FlowPrecision: Advancing FPGA-Based Real-Time Fluid Flow Estimation with Linear Quantization

2024-03-04

2024 IEEE International Conference on Pervasive Computing and Communications Workshops and other Affiliated Events (PerCom Workshops)

AI Summary

To address the challenge of real-time, high-precision fluid flow measurement in industrial and environmental monitoring, this paper proposes an FPGA-based end-to-end soft sensing method. The core innovation lies in the first adoption of linear quantization—rather than conventional fixed-point quantization—for deploying neural networks at the edge, thereby overcoming the inherent accuracy-efficiency trade-off. Integrated with a lightweight network architecture and hardware-aware FPGA co-optimization, the approach establishes a low-latency, high-accuracy, on-device estimation framework. Experimental results across multiple datasets demonstrate an average 10.10% reduction in mean squared error and a 9.39% improvement in inference speed, achieving millisecond-level latency and sub-percent estimation error. This significantly enhances both the practicality and generalizability of edge-deployed soft sensors.

Technology Category

• Machine Learning: Hardware-aware ML
• Intelligent Robots: State Estimation
• Search and Optimization: Learning to Search

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Responsible Web: Human-perceived consequences of algorithmic deployment on the web

Abstract

In industrial and environmental monitoring, achieving real-time and precise fluid flow measurement remains a critical challenge. This study applies linear quantization in FPGA-based soft sensors for fluid flow estimation, significantly enhancing Neural Network model precision by overcoming the limitations of traditional fixed-point quantization. Our approach achieves up to a 10.10% reduction in Mean Squared Error and a notable 9.39% improvement in inference speed through targeted hardware optimizations. Validated across multiple data sets, our findings demonstrate that the optimized FPGA-based quantized models can provide efficient, accurate real-time inference, offering a viable alternative to cloud-based processing in pervasive autonomous systems.

Problem

• Enhancing real-time fluid flow measurement precision
• Overcoming limitations of traditional fixed-point quantization
• Providing efficient FPGA-based alternative to cloud processing

Innovation

• Uses linear quantization for FPGA fluid flow sensors
• Reduces error and boosts speed via hardware optimization
• Enables efficient real-time inference on FPGA platforms