S. Sudharsan Deepak 
, S. Saranya, V. M. Sudharsanprakalathan, M. Sabeel Mohamed
Sri Eshwar College of Engineering, Coimbatore, Tamil Nadu, India 
Correspondence to: S. Sudharsan Deepak, sudharsandeepak12@gmail.com
Additional information
- Ethical approval: This study does not involve human participants, animals, or any clinical data. Therefore, ethical approval was not required for this research.
- Consent: This research did not involve human subjects, patient data, or volunteers; therefore, informed consent was not applicable.
- Funding: No external funding was received for the development of this research work. The project was carried out with institutional support from Sri Eshwar College of Engineering.
- Conflicts of interest: The authors declare that there are no conflicts of interest, financial or otherwise, related to this work.
- Author contribution: S. Sudharsan Deepak – Writing – original draft, Methodology; S. Saranya – Writing – review and editing, Supervision; V. M. Sudharsanprakalathan – Software, Resources; M. Sabeel Mohamed – Validation, Software
- Guarantor: S. Sudharsan Deepak
- Provenance and peer-review: This article was unsolicited and was submitted following an invitation through the STEM’25 conference for possible publication in the Premier Journal of Science
- Data availability statement: All data generated or analyzed during this study are included within the manuscript. No external datasets were used. Additional data are available from the authors upon reasonable request
Keywords: ESP32, Experimental analysis, Hazardous gas detection, IoT monitoring, MQ sensors.
Peer Review
Received: 05 March 2026
Last revised: 09 April 2026
Accepted: 10 May 2026
Version accepted: 2
Published: 17 May 2026
Plain Language Summary Infographic
Abstract
The study describes the creation and testing of an Internet of Things hazardous gas detection system, which uses MQ-series sensors together with an ESP32 microcontroller. The system detects hazardous gases and provides real-time alerts through local and wireless mechanisms. Experimental analysis was conducted under controlled indoor conditions using LPG exposure, while evaluating their dynamic response and calibration characteristics together with their recovery behavior and repeatability. The results demonstrate that the sensor exhibits fast response, stable recovery, and consistent output under repeated trials, indicating reliable performance for indicative monitoring applications. The measurements do not meet certified reference standards because they lack calibration, while cross-sensitivity and environmental conditions introduce accuracy problems. The research study presents these limitations together with solutions, which will lead to better gas monitoring systems that meet higher accuracy standards and compliance with official regulations.
Introduction
The industrial sector and technological advances have expanded their development activities, which now produce more environmental pollution that primarily consists of dangerous gases.1 The invisible nature of air pollution makes it difficult to detect without specialized sensing systems, which create a significant risk to human health and environmental safety. Industrial and domestic settings contain hazardous gases, such as liquefied petroleum gas (LPG), methane, ammonia, and hydrogen, which require continuous monitoring to prevent accidents and exposure risks. The current commercial gas detection systems face three main issues, which include high costs, large size, and limited portability that prevent their widespread use.2 Figure 1 shows the hazardous chemical accidents and fatalities that occurred over different time periods to demonstrate the need for effective monitoring and early detection systems. The existing accident-to-fatality ratios show increasing variability, which creates a demand for dependable gas detection systems that provide immediate results.
The project develops an affordable hazardous gas monitoring system which uses MQ-series sensors and an ESP32 microcontroller to create an IoT-based prototype.3 The system detects gas leakage to send real-time alerts through two different methods, which include local and wireless systems.4 The study demonstrates sensor evaluation through experimental tests, which take place in controlled environments, while existing studies focus on building system design. The proposed system uses LPG exposure for experimental analysis, which tests its performance through dynamic response assessment and calibration characteristics evaluation, recovery behavior testing, and repeatability assessment. The system serves as a prototype for monitoring purposes because its development will help identify practical limitations, which include calibration issues and cross-sensitivity, environmental factors, and subsequent gas detection system development for reliable standards-compliant operation. This work focuses on experimentally evaluating system performance using real-time measurements and controlled testing conditions.
Related Works
Various gas detection systems based on IoT technology use MQ-series sensors together with microcontrollers to create affordable monitoring systems.5 The systems use threshold-based detection systems to generate alerts, which they transmit through basic interfaces that include buzzers and displays and wireless notifications. Real-time gas monitoring systems, which these methods demonstrate as operational, fail to implement proper calibration methods and controlled testing experiments.6 The testing of current studies appears constrained because researchers have not established testing conditions, which prevents them from assessing how sensors perform in terms of response time, recovery behavior, and repeatability. The absence of certified gas detection system performance reports makes it challenging to determine how dependable and usable these systems actually are.
