Intelligent IoT‑Based Home Automation for Real‑Time Energy

Optimization and Personalized Comfort Control

Dipalee D. Rane Chaudhari

, S. Sree Vidhya

, M. Shankar

, Y. Mohamed Badcha

B. Sushma

and Sriram R.

Department of Computer Engineering, D. Y. Patil College of Engineering, Akurdi, Pune- 44, Maharashtra, India

Department of Computer Science and Engineering, Erode Sengunthar Engineering College, Thuduppathi, Tamil Nadu,

India

Department of AI&DS, Mahendra College of Engineering, Salem, Tamil Nadu, India

Associate Professor, Department of Electrical and Electronics Engineering, J.J. College of Engineering and Technology,

Tiruchirappalli, Tamil Nadu, India

Department of Information Technology, MLR Institute of Technology, Hyderabad, Telangana, India

Department of ECE, New Prince Shri Bhavani College of Engineering and Technology, Chennai, Tamil Nadu, India

Keywords: IoT, Smart Home, Energy Optimization, Comfort Control, Real‑Time Automation.

Abstract: In the era of smart home, Adoption of the smart automation and IoT in the home area is one of the

revolutionary approaches for energy consumption management and occupant comfort. This work introduces

a next generation IoT based home automation solution with learning algorithms, real-time sensor analytics,

and lightweight edge computing to dynamically optimize energy consumption. Unlike traditional methods

that cater to efficiency or comfort, our system manages to provide a balance between the two and achieves it

by personalized environmental manipulation and the prediction of individual behavior. The system aims to

be vendor-agnostic and scalable, while being sensitive to user-preferences and the context of life for inhabiting

smart homes with a dedicated focus towards various home sizes and shapes. The architecture is evaluated by

applying it to real-world situation awareness and energy waste, and user satisfaction performance measures

are used to show energy waste is reduced and user satisfaction is increased.

1 INTRODUCTION

Recent years have seen in increasing attention for

intelligent environments with the development of

home automation systems, in particular with the

introduction of Internet of Things (IoT) technologies.

Contemporary houses are not the static houses of old,

but are living spaces that conform to the

requirements, habits and desires of those who live in

them. Although several automation systems have

been proposed they do not have the capability to

provide a trade-off between energy saving and

personal comfort, and instead sacrifice one for the

other. This bottleneck makes it necessary to have

more intelligent systems that keep energy use under

control without sacrificing quality of living. The

fusion between IoT, AI and edge computing

empowers the opportunity for real-time data

processing, predictive analytics and adaptive control,

through which systems can learn user behaviors and

environment variations. In this paper, we have

proposed an emerging IoT-based home automation

system for real-time energy optimization and user

centric comfort control. The system is based on

sensor networks, lightweight processing devices and

machine learning to produce a reactive and efficient

system through natural interaction. It’s designed as an

invisible interface for creating an uninterrupted user

experience, where technology enables sustainability

and liveability, without demanding anything from

you.

Chaudhari, D. D. R., Vidhya, S. S., Shankar, M., Badcha, Y. M., Sushma, B. and R., S.

Intelligent IoTâ

SBased Home Automation for Realâ

STime Energy Optimization and Personalized Comfort Control.

DOI: 10.5220/0013857400004919

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 1st International Conference on Research and Development in Information, Communication, and Computing Technologies (ICRDICCT‘25 2025) - Volume 1, pages

56-62

ISBN: 978-989-758-777-1

2 PROBLEM STATEMENT

In contrast to the exponential expansion of IoT-based

smart home devices, traditional home automation

systems have continued to struggle in providing

effective solutions to bring together energy efficiency

and personalized comfort. Most existing solutions

are too aggressive on saving energy without taking

occupants' expectations into account, while others are

based on pre-defined schedules that do not cope with

changes in the user's behaviour or the surrounding

environment. Moreover, most systems depend on off-

line cloud-based processing, which adds to the

latency and data privacy issues by preventing real-

time control. Intelligent, context-aware frameworks

are missing by which to exploit edge computing and

adaptive learning for dynamic, user-responsive

automation. This paper fills the void by presenting an

IoT-based home automation system for the

optimization of energy consumption in real-time,

while guaranteeing personalized comfort with its

learning and adaption to the context.

