# Introduction rawn from industrial and commercial activities, the world's human population concentrates on specific regions. Such a phenomenon is known as 'urbanization'. Although the urbanization brings a higher economic development, the excessive population concentration will cause environmental damage and pollution like air pollution, noise pollution, water pollution, etc. Among various kinds of pollution, air pollution has a direct impact on our lives, because of the rapid emission of pollutants. Over the past decades, governments of many countries have imposed different regulations on air pollutants, so the severe damage brought to human health is reduced considerably. Although there may be no the immediate damage to human lives, however, air pollution still causes some chronic diseases. According to epidemiological studies [1], for example, the long-term exposure to pollutants may result in the harm to respiratory, nervous and cardiovascular systems. Therefore, the real-time monitoring of the air quality is particularly important and necessary. High urbanization in an area may lead to the absence of vegetation in that area. Air cleaning and temperature cooling become more difficult because of the lack of vegetation, and the area will be inevitably influenced by the urban heat island effect [2]. Moreover, the crowded tall buildings and heavy transportation also prevent air pollutants from dispersing. Thus, the government in Taiwan built many air quality monitoring stations to monitor the air quality in urban areas. However, the cost of building the stations is very expensive, so the deployment density is rather low. Let's take the Taipei city as an example. There are approximately 2.6 million people living in an area of 271.8 square kilometer, but only eight air quality monitoring stations are deployed in the area. The distance between the monitoring stations is more than dozens of kilometers. Such a strategy for air quality monitoring does not provide a higher spatiotemporal resolution. This paper proposes an automatic microscaled air quality monitoring system for areas with a high density of population and vehicles as shown in Figure 1. The system is based on the wireless sensor network technology and integrates with the global system for mobile communications (GSM) employed for data transmission. The monitoring system consists of a gateway and sensor nodes equipped with the air quality detection sensors. The gateway collects the data received from the sensor nodes and transmits the data to the control center with a database by the GSM network. The high-resolution meteorological data and concentrations of air pollutants can be provided in a real-time manner, including rainfall, wind speed, wind direction, temperature, humidity and the concentration of carbon monoxide (CO). For a preliminary research result, we only target the concentration of CO. In the future, however, a variety of pollutants will be considered by equipping the sensor nodes with corresponding gas sensors. CO is toxic to humans and animals. The National Fire Protection Association (NFPA) in the United States indicates that the main way that CO enters human body is through breathing. The symptoms of CO poisoning are very similar to the symptoms of getting the flu and food poisoning. The symptoms include shortness of breath, nausea, dizziness, and headaches. A high level of CO concentration is fatal, causing death within minutes if not treated [3]. Thus, in our research, a real-time monitoring system is proposed and implemented in the real environment. This system not only overcomes the problem that many monitoring systems cannot immediately send sensing data back but also provides a good way to monitor air pollution parameters such as CO concentration. This paper is Year 2014 ( D D D D D D D D ) organized as follows. Section II provides the related works. Section III outlines the architecture of the WSN and introduces our monitoring system. The section IV shows the experimental results. The final section concludes the paper. # II. Materials for Experiment and System Mapping In this study, the micro-scale air quality monitoring system is mainly composed of two parts, including a front-end automatic monitoring system and a control center, as shown in Figure 2. The front-end automatic monitoring system uses the WSN as its core technology, accompanied by the technology of Global System for Mobile Communications (GSM). This monitoring system could provide meteorological parameters and air pollution data from a small time