The microcontroller-based systems of the past, which included 89C51-based systems, dedicated their efforts to detecting gas leaks and creating alarms, but they did not develop their systems for remote monitoring and performance assessment. The new IoT systems provide better connection abilities than before, but their user experience suffers from problems related to sensor precision, environmental changes, and gas interaction. The researchers need to conduct thorough testing, which involves systematic experiments on low-cost gas detection equipment to find out how these systems function in controlled experimental settings.7 The study investigates MQ-series sensor performance through analysis of their dynamic response, calibration characteristics, and recovery and repeatability capabilities.
System Overview
The qualities of the proposed design gas leakage detection system owe some predetermined and functional features to the former systems. The proposed system extends conventional gas detection designs by integrating low-cost sensors with an ESP32 microcontroller for real-time monitoring and wireless communication.7 System design perspective is a core ESP-32 microcontroller, which is Wi-Fi8 and Bluetooth enabled, to provide better communication and processing for gas leaking stations.2
Sensor Limitations and Considerations
The gas detection systems under consideration mainly monitor hazards by MQ series sensors,9 for example, MQ-2, MQ-135, but these sensors are attached to various inherent limitations. The first of these is the sensitivity of these sensors to environmental parameters, such as temperature and humidity, which also affects the accuracy of detection. They also generally require warm-up time and often quite frequent calibration for reliable operation in the long term. Problems can also manifest themselves in the way that the many types of gases can cause cross-sensitivity to each other; ambiguous readings may occur without careful software or hardware-level filtering signal conditioning. They may also degrade in the long term, affecting the lifespan and consistency of sensors. Therefore, these limitations must be recognized for improvement in robustness potential and reliability in real-world deployments.
Gas Detection
The design uses the MQ-6 and MQ-135 gas detection system. The MQ-6 sensor can detect LPG, methane, and propane gases in a concentration range of 200 to 10000 ppm. The MQ-135 sensor is suitable for detecting ammonia and hydrogen in the range of 10–1000 ppm. All these sensors have an output in the form of an analog voltage from 0 to 1024, which is also processed by the ESP-32 for measuring the level of the gas in the air.
Table 1 compares hazardous gas detection systems along four primary facets: application environment, types of gases that can be detected by each system, sensor types used, and microcontroller platform. By comparing these parameters, the table enables users to quickly grasp distinctions as well as coverage of each method. For example, some systems are oriented toward industrial or residential environments, whereas others are specialized for applications such as sewage protection or ecosystem monitoring. The selective use of sensors and microcontrollers, such as MQ-series sensors, and trending boards, such as Arduino or ESP32, directly affects the detection of particular gases. This comparison illustrates how each system is customized to detect the most pertinent gases of its intended application so that hardware and software selection are appropriate to prospective protection requirements.
| Table 1: Comparative analysis of hazardous gas detection systems. | ||||
| Application (Multi-Use) | Gases Detected | Sensors Used | Platform/Microcontroller | Reference/Context |
| Industry, home, education | LPG, Methane (CH₄), Ammonia (NH3), Hydrogen (H2) | MQ-6, MQ-135 | ESP32 | Sewage/worker safety |
| Sewage, worker safety | Methane (CH4), Carbon monoxide (CO), Ammonia (NH3), Hydrogen sulfide (H2S) | MQ-4, MQ-135, DHT11 | Arduino UNO | Industrial, chemical/environmental |
| Smart city, indoor/outdoor | Carbon dioxide (CO2), Carbon monoxide (CO), Hydrogen (H2), Methane (CH4) | MQ-135, MQ-2, MQ-7, MQ-8, | ESP32 | Industry/fire/leak detection |
| Industry, fire, leak detection | Carbon monoxide (CO), LPG, Butane, Propane, Ammonia (NH3), Smoke | MQ-2, MQ-3, MQ-6, MQ-135 | Arduino | — |
Warning System
The system has piezo buzzers that are activated in case the normal concentration of any gas exceeds the set limits. thus, making the people who are at the site aware of the situation. The specified limits on each of the gases have been set above levels that are safe to allow the taking of necessary measures before the levels of gas become irresponsible.
Remote Communication
Thanks to the incorporation of wireless features with ESP 32, it is possible to send alerts to the system via SMS or any IoT-based app in real time, alerting those designated. In this case, the factors in question will directly inform the designated personnel about the threat posed by dangerous gases, even if they are outside the poisonous gas leak area. The current draft allows for the delivery of notifications solely via Wi-Fi through a web dashboard. The possibility of SMS delivery is contemplated as a future extension using an online gateway service, but is not implemented.
