acm-header
Sign In

Communications of the ACM

Virtual extension

Monitoring Ambient Air Quality With Carbon Monoxide Sensor-Based Wireless Network


Carbon Monoxide (CO) is a poisonous air pollutant produced from the incomplete oxidation of carbon during the combustion process. It has a direct effect on the human body due to its affinity for blood hemoglobin, which inhibits the absorption of oxygen to the blood. The formation of carboxyhemoglobin complex can profoundly affect human health both on an acute and a chronic basis. CO can also be found inside any house at the level of 0.5-30 ppm [http://www.epa.gov/iaq/co.html] because it can be produced from the combustion of household utilities such as heater, stove, fireplace and automobile exhaust in the attached household garage. As CO is a colorless and an odorless gas, CO detectors need to be installed to monitor the CO concentration in a working environment.

For an ambient environment, the most popular way of measuring CO uses the principles of nondispersive infrared absorption (NDIR). Other useful methods are Gas Chromatography with flame ionization detector (GC/FID) or Catalytic oxidation techniques. U.S. Environmental Protection Agency (USEPA) employs NDIR as a traditional reference method for CO monitoring regulation. This method is performed by an analyzer and required standard gas system, pump, monitoring station, air conditioner or heater, computing equipment with appropriate programming, and other related equipment. All the necessary equipment needs to be housed and operated inside a room, and protected from rain, dust, and sunlight. Such preventive issues make this method complicated, cumbersome, and expensive.

Recent advances in wireless sensor networks (WSNs)2,5 make them an attractive solution for monitoring air quality. For instance, a wireless system designed to monitor indoor CO2 concentration is described in the literature.6 Lindsay Seders et al.6 deployed a sensor network to monitor water quality in St. Mary's Lake on the University of Notre Dame campus. This wireless sensor network used nodes by Mica2 and MDA300 from Crossbow Inc. [http://www.epa.gov/iaq/co.html]. Cardell-Oliver et al.3 developed and evaluated a reactive sensor network for monitoring soil moisture, which can adaptively change the sampling rate based on rainfall events. The successful deployment of these systems demonstrates that WSNs can be useful for some environmental monitoring scenarios.

Very little work has been done for CO monitoring with wireless sensor networks. Agrawal et al.1 have indicated that WSNs can provide continuous, real-time data of ambient air quality. The sensor systems, combined with the wireless communication network, give the benefit of convenience in deployment, and lower operation and maintenance cost when compared with NDIR technique. The sensor nodes can be powered by either batteries and/or solar energy sources. With the objective of monitoring the area around the University of Cincinnati (UC), 5 out of 15 planned CO sensors were placed on electric poles as shown in Figure 1. This was done to check the proof of the concept and the rest of sensors will be placed in the near future.

Back to Top

Functional Units of CO-based Wireless Sensor Network

Each sensor unit consists of five functional parts: Analog sensor unit, Data acquisition board, Wireless radio module, Solar charging unit and Radio Antenna.

Analog sensor unit. The Analog sensor unit contains the CO sensor and signal amplifying board. There are three common CO detection technologies used in CO sensors: chemo-optical, semiconductor, and electro-chemical [http://www.esru.strath.ac.uk]. Chemooptical technology mimics the response of human hemoglobin to CO by detecting the amount of passing infrared light. This rate of change of infrared light through the sensor is used to indicate the level of carbon monoxide concentration. The main drawback is that the sensor can non-reversibly accumulate CO and other contaminants over time and could cause false alarms. Furthermore, the need to continually replace batteries and sensors can also be expensive. Semiconductor sensor is the oldest carbon monoxide sensor. This technology utilizes a controlled quantity of tin dioxide as a sensing element. Change in sensor resistance is based on CO concentration. The disadvantage of this sensor type is the energy consumption higher than the others. Electrochemical sensors mostly use acid as an electrolyte and platinum as a catalyst to break down carbon monoxide gas and release electrons. The chemical reaction in the case of CO gas generation is:

ueq01.gif

From this reaction, electrons induce a small current between two electrodes proportional to the CO concentration. For our ambient air monitoring application, the CO concentration is low and usually not higher than 50 ppm. The CO sensor should be very accurate and must respond to low CO concentration. Therefore, among the three sensors, electro-chemical sensors are typically found to yield more accurate CO concentrations and are inexpensive in comparison with the others. Product specifications for the CO sensor used in this study can be found elsewhere (reference product specs from "Sixth-sense" company in UK [http://www.esru.strath.ac.uk]). Under ambient environment, the direct output signal of the sensor is about 48nA/ppm. A TI operational amplifier is used to enhance the low signal from the CO sensor and to make it large enough for the A/D converter. Resistors and capacitors are used to make the circuit stable and to maintain noise at lower level.

Data acquisition board and wireless radio module. Since the CO output signal is dependent on the temperature and humidity, we need to measure these three signals (CO, Temperature and Humidity) to calibrate the CO sensor reading. For seamless integration of sensors, the MDA300 data acquisition board manufactured by Crossbow has been used, which offers multiple analog and digital channels. The wireless radio MicaZ with Chipcon CC2420 transceiver transmits humidity, temperature and CO level in a pre-programmed order and is stored in 2.4GHz MDA300 Amtel board Atmega128L microcontroller [http://www.esru.strath.ac.uk]. The sensor's data can be transmitted to other sensor nodes en route to the gateway node, which is interfaced with a laptop as a base station for real time data storage at the gateway and availability at the Internet.

Radio antenna and solar charging unit. For this application, it has been necessary to transmit the sensor signal a long distance, and different antennas and wireless radios have been evaluated. The Mica2 433MHz, MicaZ, and MicaZ with external 8.5dbi Hyperlink Antennas have been tested. An external 8.5 dbi antenna is used to increase the transmission range of 30 feet for the original MicaZ antenna to more than 200 feet. Tests have been conducted on the roof top of a building, thus providing a flat terrain. The sensor network needed to send data continuously for at least a few months, and after deployment at a height of 15 feet on the utility poles it has been not very convenient to replace the batteries. Therefore, a heliomote solar charging unit has been integrated to provide continuous energy supply. The finished design of the CO node and custom built box on an electric pole are shown in Figure 2 and Figure 3 respectively.

Back to Top

Sensor Calibration

The assembled CO-WSN node needs to be calibrated in the laboratory so as to obtain an accurate and precise CO value. An environmental exposure chamber has been constructed for testing the sensor under different CO concentrations, temperatures and humilities. Real time temperature and humidity inside the chamber are measured by NIST traceable thermo-hygrometer, and CO concentration is monitored by a CO-NDIR analyzer which is calibrated at 10 ppm and 40 ppm certified standard CO. The environmental exposure chamber of size 18″ × 9 15/16″ × 9 5/16″ and 0.9 cubic feet is made of acrylic glass to ensure homogenous distribution of CO concentration with different temperatures and relative humidity. The purified air is obtained by using an air generator, flow to the molecular sieve, hopcalite (a mixture of copper and manganese oxides), ascarite (non fibrous silicate carrier coated with sodium hydroxide) and activated carbon columns to remove moisture, carbon monoxide, carbon dioxide and residue hydrocarbon, respectively. The desired CO concentration can be produced by diluting pure carbon monoxide with purified air to the designated value through a mass flow controller. Humidity and temperature are controlled by varying a by a flow-temperature-humidity control system (HCS-401, MNR) and keeping the mixing chamber in the incubator.

Based on the standard equipments data, the CO-WSN sensor has been calibrated and the CO calibration function developed under different temperature and humidity. The CO sensor provided a linear correlation with R2= 0.9951 (with temperature =20°C, humidity=30%) for a low CO concentration of 3-12 ppm (Figure 4). After the calibration, the CO-WSN has been equipped with a rain shield and an antenna. Our CO-WSN sensor has been deployed and tested the system in actual environmental conditions, with the effect of temperature, humidity, wind, sun, and other factors. The solar panel system has been assembled to provide energy to the batteries. Data from the CO-WSN sensors have been transmitted back to a laptop as the gateway node in multi-hop fashion. For energy efficiency, the CO-WSN sensor has been set to wake up every 3 minutes, transmit data and then go back to sleep. Sensor readings have been continuously collected from June 21 to June 29, 2007 and the relationship between sensor data and CO concentration in ppm is shown in the Figure 4.