scale. It includes a gateway and wireless sensor nodes. The gateway is used to control sensor nodes, collect the sensing data from the sensor nodes, and transmit the data to the control center through the short message service (SMS) provided by GSM. The gateway, on the other hand, is equipped with a weather station module, which provides various meteorological parameters, including temperature, relative humidity, barometric pressure, rainfall, wind speed, and wind direction. Through the Lab VIEW program [9], the sensing data can be stored and integrated into the database. In addition, users can make inquiries to the historical data and the latest updated data through the web page provided by the database. [12] and a temperature-humidity sensor (Sensirion SHT11) [13] for environmental information acquisition. The memory (10 KB RAM, 48 KB ROM flash, 1 MB extend flash memory) provided by the Octopus II is rather large, so it might serve as a buffer to temporarily store sensing data when the network connection fails. In addition, the Octopus II uses a low-power CC2420 wireless transceiver adopting IEEE 802.15.4 and the Zig Bee specification [14] for wireless communication. # a) Sensor Nodes # ii. Gaseous Pollution Sensors In this paper, CO derived from the vehicle emissions is selected as the research target, since it is one of the main sources of air pollution and highly toxic to human beings. We use MiCS-5525 [15] to monitor the CO concentration. MiCS-5525 is a semiconductorbased sensor produced by the e2v Company as a miniature carbon monoxide gas sensor (CO sensor). MiCS-5525 has the merits of short warm-up time, a smaller size and high sensitivity. Figure 4 shows the conversion of CO concentration and the resistance value. R0 is a benchmark resistance of MiCS-5525 under a relative humidity of 40 % and a temperature of 25 °C. Rscan be used to estimate the CO concentration under the same environmental conditions. In general, the across voltage of Rsis measured to estimate the concentration of CO in practical applications. iii. Bridge Module The bridge module is composed of voltage regulation circuits, signal modulation circuits and power switch circuits, as shown in Figure 7. The voltage regulation circuits are mainly used to regulate the voltage fed into the wireless communication module and the CO gas sensor. The signal modulation circuits are utilized to modulate the sensing signal generated from the CO sensor, and the power switch circuits are used to turn on/off the CO sensor in order to conserve energy. Figure 8 shows the bridge module. # III. Experimental Results In order to accurately obtain the CO concentration, we conducted a linear regression analysis on the monitoring data provided by the Environmental Protection Administration (EPA) monitoring station and our sensing data generated by the MCS-5521. Based on the linear regression result, we created a calibration equation that can be used to correct the readings of CO concentration from the CO sensor. In the calibration experiment, we set up 9 sensor nodes next to the sensors from the EPA monitoring station at Yong he in New Taipei City, as shown in Figure 9. The architecture of the sensor node in our system is shown in Figure 5. The sensor node includes a bridge module, a wireless communication module (Octopus II), a temperature/humidity sensor, and a CO sensor. We use a 12 V/ 9 Ah battery as the power supply for the node. The CO sensor is also driven by the bridge module. The sensor device is placed inside a waterproof box with two hole affixed with a breathable waterproof membrane as shown in Figure 6. Figure 17 shows the linear regression results based on the 9 sensor nodes, and the points represent the averages of 30 sets of sensing data collected every hour. The horizontal axis represents the readings (V) from the MiCS-5525, and the vertical-axis represents the readings from the Yong he monitoring sensor (ppm) that measures environmental CO concentration. Table II shows the results of the linear regression. The ID denotes the sensor nodes number, and the C (v) denotes the calibration equations. Hence, the real CO concentration can be acquired through the voltage readings provided by the MiCS-5525. In this study, the sensor nodes were used to collect air pollution and meteorological data every ten minutes, including temperature, relative humidity, CO sensor voltage, wind speed, wind direction, and atmospheric pressure. The CO concentration detected by the nine sensor nodes are shown in Figure 12. This study focuses on air pollution