Scalability and Flexibility
The basic and primary objective of the system is to save human life, and therefore, such systems are required to be compact, lightweight, and inexpensive for industrial, domestic, and commercial applications. The RT frame is lightweight, but, being a box structure, it is modular. This means that the augmentation of their mass with more sensors or the addition of other sensors’ mass in use can be readily done.
System Design
The hazardous gas monitoring system uses MQ-series gas sensors together with an ESP32 microcontroller.7 The microcontroller includes Wi-Fi and Bluetooth wireless communication capabilities. The system operates to identify dangerous gas leaks while delivering immediate notification through its local and remote monitoring systems (Figure 2).
The system uses MQ-6 and MQ-135 sensors for gas detection. The MQ-6 sensor detects LPG, methane, and propane gases within the range of 200 to 10000 ppm, while the MQ-135 sensor detects ammonia and hydrogen in the range of 10 to 1000 ppm. The sensors generate analog voltage signals, which increase with rising gas levels.9 The ESP32 processes these signals through its integrated analog-to-digital converter (ADC) system. The hardware interfacing of the proposed system is illustrated in Figure 3.
Sensor Limitations
MQ-series sensors demonstrate measurement inaccuracies because they become sensitive to changes in temperature and humidity conditions.10 The sensors need a warm-up time to reach stable operation conditions, while they must undergo periodic calibration procedures to maintain proper functioning during extended use. The sensors detect multiple gas types which creates a problem because cross-sensitivity between gases produces uncertain results. The sensors experience performance degradation throughout their lifespan, which results in reduced accuracy. The limitations of the system need to be tested through experiments because they affect its ability to function in actual environments.
Alert and Communication System
The system includes a local alert mechanism using a buzzer, which is activated when gas concentration exceeds predefined threshold levels. The ESP32 device establishes wireless connections to transmit sensor information to an online monitoring system, which displays data on a remote dashboard. The current notification system uses a dashboard interface that operates through Wi-Fi connections. The system might receive future updates that will implement SMS alert capability through third-party gateway services.
System Characteristics
The system is designed to be compact, low-cost, and modular, which enables its use across industrial sites, home environments, and environmental monitoring systems. The modular design enables users to add sensors according to their requirements, which provides them with adaptability in various operational scenarios.
Challenges and Future Directions
Power Consumption
- The ESP32 microcontroller uses power-saving technology, but its power usage increases when gas sensors and wireless communication systems work continuously. This problem makes it difficult to operate in locations that lack reliable electricity.
- The approximate power consumption distribution of system components is illustrated in Figure 4. The MQ-series sensors use excessive power because their heating elements consume energy, while the ESP32 and its additional parts use power at a lower rate.
- The values represent short-term measurements combined with datasheet estimates, which should serve as general indicators instead of exact measurements.
- Upcoming developments will aim to reduce energy consumption through the implementation of duty cycling methods, together with low-power communication techniques and improved sensor management strategies.11
Sensor Calibration and Environmental Influence
The MQ-series sensors require calibration to be performed at regular intervals so they can keep their correct measurements. Temperature and humidity conditions in the environment create strong impacts on sensor accuracy, which leads to errors in measurement results. Gas cross-sensitivity creates reading confusion that requires both signal filtering and calibration methods to solve. The upcoming research will investigate adaptive calibration methods together with compensation techniques, which will enhance system dependability.
Safety and Practical Limitations
The developed system functions as a prototype system, which provides monitoring capabilities but lacks certification for use in safety-critical environments. The system fails to meet industrial safety requirements, which include ATEX and IECEx standards. Future development will work toward achieving certified component integration, which includes protective enclosures and fail-safe systems to secure operational use in dangerous work environments.
Future Enhancements
The project will develop better solutions through its IoT system connections, which provide real-time data analysis and predictive maintenance capabilities and remote notification functions. The research will focus on improving system durability, precise measurement capabilities, and the system’s capacity to expand.
Methodology
Data Collection and System Operation
The hazardous gas detection system was designed to monitor LPG, methane, ammonia, and hydrogen using MQ-series sensors and an ESP32 microcontroller.4 The system includes MQ-6 and MQ-135 sensors. These sensors generate analog signals proportional to gas concentration. The ESP32 processes the sensor outputs through its ADC. This enables continuous monitoring of gas levels. The system activates a buzzer for local alerting when the measured values exceed predefined threshold levels, and it sends data to a web-based interface for remote monitoring (Figure 5). The system performs sensor data collection every 500 milliseconds, which it uses to measure concentration levels and compare them against threshold limits for assessing dangerous and safe conditions.
