Search:
Match:
2 results

Analysis

This paper addresses the challenge of enabling physical AI on resource-constrained edge devices. It introduces MERINDA, an FPGA-accelerated framework for Model Recovery (MR), a crucial component for autonomous systems. The key contribution is a hardware-friendly formulation that replaces computationally expensive Neural ODEs with a design optimized for streaming parallelism on FPGAs. This approach leads to significant improvements in energy efficiency, memory footprint, and training speed compared to GPU implementations, while maintaining accuracy. This is significant because it makes real-time monitoring of autonomous systems more practical on edge devices.
Reference

MERINDA delivers substantial gains over GPU implementations: 114x lower energy, 28x smaller memory footprint, and 1.68x faster training, while matching state-of-the-art model-recovery accuracy.

Analysis

This paper addresses the computational challenges of large-scale Optimal Power Flow (OPF) problems, crucial for efficient power system operation. It proposes a novel decomposition method using a sensitivity-based formulation and ADMM, enabling distributed solutions. The key contribution is a method to compute system-wide sensitivities without sharing local parameters, promoting scalability and limiting data sharing. The paper's significance lies in its potential to improve the efficiency and flexibility of OPF solutions, particularly for large and complex power systems.
Reference

The proposed method significantly outperforms the typical phase-angle formulation with a 14-times faster computation speed on average.