In this article, design of the CO-WSN sensor for ambient CO monitoring has been explored. The field test shows a successful deployment of CO-WSN. The CO-WSN appears to provide a more convenient way to monitor the ambient carbon monoxide. In the future, ozone sensors may be incorporated onto the CO-WSN platform to make the system a more powerful tool for monitoring other ambient air pollutants.

This work has been partially supported by the National Science Foundation under research grant BES-0529063 and MRI equipment grant 0521189 and the Ohio Board of Regents Doctoral Enhancement Funds.

Back to Top

References

1. Agrawal, D.P., Lu, M., Keener, T.C., Dong, M., and Kumar. V. Environmental Monitoring Using Wireless Sensors. EM (2004) 35–41.

2. Agrawal, D.P. and Zeng, Q.A. Introduction to Wireless and Mobile Systems. Brooks/Cole, 438 pages, Aug. 2002, ISBN No. 0534-40851-6, second edition (2005) ISBN No. 0-534-49303-3.

3. Cardell-Oliver, R., Kranz, M., Smettem, K., and Mayer, K. A reactive soil moisture sensor network: design and field evaluation. International J. Distributed Sensor Networks 1, (2005) 149–162.

4. Chung, W.Y. and Oh, S.J. Remote monitoring system with wireless sensors module for room environment. Sensors and Actuators B: Chemical 133, 1, (2006) 64–70.

5. Cordeiro, C. and Agrawal, D.P. Ad hoc & Sensor Networks: Theory and Applications. World Scientific Publishing, (2006) ISBN No. 81-256-681-3.

6. Seders, V., Shea, V, Lemmon, M.D., Maurice, P.A. and Talley, J.W. LakeNet: an integrated sensor network for environmental sensing in lakes. Environmental Engineering Science 24, 2, (2007), 183–191.

Back to Top

Authors

Demin Wang (wangdm@cs.uc.edu) has completed his Ph.D. from the Department of Computer Science, University of Cincinnati, Cincinnati, OH, and now works for Microsoft Corp.

Dharma P. Agrawal (dpa@cs.uc.edu) is the Ohio Board of Regents Distinguished Professor in the Department of Computer Science, University of Cincinnati, Cincinnati, OH.

Wassana Toruksa is an Environmental Officer, Ambient Air Quality Division, Air Quality and Noise Management Bureau, Pollution Control Department (PCD), Bangkok, Thailand.

Chaichana Chaiwatpongsakorn (chaiwac@email.uc.edu) is a doctoral student in the Department of Civil and Environmental Engineering, University of Cincinnati, Cincinnati, OH.

Mingming Lu (lumg@ucmail.uc.edu) is an associate professor in the Department of Civil and Environmental Engineering, University of Cincinnati, Cincinnati, OH.

Tim C. Keener (keenertc@ucmail.uc.edu) is a professor in the Department of Civil and Environmental Engineering, and an Associate Dean for Graduate Studies and Research, University of Cincinnati, Cincinnati, OH.

Back to Top

Footnotes

DOI: http://doi.acm.org/10.1145/1735223.1735257

Back to Top

Figures

F1Figure 1. 5 of 15 CO Sensors Placement in the West Campus of the University of Cincinnati

F2Figure 2. Components of a CO-WSN Unit

F3Figure 3. CO-WSN Unit Mounted on a Street Light Pole

F4Figure 4. Relationship between CO-WSN signal and CO concentration

Back to top


©2010 ACM  0001-0782/10/0500  $10.00

Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.

The Digital Library is published by the Association for Computing Machinery. Copyright © 2010 ACM, Inc.


 

No entries found