Automated FPGA-Based Soft Sensors for Wastewater Flow Estimation Using Deep Learning

A prototype IoT device has been developed for automated wastewater flow estimation using Deep Learning-based soft sensors. The research addresses three key challenges: limited datasets, inconvenient toolchains for on-device AI model development, and hardware platforms not optimized for energy-efficient soft sensor applications. This end-to-end solution demonstrates promise for reliability and energy efficiency when deployed on resource-limited IoT devices. The study specifically targets the underexplored field of wastewater flow estimation. The automated approach aims to overcome existing gaps in available datasets and hardware optimization. The work was published on arXiv with the identifier 2407.05102. The research falls under the categories of Electrical Engineering, Systems Science, and Signal Processing. The automated solution represents a significant advancement in applying DL-based soft sensors to environmental monitoring applications.