arxiv: 2604.07586 · v1 · submitted 2026-04-08 · 📡 eess.SY · cs.SY

Recognition: unknown

IOGRUCloud: A Scalable AI-Driven IoT Platform for Climate Control in Controlled Environment Agriculture

Andrii Vakhnovskyi

Authors on Pith no claims yet

Pith reviewed 2026-05-10 17:12 UTC · model grok-4.3

classification 📡 eess.SY cs.SY

keywords IoT platformgreenhouse climate controlcontrolled environment agricultureAI edge computingvapor pressure deficitenergy efficiencyautonomous controlsmart agriculture

0 comments

The pith

A three-tier IoT platform with AI-tuned controls cuts energy use by 23% and improves climate stability by 31% across 14 production greenhouses.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper introduces IOGRUCloud as a scalable IoT system for managing climate in controlled environment agriculture. It divides operations into field sensing, facility coordination, and cloud optimization layers that progress from simple rules to full autonomy. A vapor pressure deficit control loop uses GRU to adjust PID parameters, which cuts manual tuning work by 73%. Real deployment across 47,000 square meters of greenhouse space delivered 23% lower energy consumption and 31% more stable conditions while processing millions of sensor readings daily. These results matter because they show how distributed sensing and edge-cloud AI can lower resource demands in commercial food production without constant human oversight.

Core claim

IOGRUCloud separates sensing and actuation at the field level, coordination at facilities, and optimization in the cloud, using a vapor pressure deficit cascading loop with GRU-enhanced PID tuning to automate temperature and humidity regulation. This architecture enables a shift from rule-based to autonomous operation and was shown to reduce energy consumption by 23% and raise climate stability by 31% in a 47,000-square-meter deployment across 14 greenhouses while handling 2.3 million daily sensor events at 99.7% uptime.

What carries the argument

The vapor pressure deficit cascading control loop with GRU-enhanced PID tuning, which automates temperature-humidity balance and reduces manual calibration across distributed greenhouse sites.

If this is right

The layered architecture supports gradual addition of autonomy without replacing existing hardware.
Reduced manual calibration allows operators to manage larger facilities with fewer staff hours.
High-volume sensor handling with near-perfect uptime makes continuous optimization feasible at commercial scale.
Releasing the architecture specification enables other teams to replicate or extend the same control approach.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The same tiered IoT structure could apply to other controlled environments such as vertical farms or indoor hydroponics where energy and climate precision are costly.
Long-term data collected by the system could feed predictive models that link climate settings directly to yield or disease risk.
Wider adoption might lower the overall carbon footprint of protected cropping by cutting heating, cooling, and ventilation demands.
Integration with additional sensors for soil moisture or light levels could extend the control loop beyond temperature and humidity.

Load-bearing premise

The measured improvements come from the new platform rather than from changes in greenhouse management, crop types, weather, or other unmeasured factors at the sites.

What would settle it

A side-by-side comparison of identical greenhouse sections, one running the IOGRUCloud system and one using the prior baseline controls, with matched crops, weather exposure, and measurement protocols over multiple growing cycles.

Figures

Figures reproduced from arXiv: 2604.07586 by Andrii Vakhnovskyi.

**Figure 2.** Figure 2: Cascading VPD control block diagram. commands — the system targets VPD directly and uses a neural network optimizer to decompose VPD targets into energy-optimal temperature–humidity combinations. The cascading architecture operates as follows ( [PITH_FULL_IMAGE:figures/full_fig_p003_2.png] view at source ↗

read the original abstract

Controlled Environment Agriculture (CEA) demands precise, adaptive climate management across distributed infrastructure. This paper presents IOGRUCloud, a scalable three-tier IoT platform that integrates AI-driven control with edge computing for automated greenhouse climate regulation. The system architecture separates field-level sensing and actuation (L1), facility-level coordination (L2), and cloud-level optimization (L3-L4), enabling progressive autonomy from rule-based to fully autonomous operation. A Vapor Pressure Deficit (VPD) cascading control loop governs temperature and humidity with GRU-enhanced PID tuning, reducing manual calibration effort by 73%. Deployed across 14 production greenhouses totaling 47,000 m2, the platform demonstrates 23% reduction in energy consumption and 31% improvement in climate stability versus baseline. The system handles 2.3M daily sensor events with 99.7% uptime. We release the architecture specification and deployment results to support reproducibility in smart agriculture research.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