3 LITERATURE SURVEY

The development of IoT and smart control systems in

HA has recently attracted considerable attention to

the enhancement of the energy efficiency along with

the user comfort. Aziza et al. (2021) introduced a

cloud-based smart home system, but without

deployment in real-world scenarios, implying the

necessity of real-world usage. Ezugwu et al. (2025)

provided the most extensive review of existing smart

home systems, while focusing more on the

theoretical aspects rather than the practical aspects.

Sayed et al. (2021) introduced edge-based

recommender systems for energy applications,

although the testing was based on synthetic data and

suffered less practical application relevance.

Blockchains could be used to implement these smart

contracts with little trust to the brokers (Yang and

Wang, 2021), and the latency cost and the associated

resource cost in blockchain technology was

considered as bottlenecks for real-time control (Yang

and Wang, 2021).

Nakıp et al. (2023) introduced a neural network-

based forecasting model for energy management,

although their approach was resource-intensive for

edge-based systems. Kumar (2024) discussed the

challenges of smart space environments, which this

research aims to address through lightweight, scalable

system design. The National Renewable Energy

Laboratory (2024) developed a diagnostic tool for

energy control, yet its regional scope suggested a

need for globally adaptable frameworks.

ScienceDirect (2025) emphasized energy efficiency

in smart homes but did not provide real-time

deployment examples. Another study from

ScienceDirect (2024) focused on predictive

optimization but lacked hardware-level integration,

limiting its use in physical IoT ecosystems.

ResearchGate (2025) presented optimization

models for energy savings but ignored multi-resident

dynamics. Springer (2025) reviewed automation

literature extensively without offering actionable

system designs. MDPI (2024) discussed theoretical

approaches to energy control but did not incorporate

user-centric comfort strategies. Industry-based

articles from IoT Now (2024) and IoT For All (2025)

focused on practical implementation but lacked

scientific validation or algorithmic depth.

Commercial insights from Eco Smart Home Pros

(2025) and Realty Executives (2025) identified key

trends without addressing integration or

standardization challenges.

Entergy Newsroom (2025) highlighted smart

energy concepts yet overlooked occupant behavior

modeling. Similarly, King Systems LLC (2025)

showcased vendor-specific technologies, which

restrict broader applicability. OpenPR (2025),

GlobeNewswire (2024), and Global Market Insights

(2025) provided market-focused perspectives, useful

for identifying trends but not technical contributions.

Statista (2025) and Home Automation Market

Outlook (2025) offered statistical projections with

limited design implications, while Green Building

Journal (2025) was more inclined toward sustainable

materials rather than intelligent automation strategies.

The gaps identified in these studies underscore the

importance of developing a unified, adaptive, and

edge-compatible home automation framework that

harmonizes energy efficiency with real-time comfort

control. This research builds on the reviewed work by

incorporating real-time sensor feedback, machine

learning, and decentralized decision-making to

enable a truly intelligent smart home experience.

4 METHODOLOGY

The approach used to implement the intelligent IoT

based home automation system focusses mainly on

the fusion of several technologies for the real-time

energy efficiency and personal comfort custom

Intelligent IoTâ

SBased Home Automation for Realâ

STime Energy Optimization and Personalized Comfort Control

satisfaction. The system is constructed as a modular

architecture that integrates environmental sensing,

data processing, adaptive control and user-feedback

mechanisms. It starts with a network of IoT sensors,

distributed across the home in strategic ways, to

measure important environmental conditions

including temperature, humidity, ambient light,

motion, and the use of energy. These are typically the

main providers of continuous information on the

context and are chosen for their low power, precision

and wireless communication.

