A Low-Power Portable Gas Sensor System with Adaptive ROIC and Wi-Fi Communication for Biomedical Applications

Abstract

This study presents a portable gas sensor system that achieves high performance while minimizing power consumption and production costs for biomedical applications. The proposed system integrates a low-power readout integrated circuit (ROIC) capable of processing large-amplitude sensor signals using a 1.2 V ADC, significantly reducing the power consumption compared with conventional high-voltage solutions. To address the inherent limitations of single-core/single-thread microcontrollers, an optimized Wi-Fi communication algorithm is implemented, enabling real-time data transmission and accurate signal reconstruction without data loss. Experimental validation using a hydrogen gas detection setup demonstrated that the system achieves less than 0.15% error in reconstructed signals, while substantially reducing overall power consumption and component cost. Comparative analysis confirms that the proposed approach achieves a performance comparable to conventional systems while offering significant reductions in energy usage and hardware expense. These results demonstrate the feasibility of a scalable, low-cost solution for portable gas sensing, particularly in biomedical applications requiring precise and reliable monitoring.

1. Introduction

The detection and analysis of gases are essential across a wide range of fields, including industrial safety, environmental monitoring, automotive technology, space exploration, and biomedical applications [1]. Among these, gas sensor systems play a particularly vital role in the biomedical domain, enabling the identification of specific gases that act as biomarkers for various diseases [2,3]. For instance, certain volatile organic compounds (VOCs) emitted by patients can serve as indicators of conditions such as diabetes or cancer, facilitating early diagnosis and timely intervention [4]. Similarly, the detection of toxic gases such as ammonia, hydrogen sulfide, and methane within the human body can assist in diagnosing gastrointestinal disorders and predicting related health conditions [5].

The development of portable gas sensor systems that balance cost-effectiveness, operational efficiency, and functional versatility is critical in addressing such urgent medical challenges [6,7]. These systems offer a promising solution by enabling precise diagnostics and real-time health monitoring in a more accessible and sustainable manner. They hold significant potential to transform healthcare delivery by making diagnostics affordable, scalable, and broadly deployable [8].

However, designing portable gas sensor systems poses two primary challenges. From a technical perspective, achieving low-power operation without compromising sensitivity and accuracy is paramount. From a commercialization standpoint, it is equally important to ensure cost-effective and robust system design. Addressing these challenges necessitates a holistic approach that includes in-depth signal analysis, the optimization of individual components, and system-level operational considerations [9,10].

To meet these demands, prior work [11] has shown that chopping large gas sensor signals into smaller components and reconstructing them via straightforward digital post-processing can effectively optimize the performance of analog-to-digital converters (ADCs) [12,13], which are traditionally among the most power-intensive and cost-constraining elements in sensor systems. Although these techniques alleviate some design constraints, they often fall short in addressing broader system-level requirements such as integrating communication modules and managing control under strict power and cost limitations.

Building on these advancements, this study introduces a comprehensive design methodology that extends beyond signal acquisition to encompass the full gas sensor system. A novel readout architecture is presented to efficiently process large sensor output signals while significantly easing ADC performance requirements. This innovation reduces both power consumption and the system cost, while maintaining sensitivity and accuracy. Moreover, a cost-efficient communication interface is implemented using a single-core microcontroller unit (MCU) [14] for real-time data handling. By analyzing the output signal patterns, the proposed method enables the use of lower-specification MCUs without sacrificing system performance, eliminating the need for high-end dual-core solutions. This reduces system complexity and the overall cost, improving scalability and applicability across diverse domains.

The proposed gas sensor system addresses key limitations in portable sensor design. It preserves sensitivity while relaxing the demands on core components such as ADCs and MCUs. Through efficient signal pattern analysis, it supports real-time data handling in low-specification hardware, achieving notable reductions in power consumption and component cost. These attributes make the system highly practical and competitive particularly in biomedical applications requiring compact, low-power, and low-cost solutions. Furthermore, the modular design of the proposed system supports a wide range of sensor types and applications. In this study, the system was prototyped and evaluated at the module level, with a particular emphasis on biomedical applications. The evaluation results demonstrate strong potential for commercialization, particularly in health diagnostics and real-time condition monitoring, while ensuring cost-efficiency and scalability.