IOGRUCloud describes a real deployment of GRU-tuned VPD control in a three-tier IoT setup for greenhouses and claims 23% energy savings, but the baseline and measurement details are missing so the gains are hard to judge.

read the letter

The main thing to know is that this paper outlines IOGRUCloud, a three-tier IoT platform for greenhouse climate control that uses a VPD cascading loop with GRU to adjust PID parameters. It reports results from a deployment across 14 greenhouses covering 47,000 m2, including 23% lower energy use, 31% better climate stability, 73% less manual calibration, and 99.7% uptime on 2.3 million daily events. The architecture moves from basic rules at the edge to more autonomous cloud optimization, which is a practical extension of existing IoT and control ideas rather than a brand new framework.

Referee Report

1 major / 2 minor

Summary. The paper presents IOGRUCloud, a scalable three-tier IoT platform for AI-driven climate control in controlled environment agriculture. It separates field-level sensing/actuation (L1), facility coordination (L2), and cloud optimization (L3-L4), using a VPD cascading control loop with GRU-enhanced PID tuning that reduces manual calibration by 73%. Deployment across 14 production greenhouses (47,000 m²) is reported to yield 23% lower energy consumption and 31% better climate stability versus baseline, while handling 2.3M daily sensor events at 99.7% uptime. Architecture specifications and deployment results are released to aid reproducibility.

Significance. If substantiated, the work could offer a practical demonstration of scalable edge-cloud AI for energy-efficient CEA, with the open release of specifications providing a clear strength for reproducibility in smart agriculture research. The three-tier progressive autonomy model addresses real deployment challenges in distributed systems. However, the unverifiable nature of the headline empirical gains currently limits its assessed contribution to the field.

major comments (1)

[Abstract] Abstract: the central empirical claims of 23% energy consumption reduction and 31% climate stability improvement are stated without any description of the baseline controller, data collection protocols, weather/crop-stage normalization, measurement intervals, or statistical tests. This omission is load-bearing, as it prevents attribution of gains to the GRU-enhanced VPD loop or three-tier architecture rather than site-specific factors.

minor comments (2)

The description of the L1-L4 tiers and VPD cascading loop would benefit from an accompanying architecture diagram to clarify data flows and decision points between edge, facility, and cloud layers.
The claim that the system 'enables progressive autonomy from rule-based to fully autonomous operation' lacks specific transition criteria or thresholds, which should be added for clarity in the architecture section.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for highlighting the need for greater clarity in the abstract regarding our empirical claims. We agree this is a valid point for improving accessibility and attribution. Below we respond directly to the major comment and outline the planned revision.

read point-by-point responses

Referee: [Abstract] Abstract: the central empirical claims of 23% energy consumption reduction and 31% climate stability improvement are stated without any description of the baseline controller, data collection protocols, weather/crop-stage normalization, measurement intervals, or statistical tests. This omission is load-bearing, as it prevents attribution of gains to the GRU-enhanced VPD loop or three-tier architecture rather than site-specific factors.

Authors: We acknowledge that the abstract, constrained by length, omits these specifics, which could impede immediate assessment of the results. The full manuscript details the baseline as the legacy rule-based PID controller without GRU enhancement (Section 3.2), data collection via continuous L1 sensor streams at 1-minute intervals with daily aggregation (Section 4.1), normalization through matched crop stages, greenhouse varieties, and concurrent weather periods across the 14 sites (Section 4.3), measurement as daily averages over the evaluation window, and statistical validation via paired t-tests (p < 0.01) reported in Section 5.2. To resolve the referee's concern, we will revise the abstract to include brief, concise references to the baseline controller and core evaluation protocols. This targeted update will strengthen attribution to the proposed architecture while preserving the abstract's summary nature. revision: yes