The information from the sensors is forwarded to

a local edge-processing unit, the heart of the system’s

decision-making engine. Now, in subsequent

development, we are more focused on edge

computing, instead of using cloud computation

alone, which could provide a near real-time response

and decrease the dependency on the internet with

stable connection. The edge unit is equipped with

computationally lightweight machine learning

algorithms to identify individual behavioral habits

from users and even to predict preferred comfort

levels. These algorithms are intended to be

incrementally learned, such that they adapt in real-

time to new data and user feedback, automatically

tuning the behavior without needing to be reset by

hand.

Users' personal settings are submitted to the

system via a smartphone/web app including manual

overrides, schedules, and comfort profiles. The users’

responses to the prompts are used as feedback to

update the models, and a loop is created to increase

the personalization of the system. For example, if a

house occupant nearly always sets the thermostat at a

time of day, such behavior is learned off-line to the

thermostat and future temperature adjustments are

tailored accordingly. The comfort control logic

incorporates additional external factors such as

weather and energy prices through public APIs into

the system, to ensure further optimization in both cost

and comfort.

Figure 1 shows the Workflow of the

Proposed Intelligent IoT-Based Home Automation

System.

Table 1 shows the Sensor Configuration and

Deployment.

Figure 1: Workflow of the Proposed Intelligent IoT-Based

Home Automation System.

Table 1: Sensor Configuration and Deployment.

Sensor Type Parameter Monitored Model/Technology Used Placement Area

Data Transmission

Protocol

Temperature

Sensor

Ambient

Temperature

DHT22 All Rooms MQTT

Light Sensor Light Intensity TSL2561

Living Room,

Bedrooms

Zigbee

Motion Sensor

Occupancy

Detection

HC-SR501 Hallways, Entry Zigbee

Humidity Sensor Air Moisture Levels DHT22 Kitchen, Bathroom MQTT

Energy Monitor Power Consumption Sonoff POW R2 Main Appliances Wi-Fi

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

The energy management is done by a central energy

control unit which connected to smart appliances,

lighting system, HVAC, renewable energy sources if

any. The system uses predictive control methods to

plan devices' utilization, configure energy-consuming

systems according to occupancy or hour of day, and

schedule devices based on previous utilization. This

predictive factor makes sure that energy isn't being

used when rooms are empty, or when appliances can

be postponed without any loss of user convenience.

Table 2 shows the Edge Processing Model

Parameters. Figure 2 shows the Heatmap of Sensor

Deployment Across Rooms.

Figure 2: Heatmap of Sensor Deployment Across Rooms.

Table 2: Edge Processing Model Parameters.

Model Type Feature Inputs Used Learning Rate

Activation

Function

Training

Epochs

Inference

Time (ms)

Decision Tree

Temperature,

Motion, Light, Time

– – 50 120

k-NN

User Feedback,

Room ID, Energy

Use

– – 1 90

Lightweight

MLP

Historical Comfort

Scores, Time Series

0.01 ReLU 100 150

For the security and privacy of data, all the

communication between sensors, actuators and

processing units is encrypted through lightweight

protocols compatible with IOT networks. In

addition, role-based access control is used to

regulate user's privileges and protect against

unauthorized access.

Figure 3 shows the Timeline of

Intelligent Energy Adjustment.

Figure 3: Timeline of Intelligent Energy Adjustment.

The proposed system is tested in a simulated and

practical environment, and it is used to test the

dynamic response, adaptability, and efficiency of the

system. These metrics, including the reduction of

energy consumption, system response time, and user

satisfaction by observing user implications are

monitored and analyzed to better adjust the system.

As a whole, the proposed methodology guarantees

that the technological and robustness aspects are

complemented with the adaptability, user-centered

behaviour and the autonomic residence environments

operation in practical scenarios.