in the urban areas. The intersection circle of Keelung Road and Roosevelt Road (hereinafter referred to as the Gong Guan roundabout) in the Taipei city has witnessed heavy traffic during the rush hours, and this area urgently requires an air quality monitoring system. We set up a test bed in the Gong Guan roundabout. As shown in Figure 10, the urban air quality monitoring system located in the Gong Guan roundabout includes a gateway and 9 sensor nodes. Figure 11 shows the deployment of the proposed system. Figure 11: Deployment of the air quality monitoring system. A sensor node was attached to the traffic signal pole There were 24 data sets per day, coming from the hourly averaged data. Two peaks were found at 7 a.m. and 7 p.m., which echoed the rush hours during the daytime. Especially, in rush hours, there were significant changes in the data collected by sensor nodes 3, 4, 5, 7 and 9, because the sensor nodes were mounted on the traffic signals near the intersection. A high concentration of pollutions was detected and continuously accumulating when motor vehicles waited for green lights. As a result, the CO concentration detected by sensor nodes 3 and 9 was up to 9 parts per million (ppm). According to the definition of air pollution indicators [20], human body will be harmed when the CO concentration reaches 9 ppm. If so, people should avoid the area or wear masks. Thus, the proposed realtime urban air quality monitoring system can provide warnings to the public by showing a red light to further traffic planning. In Figure 12, we made a comparison of the CO concentration during rush hours in the morning and afternoon (7 a.m. to 9 a.m. and 5 p.m. to 7 p.m.). The CO concentration detected by sensor nodes 5, 8 and 10 at the rush hours in the morning were higher than that in the afternoon. As mentioned earlier, the Gong Guan roundabout is an important transportation hub in the Taipei city for people who live in the Zhonghe, Yong he and Xindian Dist. and have to commute to the city through the roundabout. The readings of the CO concentration in the morning indicated that the traffic flow from Zhong he and Yong he (Zhong-Yong he Dist.) to the Taipei city through the Keelung Road was heavier than that from Xindian Dist. to Taipei. In other words, the traffic flow from Zhong-Yong he Dist. to the Taipei city generates more serious air pollution than that from Xindian Dist. to the city during morning rush hours. On the other hand, Figure 12 shows that the sensor nodes 3, 4, 6, and 9 detected higher CO concentrations during the rush hours in the afternoon compared with the morning hours, because the heavy traffic flow during the rush hours was from the city to Xindian Dist. and Zhong-Yonghe Dist. Moreover, sensor nodes 3 and 9 were located at a corner of the intersection next to the traffic signal. Comparing with the sensor nodes 4 and 6, sensor nodes 3 and 9 detected that CO concentrations were higher than 9 ppm. Based on the findings, the CO concentration was largely influenced by the direction of the traffic flow and how long cars stopped in front of the traffic signal. IV. # Conclusion This study uses wireless sensor network technologies to acquire and record monitoring data for the goal of completely automatic air-quality monitoring. On the hardware side, we integrate sensor nodes with the CO sensors to perform air-quality-monitoring tasks. The sensor nodes are able to communicate with each other based on the Zig Bee protocol. The control center, controlled by the Lab VIEW program, successfully communicates with users through sending them SMS messages. It also stores a large amount of data into the database via the My SQL program, so that experts can establish a prediction model of pollution diffusion based on the data. In addition, the monitoring data reveals high-resolution pollution conditions near the Gong Guan roundabout in the Taipei city of Taiwan. The data can be an important source when addressing the issue of the impacts of motorcycles at idles (e.g. waiting for a green light) on air quality. Moreover, to achieve real-time monitoring, the data of CO concentration in a particular place could be reviewed from mobile communication devices, such as PDAs, smart phones, and tablet PCs to help keep air quality in check. 