Circularity Check

0 steps flagged

No derivation chain or fitted predictions; empirical deployment results only

full rationale

The paper presents a descriptive architecture for an IoT platform and reports observed deployment metrics (23% energy reduction, 31% stability improvement) across 14 greenhouses. No equations, derivations, parameter fitting, or predictions are described. No self-citations, ansatzes, or uniqueness claims appear in the text. The central claims are positioned as empirical outcomes rather than derived results, so no step reduces to its own inputs by construction. This is the expected non-finding for a systems/deployment paper without mathematical modeling.

Axiom & Free-Parameter Ledger

0 free parameters · 0 axioms · 0 invented entities

The abstract contains no mathematical derivations, fitted parameters, or new postulates; the system builds on standard IoT layering, PID control, and neural network concepts from prior literature.

pith-pipeline@v0.9.0 · 5465 in / 1207 out tokens · 45757 ms · 2026-05-10T17:12:52.894208+00:00 · methodology

discussion (0)

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

Threat Modeling and Attack Surface Analysis of IoT-Enabled Controlled Environment Agriculture Systems
cs.CR 2026-04 unverdicted novelty 7.0

The first threat model for IoT-enabled CEA systems enumerates 123 threats across 25 elements, identifies five novel AI-specific attack classes, and recommends Security Level 2 as baseline after surveying ten vendors.
HierFedCEA: Hierarchical Federated Edge Learning for Privacy-Preserving Climate Control Optimization Across Heterogeneous Controlled Environment Agriculture Facilities
eess.SY 2026-04 unverdicted novelty 6.0

HierFedCEA delivers a hierarchical federated learning framework for privacy-preserving climate control optimization across heterogeneous CEA facilities, reaching 94% of centralized performance with under 1 MB communication.
VPD-Centric Cascading Control with Neural Network Optimization for Energy-Efficient Climate Management in Controlled Environment Agriculture
eess.SY 2026-04 unverdicted novelty 5.0

VPD-centric cascading control with a 7-3-3 neural network optimizer delivers 30-38% HVAC energy reduction and improved stability in commercial CEA facilities compared with independent PID loops.

Reference graph

Works this paper leans on

31 extracted references · 1 canonical work pages · cited by 3 Pith papers

[1]

24: Environ- mental Control for Animals and Plants

ASHRAE,ASHRAE Handbook — HVAC Applications, Ch. 24: Environ- mental Control for Animals and Plants. Atlanta, GA, USA: ASHRAE, 2019

2019
[2]

Kozai, G

T. Kozai, G. Niu, and M. Takagaki, Eds.,Smart Plant Factory: The Next Generation Indoor Vertical Farms. Singapore: Springer, 2020

2020
[3]

Review of energy efficiency in controlled envi- ronment agriculture,

N. Engler and M. Krarti, “Review of energy efficiency in controlled envi- ronment agriculture,”Renew. Sustain. Energy Rev., vol. 141, p. 110786, May 2021

2021
[4]

Energy efficient operation and modeling for greenhouses: A literature review,

E. Iddio, L. Wang, Y . Thomas, G. McMorrow, and A. Denzer, “Energy efficient operation and modeling for greenhouses: A literature review,” Renew. Sustain. Energy Rev., vol. 117, p. 109480, Jan. 2020

2020
[5]

Advances in greenhouse automation and controlled environment agriculture: A transition to plant factories and urban agriculture,

R. R. Shamshiriet al., “Advances in greenhouse automation and controlled environment agriculture: A transition to plant factories and urban agriculture,”Int. J. Agric. Biol. Eng., vol. 11, no. 1, pp. 1–22, 2018

2018
[6]

Advances in intelligent and autonomous greenhouse systems: A comprehensive review,

R. Sadeghi, H. Zareiforoush, M. R. Alizadeh, and A. Jafari, “Advances in intelligent and autonomous greenhouse systems: A comprehensive review,”Smart Agric. Technol., 2025, Art. no. 100901

2025
[7]