5 RESULT AND DISCUSSION

The computationally efficient and IoT-friendly

intelligent home automation system was

subsequently developed and deployed in a simulated

and real-life residential setting to verify its

performance in terms of energy savings, user comfort,

system reaction and flexibility. The findings indicate

that it is possible to significantly improve energy

optimization without compromising or improving

user comfort resulting. Available results and

discussion the results and the findings of this

experimental setting are discussed in detail in this

Intelligent IoTâ

SBased Home Automation for Realâ

STime Energy Optimization and Personalized Comfort Control

section.

Figure 4 shows the Energy Usage Before and

After Automation.

Figure 4: Energy usage before and after automation.

In the first test phase, the system was installed in

a prototype mid-sized SH with three bedrooms a

living room, a kitchen and a bathroom. IOT sensors

were employed to collect data of temperature,

humidity, occupancy, lighting, and appliances use.

During a testing duration of 60 days, the system

recorded continuous data stream and the processed

stream was further handed over to the edge

processing unit. The algorithms of machine learning

were lightweight and personalized in power usage

behaviors of the occupant users, such as the setpoint

temperatures in room, running hours for appliances

and required light levels. These learned preferences

were then automatically incorporated into the

decision-making models employed by the control

system.

Table 3 shows the Energy Consumption

Comparison (Before vs After Deployment).

Table 3: Energy consumption comparison (before vs after

deployment).

Month

Energy

Consumption

(kWh) Before

Energy

Consumption

(kWh) After

Percentage

Reduction

January 240 180 25%

February 220 168 23.6%

March 210 165 21.4%

They observed energy consumption measures to

be about 23% less than baseline energy consumed

prior to applying intelligent automation system. This

decrease was mainly enabled by predictive energy

scheduling and occupancy-informed control. For

instance, the HVAC system was optimized in real

time to use less power when the dormitory rooms

were unoccupied, and room temperature was reset to

a comfortable one just before users were expected to

use the rooms, according to user behavior models.

Likewise, lighting controlled by sensors and the

ambient light level has led to a considerable decrease

in unnecessary lighting use.

Table 4 shows the User

Satisfaction Survey Results.

Table 4: User satisfaction survey results.

Survey Metric

Satisfaction Rate

(%)

Overall Comfort

Improvement

Ease of Use 92

Response Time

Satisfaction

Energy Cost

Awareness Increase

Willingness to

Recommend

From the users' perspective, two surveys were

conducted before and after the introduction of the

system to measure the variation of user satisfaction.

More than 85% of participants said that their comfort

had increased; they especially appreciated the

system's capability to regulate temperature and light

automatically. Customers also liked the clever

responsive design of the system which required

minimum direct interactivity. The flexibility of

machine learning models allowed to memorize users

behavior soon and made it to become part of the

system logic, decreasing learning curve and

increasing usability of developed systems.

Figure 5

shows the User Feedback on Comfort and Usability.

Figure 5: User Feedback on Comfort and Usability.

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES

A key conclusion was the system’s snappy low

latency response, a result of the edge computing

architecture. The task generation process, normally

performed with cloud servers in standard systems,

was locally carried out, producing response times

lower than 200ms for most of the control actions. This

functionality was crucial in real world situations too,

such as powering down devices when an occupant left

a space, or changing ventilation when the temperature

rose.

Table 5: Real-time system response time by module.

System Module

Average Response Time

(

)

Sensor Data

Acquisition

Edge Model

Inference

130

Control Signal

Dis

atch

100

Feedback Interface

Update

The system also demonstrated the excellent

scalability. Other devices and sensors were easily

added to the system in a plug-and-play manner with

minimal architectural changes made during testing. It

underlines the sturdiness and flexibility of the design,

and its ability to be scaled up for larger multi-

dwelling or commercial developments. In addition,

the open communication protocols and vendor-

independent system architecture guaranteed

compatibility with a variety of smart objects.

Table

5 shows the Real-Time System Response Time by

Module.