The remainder of this paper is organized as follows. Section 2 describes the prototype structure of the proposed portable gas sensor system and the associated challenges. Section 3 presents the proposed ROIC design and Wi-Fi communication scheme. Section 4 reports the measurement results from the prototype. Finally, Section 5 concludes the paper with a summary of findings and a discussion of potential applications.

2. Prototype Structure of Portable Gas Sensor

The proposed portable gas sensor system was developed to address the requirements of miniaturization, low power consumption, and wireless communication for biomedical and environmental applications. Specifically designed for real-time monitoring in the biomedical field, this system combines cost-efficiency and performance to provide a practical solution for gas detection and analysis. The system integrates several critical components to achieve its functionality. It employs a commercially available Figaro TGS-2600 gas sensor [15], which detects target gases and converts their concentrations into electrical signals, ensuring precise and reliable environmental monitoring. The readout circuit, designed using the proposed readout integrated circuit (ROIC) methodology, enables efficient signal acquisition and processing while significantly reducing power consumption. This circuit is optimized to handle large gas sensor outputs effectively, ensuring the system’s reliability. To manage control and communication, a MCU [14,16] is utilized. The MCU oversees the operation of the ROIC, reconstructs digital signals, and manages wireless communication via the integrated Wi-Fi antenna. Through innovative algorithm optimization, the single-core MCU achieves a performance comparable to dual-core systems, effectively reducing hardware costs and energy consumption. The Wi-Fi antenna facilitates real-time data transmission to a custom-developed mobile application, allowing users to monitor gas concentrations, visualize sensor outputs, and receive hazard warnings based on predefined thresholds. The system is powered by a 3.7 V/2000 mAh lithium-ion battery, ensuring low-power operation suitable for portable applications. The mobile application further enhances user experience by providing an intuitive interface for the real-time visualization of gas concentration data, enabling accurate risk assessments and timely feedback. During development, significant challenges such as achieving a low power consumption and minimizing production costs were addressed. High-power and high-cost components like ADCs and MCUs were replaced with energy-efficient and cost-effective alternatives. Despite the use of lower-specification components, the system delivers a performance on par with or exceeding that of conventional high-end designs. This makes it highly practical for a wide range of biomedical diagnostics and environmental monitoring applications, demonstrating scalability and cost-effectiveness.

Figure 1 shows the overall system architecture, which integrates gas sensing, signal processing, wireless communication, and data visualization into a unified and efficient workflow. The architecture consists of four main functional blocks: (1) the gas sensing unit, (2) the ROIC, (3) the MCU, (4) the wireless communication module connected to a mobile application. The gas sensing unit includes a gas sensor and a Resistance-to-Voltage (R-V) converter, which transforms the resistive output of the gas sensor into a proportional voltage signal. The sensor detects target gases such as VOCs and outputs resistance values that vary with gas concentration. These are then converted to voltage signals and passed to the ROIC. The ROIC is designed to acquire these analog signals efficiently while operating under strict power constraints. It integrates key components including a Bandgap Reference (BGR) for stable voltage generation, a Low-Dropout Regulator (LDO) for power regulation, and a Delay-Locked Loop (DLL) with a Voltage-Controlled Oscillator (VCO) for precise clock synchronization. The analog signal is digitized using a low-power Successive Approximation Register (SAR) ADC, chosen for its efficiency in portable systems. The MCU acts as the central processor. It receives digitized data from the ROIC, reconstructs gas concentration values using an optimized algorithm, and performs system control, wireless communication, and data logging. The processed data is transmitted via a Wi-Fi-enabled RF transceiver to the mobile application, which presents real-time gas concentration levels, safety thresholds, and hazard alerts. This interface provides users with a convenient and intuitive platform for continuous gas monitoring and risk management.