Calibration and Measurement Considerations
The MQ-series sensors used in this prototype were operated using baseline values suggested by the manufacturer. The threshold levels were established through initial field tests instead of using standardized calibration methods. The testing process used non-certified gas chambers, which do not include reference instruments and laboratory calibration setups. Therefore, the obtained readings should be interpreted as indicative trends rather than precise gas concentration measurements.
System Architecture
The system hardware architecture includes an ESP32 microcontroller connected to MQ-6 and MQ-135 gas sensors, along with a buzzer for alert generation. The sensors deliver their outputs as analog signals, which the ESP32 processes to establish gas concentration levels. The block diagram of the system architecture is shown in Figure 6.
Experimental Procedure
The experimental evaluation of the proposed hazardous gas detection system was conducted using a controlled prototype setup under indoor conditions. The system used MQ-6 and MQ-135 sensors together with an ESP32 microcontroller, which established its initial measurements in clean ambient air. The gas source for hazardous conditions simulation used Liquefied Petroleum Gas (LPG), which was obtained from a standard lighter. The sensor received gas releases that occurred from a fixed distance of approximately 5 cm during controlled exposure periods of 5, 10, 15, and 20s.
The sensor outputs were recorded at regular intervals of 500 milliseconds through the ESP32 system which used its 12-bit ADC to measure data. The researchers repeated each experiment several times to verify results, and they used average measurements for their study. The sensor went through a restoration process after gas exposure when it stayed in ambient air, and researchers recorded the sensor output decrease to measure its recovery ability. All experiments were conducted at room temperature (approximately 27°C) and relative humidity levels of 50%–60%.
Experimental Results and Discussion
Dynamic Response Analysis
The dynamic response of the MQ-6 sensor was evaluated by exposing it to LPG gas and recording the sensor output over time. The metal-oxide semiconductor gas sensor response shows two stages, which begin with a fast increase and end with a stable state. The response behavior is modelled using an exponential function:
S(t) = A(1 − e^(−t/τ)) + C (1)
The equation describes sensor output through S(t), while τ represents the response time constant and A functions as the signal’s maximum value, and C acts as the initial baseline measurement. The fitted model shows good agreement with the observed data, which demonstrates that the system reaches its maximum gas response within a brief period, followed by a period of stable operation (Figure 7).
Calibration Characteristics
The research team used a logarithmic calibration method to study how sensor output responds to different gas concentration levels. The experimental data exhibit a near-linear trend on a log–log scale, which demonstrates a power–law relationship between gas concentration and sensor response. The calibration model can be expressed as:
ADC ∝ (ppm)^n (2)
The sensitivity exponent of the sensor appears as n in this equation. The fitted calibration curve shows consistent behavior across the tested concentration range, enabling approximate estimation of gas concentration from sensor output values. The process of conducting precise calibration requires the use of controlled gas environments, together with certified reference standards. The calibration characteristics are illustrated in Figure 8.
Recovery Analysis
The evaluation of MQ-6 sensor recovery characteristics occurred through gas source elimination and continuous sensor output monitoring. The response pattern shows an exponential decay function, which demonstrates that gas molecules are slowly leaving the sensor. The recovery behavior can be modelled as:
S(t) = A e^(−t/τ) + C (3)
The formula S(t) defines sensor output, while τ represents the recovery time constant and A stands for amplitude and C indicates the baseline offset, which remains after measuring. The fitted model shows good agreement with the observed data, confirming that the sensor gradually returns to its baseline condition after gas exposure. The gas sensors using metal-oxide semiconductor technology displayed recovery times that exceeded response times according to industry standards. The recovery characteristics are illustrated in Figure 9.
Repeatability Analysis
Repeatability tests were conducted by multiple experiments, which used the same testing conditions. The sensor output values obtained across different trials exhibited minimal variation, indicating stable and consistent performance. The mean sensor output and standard deviation were calculated to quantify measurement variability. The coefficient of variation was determined to measure how sensor readings spread in relation to the total measurement of sensor readings. The sensor demonstrates dependable performance through its measurements because both standard deviation and coefficient of variation remain low, which confirms measurement accuracy during continuous monitoring tests. The repeatability characteristics are illustrated in Figure 10.
Table 2 displays the timing characteristics, which measure the system’s timing performance. The values represent the time required for the ESP32 firmware to detect threshold crossings, process alert logic, and activate the buzzer and notification mechanisms.
| Table 2: Time calibration for alerts. | |||
| Gas Type | Detection Time (ms) | Alert Activation Time (ms) | Response Time (ms) |
| LPG | 450 | 560 | 110 |
| Methane | 490 | 590 | 100 |
| Ammonia | 520 | 610 | 90 |
| Hydrogen | 530 | 640 | 110 |
| Carbon monoxide | 480 | 590 | 110 |
The values represent software response times because they do not account for the basic response time of MQ-series sensors. The actual MQ sensors need extra time to adsorb gas and reach stability because laboratory testing is required to obtain accurate measurements of these processes. Overall, the results confirm that the system provides reliable indicative monitoring performance under controlled conditions (Figure 11).