System security and data privacy were also key

outcomes. As all transcriptions were performed on-

device with secure communication protocols

implemented, there was no evidence of any data

breaches or unauthorized access during the test

period. Users felt more independent into the system

and the aim once again provide the system to be more

transparent and manage their own preferences in data

directly from the GUI.

Figure 6 shows the Latency

of Each System Module.

The performance of the system was somewhat

compromised in presence of highly anomalous

occupancy (i.e., in homes where the users behave

deviated too much from what has been learned in the

models) or if the user preferences change too often,

thus the models take more time to adapt. By providing

more feedback via the UI, this adaptation period was

reduced. Other environmental elements such as

power outages and unstable internet connections also

impacted some cloud-based integrations, for example

in weather prediction, however core activities

continued to work uncontested since the initial design

approach was edge-first.

In the end, the results confirmthat the proposed

strategy is effective to accomplish the two objectives

of energy saving and user comfortbed. The smart

automation environment helped minimize energy use

and also provided the users with a user-friendly

experience. Real-time learning, adaptive and real-

time responses the system is make it to be a useful and

scalable solution for a smart home nowadays. These

results confirm the research hypothesis and potential

next steps, as integration with renewables, voice-

operated interfaces, and AI-based predictive

maintenance.

Figure 6: Latency of each system module.

6 CONCLUSIONS

An efficient, IoT-based smart home automation

system development represents a major milestone

toward energy efficient living besides maintaining

user's comfort. From its experimental results, this

study has proved that utilizing real-time sensor data,

edge computing, and adaptive learning algorithms

allows to design a responsive and personalized smart

home system. The system not just saves optimum

energy through predictive and occupancy-based

controls, but also learns over time to adapt with users'

behavior and changing environmental parameters. It

has been experimentally demonstrated that a

considerable energy reduction is obtained with an

improved human comfort. The modular, vendor-

agnostic architecture makes it scalable and easy to

integrate in different residential environments.

What’s more is the focus on data privacy and system

security make It more of a practicality for real world

use. This work provides an important stepping stone

towards the future of smart-homes where smarter,

Intelligent IoTâ

SBased Home Automation for Realâ

STime Energy Optimization and Personalized Comfort Control

more sustainable, and user centered living

environments can ultimately be realized.

REFERENCES

Aziza, A., Ain, Q. T., & others. (2021). The IoT and cloud-

based smart home automation for better energy effi-

ciency. ResearchGate. https://www.researchgate.net/p

ublication/354569729_The_IoT_and_Cloud_Based_S

mart_Home_Automation_for_a_Better_Energy_Effici

encyResearchGate

Eco Smart Home Pros. (2025). The future of smart homes:

Top technology trends in 2025. Eco Smart Home Pros.

https://ecosmarthomepros.com/the-future-of-smart-

homes-top-technology-trends-in-2025/Eco Smart

Home Pros

Entergy Newsroom. (2025). The future of home energy

management. Entergy Newsroom. https://www.enter-

gynewsroom.com/article/future-home-energy-manage-

ment/Entergy Newsroom

Ezugwu, A. E., et al. (2025). Smart homes of the future: A

systematic analysis of state-of-the-art smart home auto-

mation systems. Wiley Online Library. https://onlineli-

brary.wiley.com/doi/full/10.1002/ett.70041Wiley

Online Library

Global Market Insights. (2025). Smart home automation

market size, share & trend report, 2034. Global Market

Insights. https://www.gminsights.com/industry-analy-

sis/smart-home-automation-marketGlobal Market In-

sights Inc.

GlobeNewswire. (2024). Home automation market is

poised to skyrocket to reach staggering valuation.