1![Figure 1: The architecture of the proposed automatic micro-scaled air quality monitoring system](image-2.png "Figure 1 :") 2![Figure 2 : The architecture of the micro-scale air quality monitoring system](image-3.png "Figure 2 :") 3![Figure 3 : Octopus II wireless communication module utilized in this study. (a) The front of the Octopus II; (b)The back of the Octopus IIi. Wireless Communication ModuleThe micro-scale air quality monitoring system employs the Octopus II, developed by the Department of Computer Science at National Tsing Hua University in Taiwan, as its wireless sensor nodes. The configuration of the Octopus II[10] is shown in Figure3. The Octopus II is based upon the Texas Instruments MSP430F1611 architecture[11]. Fifty pins are available for general purpose input/output and dual 12-bit digital-to-analog (D/A) converters with synchronization. The Octopus II contains a light sensor (Hamamatsu S1087)[12] and a temperature-humidity sensor (Sensirion SHT11)[13] for environmental information acquisition. The memory (10 KB RAM, 48 KB ROM flash, 1 MB extend flash memory) provided by the Octopus II is rather large, so it might serve as a buffer to temporarily store sensing data when the network connection fails. In addition, the Octopus II uses a low-power CC2420 wireless transceiver adopting IEEE 802.15.4 and the Zig Bee specification[14] for wireless communication.](image-4.png "Figure 3 :") 456![Figure 4 : Conversion of the CO gas concentration and the ratio of Rsto R0](image-5.png "Figure 4 :Figure 5 :Figure 6 :") 78![Figure 7 : The three sub-circuits of the bridge module](image-6.png "Figure 7 :Figure 8 ;") ![Figure 16 demonstrates that our sensor nodes and the sensors from the Yong he monitoring station were located at the same height. The experiment was Year 2014](image-7.png "") 9![Figure 9 : The location of the Yong he monitoring station](image-8.png "Figure 9 :") 10![Figure 10 : Locations of sensor nodes of the proposed system](image-9.png "Figure 10 :") 12![Figure 12 : CO concentrations on September 6, 2010, detected by each sensor node](image-10.png "Figure 12 :") © 2014 Global Journals Inc. (US) ## Acknowledgment This work was financially supported in part by the National Science Council, Taiwan, under contract no. NSC 100-2118-E-002-005 and NSC 100-2218-E-002-006. This work was also supported by the National Science Council, National Taiwan University and Intel Corporation under Grants NSC99-2911-I-002-201, NSC100-2911-I-002-001, and 10R70501.The authors would also like to thank the Council of Agriculture of the Executive Yuan, Taiwan, for its financial support under contracts no. 100AS-6.1.2-BQB1, and 100AS-6.1.2-BQ-B2. * Effects of air pollution on daily clinic visits for lower respiratory tract illness JSHwang CCChen American Journal of Epidemiology 155 1 2002 * A System Dynamic Model of Sustainable Urban Development: Assessing Air Purification Policies at Taipei City MCChen THo CGJan Asian Pacific Planning Review 4 1 Dec. 2006 * USA: National Fire Protection Association official website * Air Quality Monitoring with Sensor Map PVölgyesi ANádas XKoutsoukos ÁLédeczi Proc of the 7th International Conference on Information Processing in Sensor Networks of the 7th International Conference on Information essing in Sensor Networks 2008 * Micro Sensor Node for Air Pollutant Monitoring: Hardware and Software Issues SChoi NKim HCha RHa Sensors 9 10 2009 * Low-Cost Sensor Units for Measuring Urban Air Quality OA MPopoola MIMead GStewart JJBaldovi THodgson MMcleod RLJones PLandshoff MSimmons MHayes MCalleja Proc of the 24th Association of Academic Museums and Galleries of the 24th Association of Academic Museums and Galleries 2010 * ICT. 2009 /ProjectDetails.aspx? Project Id=88f4b9e7d120 4056bf8a660315bfc843 Accessed February 6, 2012 * GMV: Emergency drill under the Osiris project Belgium: EARSC. 2008 February 6,2012 Accessed * Lab VIEW Usa: Ni 2011. February 6, 2012 * Octopus II Datasheet Taiwan February 6, 2012 Available at * MSP430F1611 Datasheet 11. USA: TI 2011. February 6, 2012 Available at * SHT11 Datasheet :Switzerland Sensirion 2010. February 6, 2012 Available at. Accessed at * Acce-ssed at Usa: Ti CC2420-IEEE 802.15.4 2006. February 6, 2012 Datasheet * The weather station module WS-2316 Datasheet MiCS-5525 Datasheet 2009. February 6, 2012. 2011. February 6, 2012 16 USA: LA CROSSE TECHNOLOGY * ARK-3360L IPC Datasheet :Taiwan Advantech * Accessed at B5e889f55 February 6,2012 * A Wireless Agricultural Information Monitoring System Using GSM Technology JAJiang CLTseng CHChang YSChen WKPeng FMLu Journal of the Agricultural Association of China 7 1 2006 in Chinese * Acc-essed at: February 6 RO CEnvironmental Protection Administration Executive Yuan Taiwan 2012 * S1087 Datasheet :Japan Hamamatsu 2001. February 6 20 12 Available at. Accessed at