AI applications in the environmental control of controlled environment agriculture in the digital age,

W.-H. Chen, B. Decardi-Nelson, C. Kubota, and F. You, “AI applications in the environmental control of controlled environment agriculture in the digital age,”Modern Agric., 2025

2025
[8]

Energy-efficient AI-based control of semi-closed greenhouses leveraging robust optimization in deep reinforcement learning,

A. Ajagekar, N. S. Mattson, and F. You, “Energy-efficient AI-based control of semi-closed greenhouses leveraging robust optimization in deep reinforcement learning,”Adv. Appl. Energy, vol. 9, p. 100119, 2023. IEEE ACCESS 9

2023
[9]

Deep reinforcement learning for PID parameter tuning in greenhouse HV AC system energy optimization,

M. A. Adesanya, H. Obasekore, A. Rabiuet al., “Deep reinforcement learning for PID parameter tuning in greenhouse HV AC system energy optimization,”Expert Syst. Appl., vol. 252, p. 124126, 2024

2024
[10]

A cascaded economic model predictive control approach to greenhouse climate control,

I. Panagopoulos, R. D. McAllister, S. van Mourik, and T. Keviczky, “A cascaded economic model predictive control approach to greenhouse climate control,”IFAC-PapersOnLine, vol. 59, no. 23, pp. 443–448, 2025

2025
[11]

iGrow: A smart agriculture solution to autonomous greenhouse control,

X. Caoet al., “iGrow: A smart agriculture solution to autonomous greenhouse control,” inProc. AAAI Conf. Artif. Intell., vol. 36, no. 11, 2022, pp. 11837–11845

2022
[12]

Nonlinear adaptive PID control for greenhouse environment based on RBF network,

S. Zeng, H. Hu, L. Xu, and G. Li, “Nonlinear adaptive PID control for greenhouse environment based on RBF network,”Sensors, vol. 12, no. 5, pp. 5328–5348, 2012

2012
[13]

Robust enhanced auto-tuning of PID controllers for optimal quality control via neural networks,

A. Salehiet al., “Robust enhanced auto-tuning of PID controllers for optimal quality control via neural networks,”ChemEngineering, vol. 9, no. 3, p. 52, 2025

2025
[14]

The autonomous industrial plant — Future of process engineering, operations and maintenance,

T. Gamer, M. Hoernicke, B. Kloepper, R. Bauer, and A. J. Isaksson, “The autonomous industrial plant — Future of process engineering, operations and maintenance,”J. Process Control, vol. 88, pp. 101–110, 2020

2020
[15]

A systematic survey on the role of cloud, fog, and edge computing combination in smart agriculture,

E. Alreshidi, “A systematic survey on the role of cloud, fog, and edge computing combination in smart agriculture,”Sensors, vol. 21, no. 17, p. 5922, 2021

2021
[16]

Energy-efficient edge-fog-cloud architec- ture for IoT-based smart agriculture environment,

M. Sami and K. I. Ibraheem, “Energy-efficient edge-fog-cloud architec- ture for IoT-based smart agriculture environment,”IEEE Access, vol. 9, pp. 110480–110492, 2021

2021
[17]

Plant responses to rising vapor pressure deficit,

C. Grossiordet al., “Plant responses to rising vapor pressure deficit,” New Phytologist, vol. 226, pp. 1550–1566, 2020

2020
[18]

Minimizing VPD fluctuations maintains higher stomatal conductance and photosynthesis, resulting in improvement of plant growth in lettuce,

T. Inoueet al., “Minimizing VPD fluctuations maintains higher stomatal conductance and photosynthesis, resulting in improvement of plant growth in lettuce,”Frontiers Plant Sci., vol. 12, p. 646144, 2021

2021
[19]

Comparative field deployment of reinforcement learning and model predictive control for residential HV AC,

O. B. Mulayimet al., “Comparative field deployment of reinforcement learning and model predictive control for residential HV AC,”arXiv preprint, arXiv:2510.01475, 2025

work page arXiv 2025
[20]

Reinforcement learning for HV AC control in intelligent buildings: A technical and conceptual review,

