arXiv subjects

Branko Kerkez

Publications and source records attributed to Branko Kerkez.

8 recordsLinked to original sources

arXiv2026

Performance-Guaranteed Reference Tracking With Power Directionality Constraints: Application to Controlled Stochastic Watersheds

Modern stormwater infrastructure faces increased demands that require a corresponding increase in capacity. Traditionally, these demands have been met by constructing new infrastructure assets, which is a costly endeavor. More recently, many system operators have achieved great success in employing feedback control techniques to improve system performance. However, the resulting closed-loop system exhibits power directionality constraints that introduce nonlinear constraints in feedback synthesis. In this work, we develop a stochastic control synthesis procedure with provable performance bounds on mean-square reference tracking for a general class of problems in which power directionality constraints arise. The proposed method is then applied to a flood mitigation example using a numerical model of a real-world smart water system. The key result is an extension of the performance-guaranteed control (PGC) framework, which was originally designed for disturbance rejection, to accommodate reference tracking control objectives.

eess.SY↗

arXiv2023

Identification of stormwater control strategies and their associated uncertainties using Bayesian Optimization

Dynamic control is emerging as an effective methodology for operating stormwater systems under stress from rapidly evolving weather patterns. Informed by rainfall predictions and real-time sensor measurements, control assets in the stormwater network can be dynamically configured to tune the behavior of the stormwater network to reduce the risk of urban flooding, equalize flows to the water reclamation facilities, and protect the receiving water bodies. However, developing such control strategies requires significant human and computational resources, and a methodology does not yet exist for quantifying the risks associated with implementing these control strategies. To address these challenges, in this paper, we introduce a Bayesian Optimization-based approach for identifying stormwater control strategies and estimating the associated uncertainties. We evaluate the efficacy of this approach in identifying viable control strategies in a simulated environment on real-world inspired combined and separated stormwater networks. We demonstrate the computational efficiency of the proposed approach by comparing it against a Genetic algorithm. Furthermore, we extend the Bayesian Optimization-based approach to quantify the uncertainty associated with the identified control strategies and evaluate it on a synthetic stormwater network. To our knowledge, this is the first-ever stormwater control methodology that quantifies uncertainty associated with the identified control actions. This Bayesian optimization-based stormwater control methodology is an off-the-shelf control approach that can be applied to control any stormwater network as long we have access to the rainfall predictions, and there exists a model for simulating the behavior of the stormwater network.

cs.LG↗

arXiv2023

Measuring city-scale green infrastructure drawdown dynamics using internet-connected sensors in Detroit

The impact of green infrastructure (GI) on the urban drainage landscape remains largely unmeasured at high temporal and spatial scales. To that end, a data toolchain is introduced, underpinned by a novel wireless sensor network for continuously measuring real-time water levels in GI. The internet-connected sensors enable the collection of high-resolution data across large regions. A case study in Detroit (MI, US) is presented, where the water levels of 14 GI sites were measured in-situ from June to September 2021. The large dataset is analyzed using an automated storm segmentation methodology, which automatically extracts and analyzes individual storms from measurement time series. Storms are used to parameterize a dynamical system model of GI drawdown dynamics. The model is completely described by the decay constant α, which is directly proportional to the drawdown rate. The parameter is analyzed across storms to compare GI dynamics between sites and to determine the major design and physiographic features that drive drawdown dynamics. A correlation analysis using Spearman's rank correlation coefficient reveals that depth to groundwater, imperviousness, longitude, and drainage area to surface area ratio are the most important features explaining GI drawdown dynamics in Detroit. A discussion is provided to contextualize these finding and explore the implications of data-driven strategies for GI design and placement.

eess.SY↗

arXiv2021

pystorms: A simulation sandbox for the development and evaluation of stormwater control algorithms

Recent accessibility of affordable sensing technologies, microcontrollers, and wireless communication technology has made it possible for stormwater systems to be retrofitted with an assortment of sensors and actuators. These smart stormwater systems have enabled the real-time sensing of their surrounding environmental dynamics, and subsequently, provide the basis for autonomous and adaptive operational control strategies. Additionally, these systems allow for inexpensive and minimally-invasive stormwater control interventions (e.g. hydraulic valve operated by cellularly-connected actuator) in lieu of new construction. However promising this area of smart stormwater control, there still remain barriers -- for experts and novices alike -- to access a set of shared tools and methods for investigating, developing, and contributing to it. In an effort to make smart stormwater control research more methodical, efficient, and accessible, we present pystorms, a Python-based simulation sandbox that facilitates the quantitative evaluation and comparison of control strategies. The pystorms simulation sandbox comes with (i) a collection of real world-inspired stormwater control scenarios on which any number of control strategies can be applied and tested via (ii) an accompanying Python programming interface coupled with a stormwater simulator. For the first time, pystorms enables rigorous and efficient evaluation of smart stormwater control methodologies across a diverse set of watersheds with only a few lines of code. We present the details of pystorms here and demonstrate its usage by applying and evaluating two stormwater control strategies.

