Search for: All records

Model Recovery (MR) enables safe, explainable decision-making in mission-critical autonomous systems (MCAS) by learning governing dynamical equations, but its deployment on edge devices is hindered by the iterative nature of neural ordinary differential equations (NODE), which are inefficient on FPGAs. Memory and energy consumption are the main concern of applying MR on edge devices for real-time running MR. We propose MERINDA, a novel FPGA-accelerated MR framework that replaces iterative solvers with a parallelizable neural architecture equivalent to NODEs. MERINDA achieves nearly 11× lower DRAM usage and 2.2× faster runtime compared to mobile GPUs. Experiments reveal an inverse relationship between memory and energy at fixed accuracy, highlighting MERINDA’s suitability for resource-constrained, real-time MCAS. “The implementation and datasets are publicly available at github.com/ImpactLabASU/ECAI2025.” 

The paper presents AIIM, an Artificial Intelligence (AI) enabled personalIzation Management software for human-in-the-loop, human-in-the-plant Learning enabled systems (LES). AIIM can be integrated with LES software to aid a human user to achieve safe and effective operation under dynamically changing contexts. AIIM consists of: A) an AI technique to derive model coefficient of a physics guided surrogate model from operational data shared following privacy norms, and b) continuous model conformance to identify key changes in LES operational behavior that may jeopardize safety. We demonstrate two capabilities of AIIM, personalization and unknown error detection, through case studies that span a significant breadth of dynamic context change scenarios including: a) involuntary change in user context such as medication induced glucose metabolism change in automated insulin delivery (AID), b) actuation failure such as cartridge blockage in AID, c) latent sensor error in aviation, and d) unknown coding error in autonomous car software patches. We compare AIIM personalization with human-in-the-loop and self-adaptive model-predictive control design on real-life and simulation settings, to show safe and improved diabetes management. 

The objective of our research is to create efficient methods and tools for the quick and thorough assessment of emerging digital circuit devices, facilitating the adoption of promising ones. In this work, we develop methods and tools for hybrid technology that combines memristors with MOS transistors and demonstrates their effectiveness. Although several types of memristor-transistor logic have been proposed, 15 years of research has created a small set of logic cells. We propose a systematic method for generating new and efficient memristor-transistor single-phase combinational logic cells. At the core of our approach is a cell enumerator, which enables us to explore a wide range of cell designs, including nonintuitive ones, and a data-driven inductive learning method, which identifies new properties of such cells and scales up our explorations. In conjunction with other completely new tools, these create a comprehensive and definitive library of logic cells. Our new cells provide significant improvements or significantly distinct Pareto-optimal alternatives for the few logic functions for which prior research has created cells. Importantly, our methods enable us to discover a previously unknown synergistic operation between memristors and transistors that occurs for specific cell topologies. We harness this synergy to develop a method for adding memristors to low-area pass-transistor circuits such that they produce strong output voltages and low power, including for patterns that cause ratioed operation. We have also developed a new memristor-transistor logic family, namely controlled-AND (cAND)/controlled-OR (cOR), which includes many of the best cells. We have also developed a constructive method for designing such cells.