3. Proposed ROIC Design and Wi-Fi Communication Scheme

3.1. Conventional Gas Sensor System

Figure 2 illustrates the signal processing flowchart of a conventional gas sensor readout, highlighting the major stages involved in converting the sensor output into digital form. The gas sensor acts as a variable resistor whose resistance varies with gas concentration, generating an output voltage (V_GS). This voltage reflects the chemical reactions occurring on the sensor surface and serves as the input for the readout circuitry. However, V_GS typically includes unwanted noise components such as power supply noise, ground fluctuations, and thermal noise. To mitigate these, a low-pass filter (LPF) removes high-frequency noise, producing a filtered signal (V_{IA_IN}) suitable for amplification. This step is essential to preserve signal fidelity. The filtered signal is then amplified by an instrumentation amplifier (IA), which emphasizes the signal variation relative to a reference voltage. The resulting signal (V_{ADC_IN}) is adjusted to match the ADC’s input range, ensuring the optimal usage of ADC resolution. Finally, the ADC digitizes this analog signal (V_{ADC_IN}) into a digital output (D_OUT), allowing further processing, storage, and communication via a microcontroller or other digital systems. A critical aspect of this signal chain is the alignment of the power supply voltage (VDD) with the operating range of the readout system. Proper VDD selection is essential to preserve the sensor’s full dynamic range. If the VDD is insufficient, the ADC may saturate when the sensor output exceeds its input range, causing signal clipping or loss. While increasing the VDD can mitigate this issue, it comes at the cost of higher power consumption—an unfavorable trade-off for portable applications. Commercial gas sensors often require high operating voltages (up to 15 V) [17], complicating integration with low-power systems. Despite these challenges, sensor development often prioritizes sensing performance without considering system-level constraints, leading to suboptimal integration and power inefficiencies.

Furthermore, semiconductor gas sensors exhibit significant baseline variability between individual units and are sensitive to environmental conditions. To investigate this, a set of 50 TGS2600 gas sensor samples were tested under various temperature and humidity settings. Figure 3a shows the baseline distribution at a hydrogen concentration of 50 ppm and 45% relative humidity, as the temperature varies (10 °C, 20 °C, 30 °C). The x-axis represents initial baseline values; the y-axis represents mean absolute deviation (MAD) in LSBs (12-bit ADC, 3.3 V range). As temperature increases, MAD values become more dispersed: from a maximum of 50.9 LSB (mean: 41.7) at 10 °C, to 60.4 LSB (mean: 76.4) at 20 °C, and up to 99.5 LSB (mean: 124.3) at 30 °C. Similarly, Figure 3b presents the baseline variation under 50 ppm hydrogen and a fixed temperature (20 °C), while humidity varies from 40% to 85%. The MAD increases with humidity: 29 LSB (mean: 47.3) at 40% RH, 60.4 LSB (mean: 76.4) at 65% RH, and 126.9 LSB (mean: 81.7) at 85% RH. These results confirm that both temperature and humidity exacerbate baseline dispersion, introducing signal inconsistency. Figure 3c illustrates a major limitation of conventional portable gas sensor systems: the mismatch between the wide dynamic range of gas sensor outputs and the limited input range of ADCs. When the sensor output exceeds the ADC’s range (D_Min to D_Max), signal clipping occurs, resulting in data loss and degraded accuracy. Expanding the ADC range could mitigate this, but would necessitate higher supply voltages, leading to increased power consumption and cost. Moreover, such expansion often requires performance trade-offs. An alternative approach is to scale down the sensor output to fit within the ADC’s input range. However, this compromises the sensor’s sensitivity and resolution, limiting its ability to detect subtle concentration changes. Therefore, conventional methods either expanding the ADC input range or reducing the sensor signal magnitude struggle to maintain a high sensitivity under low-power constraints. These observations clearly underscore the need for advanced readout architectures that minimize data loss, preserve sensitivity, and maintain system efficiency across environmental variations and wide signal ranges.