eess.SY↗

arXiv2018

Hydrograph peak-shaving using a graph-theoretic algorithm for placement of hydraulic control structures

The need to attenuate hydrograph peaks is central to the design of stormwater and flood control systems. However, few guidelines exist for siting hydraulic control structures such that system-scale benefits are maximized. This study presents a graph-theoretic algorithm for stabilizing the hydrologic response of watersheds by placing controllers at strategic locations in the drainage network. This algorithm identifies subcatchments that dominate the peak of the hydrograph, and then finds the "cuts" in the drainage network that maximally isolate these subcatchments, thereby flattening the hydrologic response. Evaluating the performance of the algorithm through an ensemble of hydrodynamic simulations, we find that our controller placement algorithm produces consistently flatter discharges than randomized controller configurations---both in terms of the peak discharge and the overall variance of the hydrograph. By attenuating flashy flows, our algorithm provides a powerful methodology for mitigating flash floods, reducing erosion, and protecting aquatic ecosystems. More broadly, we show that controller placement exerts an important influence on the hydrologic response and demonstrate that analysis of drainage network structure can inform more effective stormwater control policies.

physics.soc-ph↗

arXiv2018

Vehicles as sensors: high-accuracy rainfall maps from windshield wiper measurements

Connected vehicles are poised to transform the field of environmental sensing by enabling acquisition of scientific data at unprecedented scales. Drawing on a real-world dataset collected from almost 70 connected vehicles, this study generates improved rainfall estimates by combining weather radar with windshield wiper observations. Existing methods for measuring precipitation are subject to spatial and temporal uncertainties that compromise high-precision applications like flash flood forecasting. Windshield wiper measurements from connected vehicles correct these uncertainties by providing precise information about the timing and location of rainfall. Using co-located vehicle dashboard camera footage, we find that wiper measurements are a stronger predictor of binary rainfall state than traditional stationary gages or radar-based measurements. We introduce a Bayesian filtering framework that generates improved rainfall estimates by updating radar rainfall fields with windshield wiper observations. We find that the resulting rainfall field estimate captures rainfall events that would otherwise be missed by conventional measurements. We discuss how these enhanced rainfall maps can be used to improve flood warnings and facilitate real-time operation of stormwater infrastructure.

eess.SP↗

arXiv2017

Open storm: a complete framework for sensing and control of urban watersheds

Leveraging recent advances in technologies surrounding the Internet of Things, "smart" water systems are poised to transform water resources management by enabling ubiquitous real-time sensing and control. Recent applications have demonstrated the potential to improve flood forecasting, enhance rainwater harvesting, and prevent combined sewer overflows. However, adoption of smart water systems has been hindered by a limited number of proven case studies, along with a lack of guidance on how smart water systems should be built. To this end, we review existing solutions, and introduce open storm---an open-source, end-to-end platform for real-time monitoring and control of watersheds. Open storm includes (i) a robust hardware stack for distributed sensing and control in harsh environments (ii) a cloud services platform that enables system-level supervision and coordination of water assets, and (iii) a comprehensive, web-based "how-to" guide, available on open-storm.org, that empowers newcomers to develop and deploy their own smart water networks. We illustrate the capabilities of the open storm platform through two ongoing deployments: (i) a high-resolution flash-flood monitoring network that detects and communicates flood hazards at the level of individual roadways and (ii) a real-time stormwater control network that actively modulates discharges from stormwater facilities to improve water quality and reduce stream erosion. Through these case studies, we demonstrate the real-world potential for smart water systems to enable sustainable management of water resources.

eess.SY↗

arXiv2017

Distance-Penalized Active Learning Using Quantile Search

Adaptive sampling theory has shown that, with proper assumptions on the signal class, algorithms exist to reconstruct a signal in $\mathbb{R}^{d}$ with an optimal number of samples. We generalize this problem to the case of spatial signals, where the sampling cost is a function of both the number of samples taken and the distance traveled during estimation. This is motivated by our work studying regions of low oxygen concentration in the Great Lakes. We show that for one-dimensional threshold classifiers, a tradeoff between the number of samples taken and distance traveled can be achieved using a generalization of binary search, which we refer to as quantile search. We characterize both the estimation error after a fixed number of samples and the distance traveled in the noiseless case, as well as the estimation error in the case of noisy measurements. We illustrate our results in both simulations and experiments and show that our method outperforms existing algorithms in the majority of practical scenarios.

stat.ML↗