning techniques that reduce overhead while maintaining performance in constrained environments. The project has broad benefits across education, society, and the economy. IT enables reliable computing in areas with limited connectivity, helping applications such as rural communication, disaster recovery, transportation systems, and space exploration. Educational activities include hands on learning with robotics and embedded systems, outreach programs for pre college students, and interactive events that increase interest in science and engineering. The work also contributes to workforce development by training students in emerging areas of edge computing and artificial intelligence. In the long term, this research can support new business opportunities, strengthen national security by reducing reliance on centralized infrastructure, and help create more resilient and accessible technology systems. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the foundation's intellectual merit and broader impacts review criteria.- subawards are not planned for this award.