3.2. Proposed Sensor System Structure

Figure 4 illustrates the architecture and operational principle of the proposed ROIC, which was designed for efficient signal acquisition from gas sensors. The ROIC is composed of two main components: an IA and an adaptive reference generator. The adaptive reference generator comprises four sampling switches (two Ø_RS1 and two Ø_RS2) and two sampling capacitors (C_S1 and C_S2). It operates in alternating odd [2n − 1] and even [2n] phases, enabling the continuous generation of reference voltages tailored to the input signal trajectory. The IA’s gain is programmable through a 3-to-8 decoder that selects among multiple gain resistors (R_IG), allowing dynamic adjustment to accommodate various signal levels. Internally, the IA adopts a three-op-amp configuration, amplifying the difference between the filtered sensor output (V_SO) and the reference voltage (V_REF) generated by the adaptive reference generator, using a common-mode voltage (V_CM) as the baseline. The amplified signal (V_{ADC_IN}) is then delivered to the ADC for digitization. The adaptive reference generation process operates as follows. During the odd phase (Ø_RS1), C_S1 samples V_SO to generate V_{REF [2n−1]}, while C_S2 tracks V_SO in preparation for the next even phase. During the even phase (Ø_RS2), C_S2 samples V_SO to generate V_{REF [2n]}, while C_S1 tracks V_SO to prepare V_{REF [2n−1]} for the subsequent odd phase. This alternating mechanism produces reference voltages that continuously adapt to the input signal, enabling the IA to extract the differential signal ΔV_SO = V_SO − V_REF with high temporal precision. This dynamic signal centering method significantly reduces the input range requirement of the ADC. As a result, a low-power ADC with a narrow input range can be employed, which decreases the overall power consumption and system cost. Furthermore, this approach preserves the original sensitivity of the gas sensor, as it avoids signal attenuation through analog scaling. The IA is equipped with an internal gain controller that enables flexible amplification or attenuation. This is implemented via a resistor-based digital-to-analog converter (DAC) structure [18], offering 3-bit control and eight selectable gain levels. This programmability enhances the adaptability of the ROIC across a variety of gas sensor types and signal magnitudes. Once the IA outputs the conditioned signal, it is digitized by the ADC and transferred to the MCU. A simple digital signal reconstruction algorithm is used to recover the original waveform from the digitized differential data. The following are illustrated in Figure 5:

• Figure 5a: Shows the timing diagram of the two-phase sampling process. During each clock cycle (t_n), the adaptive reference generator alternates between Ø_RS1 and Ø_RS2, generating V_REF[n] for each respective phase.
• Figure 5b: Illustrates the computation of ΔV_SO[n] = V_SO[n] − V_REF[n] at each phase. These values are then digitized and sent to the MCU.
• Figure 5c: Describes the reconstruction algorithm. If the previous data point at t_n−1 is D_[n−1], and the differential data at t_n is ΔD_[n], the reconstructed signal D_[n] is obtained by D_[n] = D_[n−1] + ΔD_[n]. This process iterates over successive samples, reconstructing a continuous waveform.

The main advantage of this method is its ability to reduce ADC dynamic range requirements without sacrificing signal fidelity. By adaptively referencing the input signal, the need for a high-range ADC is eliminated, resulting in substantial reductions in both power consumption and circuit complexity. This technique is particularly beneficial in gas sensor applications, where signal variations can be large and unpredictable. By integrating the proposed ROIC and the reconstruction algorithm, the system achieves enhanced energy efficiency, improved signal accuracy, and a simplified overall architecture, making it well-suited for portable, low-power, and cost-sensitive environments.

