A Resilient Solution for Sewer Overflow Monitoring across Cloud and Edge
A Resilient Solution for Sewer Overflow Monitoring across Cloud and Edge
Authors: Vipin Singh, Tianheng Ling, Peter Ghaly, Felix Grimmeisen, Gregor Schiele, Felix Bießmann Date: 2026-05-11 Paper ID: openalex:2605.10592
Summary
This paper introduces a resilient monitoring framework for combined sewer systems, designed to address the environmental risks posed by extreme rainfall. The system utilizes deep learning-based forecasting for sewer basin filling dynamics, implemented across a hybrid cloud-edge architecture. By ensuring local processing at the edge, the framework maintains functionality even during network outages, providing a robust tool for real-time monitoring and preventive intervention.
Key Contributions
- Developed a web-based monitoring system for combined sewer overflows that deploys deep learning models across both cloud and edge environments.
- Ensures system resilience against network outages by enabling local inference capabilities at the edge.
- Provides an interactive dashboard for visualizing filling dynamics and capacity exceedance in sewer basins.
Open Questions & Future Work
Archivist Review
The paper describes a deployment architecture for sewer monitoring, which focuses on engineering integration rather than a novel methodological contribution to the field of time-series forecasting. The open question regarding data robustness was refined to focus on critical infrastructure, as sensor degradation in such environments is a persistent and non-trivial problem.
Approved Open Questions
- Robustness to data degradation: Ensuring model robustness against real-world sensor degradation is critical for preventing false positives or missed overflow warnings in safety-critical infrastructure.
Rejected Candidates
- [concept] Cloud-Edge Monitoring Architecture (
cloud-edge-monitoring-architecture) - not_novel: General architecture pattern without specific novel mechanism.
Links
Metadata & Links
- url
- https://arxiv.org/abs/2605.10592
- paper_id
- 2605.10592
- paper_source
- openalex
- domain
- time-series
- tags
-
- architectures
-
- datasets
-
- concept_slugs
-
- dataset_slugs
-
- skill
- TimeSeriesSkill
- created_at
- 2026-05-14T06:10:26Z