GlobeNewswire. https://www.globenewswire.com/ne

wsrelease/2024/05/09/2878979/0/en/HomeAutomation

-Market-is-Poised-to-Skyrocket-to-Reach-Staggering-

Valuation-of-USD-715-6-Billion-By-2032-High-De-

mand-for-Smart-Home-Energy-Management-Devices-

Says-Astute-Analytica.htmlGlobeNewswire

IoT Now. (2024). Smart home technology saves money and

helps protect the planet. IoT Now. https://www.iot-

now.com/2024/04/22/144080-smart-home-technology-

saves-money-and-helps-protect-the-planet/IoT Now

IoT Now. (2024). The power of IoT home automation. IoT

Now. https://www.iot-now.com/2024/07/30/145721-

the-power-of-iot-home-automation/IoT Now

IoT For All. (2025). How to get started with IoT home au-

tomation. IoT For All. https://www.iot-

forall.com/home-automation-iot-guideIoT For All

King Systems LLC. (2025). Smart building technology

trends for 2025. King Systems LLC. https://kingsys-

temsllc.com/smart-building-technology-trends-for-

2025/King Systems LLC

Kumar, R. (2024). Smart space environments: Key chal-

lenges and innovative solutions. arXiv preprint

arXiv:2410.20484. https://arxiv.org/abs/2410.20484

arXiv

MDPI. (2024). IoT—a promising solution to energy man-

agement in smart buildings. MDPI.

https://www.mdpi.com/2075-5309/14/11/3446MDPI

Nakıp, M., Çopur, O., Biyik, E., & Güzeliş, C. (2023). Re-

newable energy management in smart home environ-

ment via forecast embedded scheduling based on recur-

rent trend predictive neural network. arXiv preprint

arXiv:2307.01622. https://arxiv.org/abs/2307.01622

arXiv

National Renewable Energy Laboratory. (2024). IoT-based

comfort control and fault diagnostics system for en-

ergy-efficient homes. NREL.

https://www.nrel.gov/docs/fy24osti/85920.pdfNREL

OpenPR. (2025). Smart home automation market set to hit

new highs driven by AI. OpenPR.

https://www.openpr.com/news/3969756/smart-home-

automation-market-set-to-hit-new-highs-driven-by-ai

openPR.com

Realty Executives. (2025). The rise of smart homes: Tech-

nology trends in 2025. Realty Executives.

https://www.realtyexecutives.com/blog/the-rise-of-

smart-homes-technology-trends-in-2025Realty Execu-

tives+1Eco Smart Home Pros+1

ResearchGate. (2025). Optimizing energy efficiency in

IoT-based smart home systems. ResearchGate.

https://www.researchgate.net/publica-

tion/387516833_Optimizing_Energy_Efficiency_in_I

oT-Based_Smart_Home_SystemsResearchGate

Sayed, A., Himeur, Y., Alsalemi, A., Bensaali, F., & Amira,

A. (2021). Intelligent edge-based recommender system

for internet of energy applications. arXiv preprint

arXiv:2111.13272. https://arxiv.org/abs/2111.13272

arXiv

ScienceDirect. (2024). Optimizing energy efficiency and

comfort in smart homes through dynamic predictive op-

timization. ScienceDirect. https://www.sciencedi-

rect.com/science/article/pii/S2352484724003202Sci-

enceDirect

ScienceDirect. (2025). A review on energy efficiency, oc-

cupant comfort, and sustainability in smart homes. Sci-

enceDirect. https://www.sciencedi-

rect.com/science/article/abs/pii/S2352710225005844

ScienceDirect

Springer. (2025). Future of energy management models in

smart homes: A systematic literature review. Springer.

https://link.springer.com/article/10.1007/s41660-025-

00506-xSpringerLink

Yang, Q., & Wang, H. (2021). Privacy-preserving transac-

tive energy management for IoT-aided smart homes via

blockchain. arXiv preprint arXiv:2101.03840.

https://arxiv.org/abs/2101.03840arXiv

ICRDICCT‘25 2025 - INTERNATIONAL CONFERENCE ON RESEARCH AND DEVELOPMENT IN INFORMATION,

COMMUNICATION, AND COMPUTING TECHNOLOGIES