Figure 6 presents the overall block diagram of the proposed gas sensor system, detailing the signal processing flow and the roles of each subsystem. The system comprises two primary components: the ROIC and the digital control logic implemented within the MCU. These components work in concert to ensure accurate signal acquisition, minimal power consumption, and reliable digital reconstruction. The signal chain begins with the gas sensor output (V_SO), which is first passed through the LPF to suppress high-frequency noise and external interference. The filtered signal is then processed by the adaptive reference generator, which alternates between odd and even sampling phases to dynamically generate reference voltages (V_REFn). This mechanism allows the system to track variations in the sensor baseline and compress the signal swing prior to amplification. Next, the IA amplifies the differential signal ΔV_SO, effectively centering the input relative to the reference and preparing it for A/D conversion. The conditioned signal is digitized by a low-power ADC and transmitted to the MCU via a parallel data bus. Within the MCU, the digitized output is first stored in SRAM, where it is processed using a lightweight digital reconstruction algorithm. This algorithm incrementally restores the original waveform by summing the differential data ΔD_[n] with the previously reconstructed point D_{[n−1, m]}, thereby preserving the dynamic behavior of the sensor output. The reconstructed signal is then processed by a digital common generator and temporarily buffered before being transferred to a storage unit or communication interface for downstream processing. To ensure synchronized operation among the subsystems, the system utilizes a control bus architecture. The adaptive reference generator control unit produces the timing signals necessary for interleaved sampling, while the ADC control unit handles precise conversion timing. Additional units including the SRAM control and logic control blocks coordinate memory access and data transfer within the MCU. A system-wide master clock is provided by an on-chip oscillator (OSC) and phase-locked loop (PLL), which deliver stable and precise timing for all functional blocks. By tightly integrating the adaptive analog front-end with a low-power digital reconstruction flow, the proposed system effectively reduces the required ADC dynamic range and power consumption.

3.3. Proposed Wi-Fi Communication Scheme

Figure 7 illustrates the timing diagrams for Wi-Fi communication and data readout across various MCU configurations: single-core/single-thread, single-core/multi-thread, and dual-core/multi-thread architectures [14,16]. These diagrams highlight the limitations associated with task concurrency in gas sensor systems and their impact on data integrity. In single-core/single-thread MCUs, Wi-Fi communication and data readout are executed sequentially. This constraint prohibits parallel task execution, often resulting in data loss when communication overlaps with signal readout. As shown in the timing diagram, communication delays commonly coincide with readout intervals, leading to unrecoverable signal loss. Consequently, this architecture is inadequate for real-time applications requiring continuous, accurate data acquisition. To mitigate these issues, single-core/multi-thread configurations allow the interleaved execution of tasks. This reduces data loss compared with single-threaded systems but still suffers from occasional contention due to the single-core limitation. As a result, performance remains suboptimal for applications requiring a high data throughput and minimal latency. In contrast, dual-core/multi-thread MCUs allocate separate cores to Wi-Fi communication and data readout, enabling true parallel processing. This architecture effectively eliminates data loss, as shown in the timing diagram, making it ideal for real-time sensing. However, dual-core solutions incur increased power consumption and production costs, which are less favorable for portable and energy-constrained systems. Although the proposed ROIC facilitates low-voltage signal extraction within a narrow dynamic range, the simultaneous execution of signal reconstruction and data transmission remains critical for maintaining system accuracy. The reconstruction algorithm relies on the incremental accumulation of extracted signal components. If the MCU fails to execute both signal processing and communication concurrently, the loss of critical data points, particularly the reference or boundary values, can significantly degrade the quality of reconstructed signals, compromising real-time monitoring. To address this challenge, the proposed communication scheme restructures task execution within the MCU to support single-core/single-thread operation without data loss. By optimizing task scheduling and reducing communication latency, the system ensures reliable signal reconstruction even under sequential processing constraints.