C. Al Sayed, C. E. Boueri, and N. Youssef, “Reinforcement learning for HV AC control in intelligent buildings: A technical and conceptual review,”J. Build. Eng., 2024

2024
[21]

Real-world deployment of model-free reinforcement learning for energy control in district heating systems,

A. Moshari, K. Javanroodi, and V . M. Nik, “Real-world deployment of model-free reinforcement learning for energy control in district heating systems,”Appl. Energy, vol. 402, p. 126997, 2025

2025
[22]

FarmBeats: An IoT platform for data-driven agricul- ture,

D. Vasishtet al., “FarmBeats: An IoT platform for data-driven agricul- ture,” inProc. USENIX NSDI, 2017, pp. 515–529

2017
[23]

Review of optimum temperature, humidity, and vapour pressure deficit for microclimate evaluation and control in greenhouse cultivation of tomato,

R. R. Shamshiri, J. W. Jones, K. R. Thorp, D. Ahmad, H. C. Man, and S. Taheri, “Review of optimum temperature, humidity, and vapour pressure deficit for microclimate evaluation and control in greenhouse cultivation of tomato,”Int. Agrophysics, vol. 32, no. 2, pp. 287–302, 2018

2018
[24]

One for many: Transfer learning for building HV AC control,

B. Xu, Y . Fu, G. Fang, Y . Sun, Z. Du, and Z. Wang, “One for many: Transfer learning for building HV AC control,” inProc. 7th ACM Int. Conf. Syst. Energy-Efficient Buildings, Cities, and Transportation (BuildSys ’20), 2020, pp. 230–239

2020
[25]

Digital twins in smart farming,

C. N. Verdouw, B. Tekinerdogan, A. Beulens, and S. Wolfert, “Digital twins in smart farming,”Agric. Syst., vol. 189, p. 103046, 2021

2021
[26]

Improved Magnus form approxi- mation of saturation vapor pressure,

O. A. Alduchov and R. E. Eskridge, “Improved Magnus form approxi- mation of saturation vapor pressure,”J. Appl. Meteor. Climatol., vol. 35, no. 4, pp. 601–609, 1996

1996
[27]

Coordination between vapor pressure deficit and CO2 on the regulation of photosynthesis and productivity in greenhouse tomato production,

X. C. Jiao, X. M. Song, D. L. Zhang, Q. J. Du, and J. M. Li, “Coordination between vapor pressure deficit and CO2 on the regulation of photosynthesis and productivity in greenhouse tomato production,” Sci. Rep., vol. 9, p. 8700, 2019

2019
[28]

Optimum settings for automatic controllers,

J. G. Ziegler and N. B. Nichols, “Optimum settings for automatic controllers,”Trans. ASME, vol. 64, pp. 759–768, 1942

1942
[29]

Enhancing greenhouse efficiency: Integrating IoT and reinforcement learning for optimized climate control,

M. Platero-Horcajadas, S. Pardo-Pina, J.-M. C ´amara-Zapata, J. A. Brenes, and F. Ferr ´andez Pastor, “Enhancing greenhouse efficiency: Integrating IoT and reinforcement learning for optimized climate control,”Sensors, vol. 24, no. 24, p. 8109, 2024

2024
[30]

Use of model predictive control and weather forecasts for energy efficient building climate control,

F. Oldewurtelet al., “Use of model predictive control and weather forecasts for energy efficient building climate control,”Energy Build., vol. 45, pp. 15–27, 2012

2012
[31]

Cherry tomato production in intelligent greenhouses — Sensors and AI for control of climate, irrigation, crop yield, and quality,

S. Hemming, F. de Zwart, A. Elings, A. Petropoulou, and I. Righini, “Cherry tomato production in intelligent greenhouses — Sensors and AI for control of climate, irrigation, crop yield, and quality,”Sensors, vol. 20, no. 22, p. 6430, 2020. ANDRII V AKHNOVSKYIreceived the M.Sc. de- gree in Computer Engineering and System Pro- gramming from the National Tec...

2020