Conclusion
The research developed and tested a low-cost Internet of Things hazardous gas monitoring system, which uses MQ-series sensors and ESP32 microcontroller technology. The system successfully detected hazardous gases while delivering real-time warnings through both its local system and online web interfaces. The system demonstrated quick response times, which it maintained through stable operation in indoor settings while conducting tests that showed the same results multiple times. The calibration analysis established a predictable relationship between sensor output and gas concentration, which allowed safe monitoring of hazardous gas levels.
The system lacks calibration to certified reference standards, while its performance depends on environmental factors and sensor cross-sensitivity. The prototype serves as a monitoring tool that delivers initial results but lacks the necessary features for safety-critical applications. The study presents practical knowledge about how low-cost gas detection systems operate while demonstrating their limitations and establishing experimental validation as essential for reliability improvement and future development guidance.
Future Enhancements
The upcoming research will concentrate on enhancing the system’s precision and dependable operation and its capacity to handle growing demands. The primary focus area involves sensor calibration work, which uses controlled gas environments and certified reference standards to boost measurement accuracy. The research will investigate power optimization methods, which include duty cycling and low-power communication techniques to achieve better energy efficiency during extended operational periods. The implementation of environmental compensation techniques will assist in reducing temperature and humidity impacts on sensor performance.
The project will develop additional features, enabling integration with cutting-edge IoT platforms to deliver real-time data analysis and predictive system monitoring and remote operation control. The implementation of industrial-grade sensors, together with selective sensors, will enhance detection performance, which enables usage in critical safety environments.
References
- Chen C, Reniers G. Chemical industry in China: the current status, safety problems, and pathways for future sustainable development. Saf Sci. 2020;128:104741. https://doi.org/10.1016/j.ssci.2020.104741
- Sivasangari A, Varalakshmi V, Mishra R, Murali K. IoT-based detection and alerting of hazardous gas detection for the welfare of sewer labourers. Nanotechnol. Percept. 2024;20(S4):39–49. https://doi.org/10.62441/nano-ntp.v20is4.4
- Flores-Cortez O, Cortez R, González B. Design and implementation of an IoT-based LPG and CO gases monitoring system. arXiv. Preprint. 2021. https://doi.org/10.48550/arXiv.2107.07406
- Gupta R, Singh A, Verma P. Design and development of a
low-cost IoT-based gas monitoring system for industrial safety. J Ind Inf Integr. 2023;35:100456. https://doi.org/10.1016/j.jii.2023.100456 - World Health Organization. WHO global air quality guidelines: particulate matter (PM2.5 and PM10), ozone, nitrogen dioxide, sulphur dioxide and carbon monoxide. WHO Press; 2021.
https://www.who.int/publications/i/item/9789240034228 - Praveenchandar J, Vetrithangam D, Kaliappan S, et al. IoT-based harmful toxic gases monitoring and fault detection using deep learning techniques. Sci Program. 2022;2022:7516328.
https://doi.org/10.1155/2022/7516328 - Easterline LM, Putri AAZR, Atmaja PS, Dewi AL, Prasetyo A. Smart air monitoring with IoT-based MQ sensors using ESP32.
Procedia Comput Sci. 2024;245:815–824. https://doi.org/10.1016/j.procs.2024.10.308 - Saranya S, Sharmila B, Shashank RS, Nesasudha M, Karthikeyan TA. Optimized substrate integrated waveguide (SIW) antenna for wearable health monitoring applications. 2024 5th International Conference on Smart Sensors and Application (ICSSA). IEEE;2024:1–6. https://doi.org/10.1109/ICSSA62312.2024.10788602
- Dey A. Semiconductor metal oxide gas sensors: a review. Mater. Sci Eng. B. 2018;229:206–217. https://doi.org/10.1016/j.mseb.2017.12.036
- Bhattacharjee D, Ghosh S. Environmental factors affecting gas sensor performance and calibration techniques. Sens Actuator B. 2021;345:130344. https://doi.org/10.1016/j.snb.2021.130344
- Al-Fuqaha A, Guizani M, Mohammadi M, Aledhari M, Ayyash M. Internet of things: a survey on enabling technologies, protocols, and applications. IEEE Commun. Tutor. 2015;17(4):2347–2376. https://doi.org/10.1109/COMST.2015.2444095