Figure 8 illustrates the proposed compensation strategy for mitigating data loss during Wi-Fi communication in single-core/single-thread MCUs. In conventional implementations, Wi-Fi communication is initiated at the beginning of each readout cycle, as shown in Figure 8a. This timing overlaps with signal acquisition, causing the loss of initial or mid-cycle data points. Since the reconstruction algorithm relies on the initial sample of each cycle as a reference voltage (V_CM), missing this value prevents the accurate calculation of differential signal components (ΔV_SO), introducing significant reconstruction error. Despite this, the ROIC possesses an inherent compensatory feature: since signal extraction is performed relative to V_CM, subsequent samples still reflect accumulated changes even when intermediate values are missing. Building on this characteristic, the proposed strategy reschedules Wi-Fi communication to occur during the middle of each readout cycle, as illustrated in Figure 8b. This preserves the first and last data points, which are critical for maintaining signal continuity and reference alignment. To further address mid-cycle data loss, interpolation techniques are employed to estimate missing values. The waveform comparison in Figure 8b confirms that the proposed method yields a reconstructed signal closely matching the actual sensor output. In contrast, the conventional method shown in Figure 8a exhibits pronounced deviations due to unmitigated data loss. By strategically shifting the communication window and integrating interpolation-based compensation, the proposed scheme enables accurate signal reconstruction without requiring additional hardware resources. This ensures that low-power, low-cost single-core/single-thread MCUs remain viable for portable gas sensing applications, preserving energy efficiency without sacrificing signal fidelity.

4. Measurement Results and Discussion

Figure 9a illustrates the overall measurement system configuration and the flowchart of the testing procedure, while Figure 9b presents the actual experimental setup used to evaluate the proposed system’s hydrogen gas detection performance. The test environment comprises several key components: a mass flow controller (MFC) for gas flow regulation, a bubbler for humidity control, and a sealed chamber capable of accommodating multiple gas sensors simultaneously. The system uses hydrogen gas and clean air, with gas injection timing controlled via solenoid air valves managed by a system control PC. The PC also regulates the MFCs to precisely control gas concentrations inside the chamber. Humidity control is achieved using two separate clean air lines, each equipped with its own MFC. One pathway is connected to a bubbler, introducing moisture into the air stream. The humidified and dry air flows are then combined in controlled proportions to achieve the desired humidity level. Hydrogen gas is subsequently introduced and mixed with the conditioned air within a mixer unit, allowing the precise control of gas concentration, humidity, and clean air dilution. In this experiment, the hydrogen flow rate was set to 100 sccm, and the gas concentration was fixed at 100 ppm. Gas was injected for 500 s, followed by a 1200 s fresh air purging cycle to restore baseline conditions. All measurements were conducted in an indoor environment maintained at approximately 24 °C ± 2 °C and 50% relative humidity (RH). A 10 kΩ load resistor was applied to the sensor during testing. The total test duration was 3400 s, encompassing two complete sensing cycles. Also Figure 9b used the custom mobile application developed for the system, which enables real-time signal acquisition and user notifications. The app provides key functionalities such as live hydrogen concentration monitoring, threshold-based alert messaging, and automatic data logging for long-term tracking. This integration confirms the system’s ability to support real-time wireless communication and user feedback under practical usage scenarios. The PCB-level prototype used in the experiments measures 50 mm in width and 23 mm in height, as shown alongside a 19.5 mm coin for scale. Despite incorporating both the sensor readout circuit and the wireless communication module, the prototype maintains a compact form factor, demonstrating suitability for portable gas sensing applications.

Figure 10 shows the measured transient response of the proposed gas sensor system under controlled environmental conditions (50% RH, 24 °C). The experiment evaluates the system’s dynamic behavior during hydrogen gas injection and its capability to extract accurate gas concentration data in real time. Figure 10a presents the output of the instrumentation amplifier (V_IA) during the injection phase. The baseline common-mode output voltage is approximately 0.74 V, with a noticeable voltage drop (ΔV_IA) occurring in response to hydrogen exposure. This behavior corresponds to the alternating operation of the adaptive reference generator, which switches between odd and even readout phases at 3 s intervals. Following the end of the gas injection period, the introduction of fresh air leads to a gradual increase in V_IA, indicating the sensor’s recovery and gas clearance. Figure 10b displays the real-time digital output (D_OUT) reconstructed by the system. The waveform shows a clear separation between the gas injection and air purge phases, accurately reflecting the transient characteristics of the sensor signal. The ΔV_IA values collected during measurement are translated into concentration data through digital reconstruction, verifying that the system provides high signal integrity and real-time monitoring capabilities suitable for practical applications.

Figure 11 provides a comparative analysis of hydrogen gas signal reconstruction under three different system configurations, all evaluated under the same environmental conditions (50% RH, 24 °C). Figure 11a shows the reference signal obtained using a conventional ROIC with a 3.3 V supply. The original hydrogen signal peaks at 3.04 V, demonstrating the full dynamic response of the gas sensor. This configuration requires a high supply voltage to prevent signal clipping and to maintain accuracy. Figure 11b demonstrates the limitations of a commercial ROIC operating at a reduced 1.2 V supply. Due to the constrained ADC input range, parts of the sensor signal exceed the dynamic range and become saturated. This leads to significant signal distortion and loss. For example, a 100 ppm hydrogen gas input may be interpreted as a 5 ppm signal, resulting in critical concentration estimation errors. These results underscore the challenge of using conventional architectures at low voltages, where signal compression or clipping severely impacts detection accuracy. In contrast, Figure 11c shows the digitally reconstructed signal obtained using the proposed system, also operating at 1.2 V. Unlike Figure 11b, the proposed approach preserves the full sensor response through adaptive signal extraction and reconstruction. The system maintains high fidelity even under low-power conditions, avoiding saturation while accurately tracking gas concentration changes. This comparison confirms the proposed system’s ability to operate with a reduced supply voltage and input range while maintaining signal accuracy. By minimizing the required ADC dynamic range through ROIC-level preprocessing, the system achieves significant reductions in power consumption and hardware complexity. These attributes make the architecture highly suitable for portable and battery-powered gas sensor applications, where energy efficiency and costs are critical.

Figure 12 compares the reconstructed output waveforms obtained using the conventional communication method and the proposed communication algorithm, under identical environmental conditions. This evaluation focuses on signal integrity during wireless data transmission using a single-core/single-thread MCU architecture. Figure 12a illustrates the limitations of conventional Wi-Fi communication scheduling. Due to task conflicts, significant data loss occurs during signal readout, resulting in distorted waveform reconstruction. Specifically, the reconstructed signal exhibits a peak voltage of only 1.85 V, which is 1.19 V lower than the original 3.04 V reference signal, indicating a substantial error that can severely misrepresent the actual gas concentration. Figure 12b, in contrast, demonstrates the effectiveness of the proposed communication scheduling algorithm optimized for single-core/single-thread systems. By avoiding overlap between communication and critical signal acquisition intervals, the system preserves signal fidelity. The reconstructed waveform closely matches the original sensor signal, with a peak voltage deviation of only 0.15%, which is well within acceptable tolerance levels for real-time monitoring. These results confirm that the proposed algorithm reliably prevents signal degradation without requiring additional hardware, making it a practical and efficient solution for resource-constrained embedded systems.

Figure 13a compares two commercially available ADCs operating at 3.3 V and 1.2 V supply voltages [19,20]. The 3.3 V ADC offers a wide input range but consumes 13.2 mW and costs approximately USD 30.44. In contrast, the 1.2 V ADC significantly reduces the power consumption to 0.22 mW and the cost to USD 13.20, but it suffers from a narrower input range, limiting its applicability in systems requiring large signal handling. The proposed ROIC overcomes this limitation by adaptively scaling the sensor signal to fit within the narrow input range of the 1.2 V ADC. As shown in previous experiments, this approach enables the use of low-power ADCs without sacrificing signal fidelity. Thus, high-end ADC performance can be approximated under significantly reduced power and cost constraints, improving overall system-level efficiency. Figure 13b presents a comparison of MCU intellectual property (IP) cores with varying core/thread configurations in terms of power consumption and cost [21,22,23,24,25,26,27,28,29,30,31,32]. Single-core/single-thread MCUs consume as little as 0.506 mW and cost USD 2.80, offering a highly efficient and cost-effective option for portable applications. In contrast, dual-core/multi-thread MCUs consume up to 1320 mW and cost USD 41.91, reflecting the trade-offs associated with higher performance architectures. Although single-core/single-thread MCUs offer exceptional efficiency, their inability to simultaneously handle data acquisition and wireless transmission can result in signal loss. This challenge is critical in gas sensor systems, where the accurate reconstruction of transient signals is essential. The proposed communication algorithm addresses this issue by timing data transmission during low-sensitivity intervals of the signal acquisition cycle. This scheduling ensures that essential data points are preserved and that overall signal fidelity is maintained. As a result, the system can leverage the advantages of low-cost, low-power MCU architectures while achieving performance levels comparable to those of high-end solutions.

Table 1. Performance summary and comparison.

		This Work	Commercial Product		Research Prototype
Index	Ref.	-	[33]	[34]	[35]	[36]
MCU	Product Name	ESP32-H2FH2	STM32F439	ESP32-D0WDQ6	CC2640R2F	BCM2835
	Core/Thread	1/1	1/1	Multi/Multi	Multi/Multi	1/1
	Core Speed	96 MHz	180 MHz	240 MHz	48 MHz	1 GHz
	Wireless Method	Wi-Fi	Wi-Fi	Wi-Fi	BLE 5.1	Wi-Fi
	Power Consumption	85.8 mW	858 mW	1200 mW	27 mW	700 mW
ADC	Product Name	ADS7866	Internal ADC	Internal ADC	Internal ADC	MCP3008
	Resolution	12−bit	12−bit	12−bit	12−bit	10−bit
	Supply Voltage	1.2 V	3.3 V	3.3 V	1.8–3.3 V	5 V
	Input Range	1.2 V	3.3 V	3.3 V	3.0 V	5 V
	Power Consumption	0.22 mW	5.94 mW	-	2.475 mW	2.750 mW
Support Feature	Real Time Monitoring	O	O	O	O	O
	Data Logging	O	O	O	O	O
	Alarm	O	O	O	O	O
	Application Type	Portable Indoor	Indoor	Indoor	Wearable	Indoor

summarizes and compares the performance of the proposed gas sensor system with selected commercial products and research prototypes. The table presents key specifications, including MCU architecture, ADC characteristics, wireless communication methods, power consumption, and application suitability. The commercial product [33] adopts the same single-core/single-thread MCU architecture as the proposed system but operates at nearly twice the clock speed (180 MHz vs. 96 MHz). Although it utilizes a 12−bit ADC, its 3.3 V input range leads to a significantly higher power consumption compared with the proposed 1.2 V implementation. While functionally similar, this product is less favorable for portable or battery-powered environments due to its energy profile. Another commercial product [34] employs a multi-core/multi-thread MCU configuration, enabling concurrent signal conversion and data transmission. However, this architectural benefit comes at the cost of considerably increased power consumption, rendering it unsuitable for low-power sensing applications. Research prototypes [35] incorporate Bluetooth Low Energy (BLE) for wireless communication. BLE reduces power consumption relative to Wi-Fi; however, its limited communication range makes it less viable for extended-range portable applications. In contrast, the prototype proposed in [36] employs Wi-Fi communication, enabling longer-range connectivity, but its system design is optimized for general IoT usage rather than energy efficiency, resulting in higher overall power consumption. The proposed system achieves a favorable balance between performance and power efficiency by combining a 1.2 V ADC and a Wi-Fi-enabled MCU, while maintaining real-time monitoring, data logging, and alarm functions. Unlike the higher-cost commercial systems, the proposed design offers comparable functionality with significantly reduced power and manufacturing cost. This makes it particularly well-suited for portable biomedical gas sensing applications.

5. Conclusions

This study proposed a cost-effective and energy-efficient portable gas sensor system based on a novel ROIC design and an optimized wireless communication algorithm. Experimental validation confirmed that the proposed system maintains high signal fidelity while significantly reducing power consumption and production cost. The integration of an adaptive signal extraction technique and a timing-optimized Wi-Fi communication scheme enables reliable real-time data transmission, even under resource-constrained conditions. The proposed architecture provides a scalable and practical solution for real-time gas sensing, offering a promising direction for the development of low-cost, high-performance environmental and healthcare monitoring technologies.