A Wireless And Batteryless Intelligent Carbon Monoxide Sensor

Transcript

sensors Article A Wireless and Batteryless Intelligent Carbon Monoxide Sensor Chen-Chia Chen *, Gang-Neng Sung, Wen-Ching Chen, Chih-Ting Kuo, Jin-Ju Chue, Chieh-Ming Wu and Chun-Ming Huang National Chip Implementation Center, National Applied Research Laboratories, 7F, No. 26, Prosperity Rd. I, Science Park, Hsinchu City 30078, Taiwan; [email protected] (G.-N.S.); [email protected] (W.-C.C.); [email protected] (C.-T.K.); [email protected] (J.-J.C.); [email protected] (C.-M. W.); [email protected] (C.-M.H.) * Correspondence: [email protected]; Tel.: +886-3-577-3693 Academic Editor: Teen-Hang Meen Received: 29 May 2016; Accepted: 19 September 2016; Published: 23 September 2016 Abstract: Carbon monoxide (CO) poisoning from natural gas water heaters is a common household accident in Taiwan. We propose a wireless and batteryless intelligent CO sensor for improving the safety of operating natural gas water heaters. A micro-hydropower generator supplies power to a CO sensor without battery (COSWOB) (2.5 W at a flow rate of 4.2 L/min), and the power consumption of the COSWOB is only ~13 mW. The COSWOB monitors the CO concentration in ambient conditions around natural gas water heaters and transmits it to an intelligent gateway. When the CO level reaches a dangerous level, the COSWOB alarm sounds loudly. Meanwhile, the intelligent gateway also sends a trigger to activate Wi-Fi alarms and sends notifications to the mobile device through the Internet. Our strategy can warn people indoors and outdoors, thereby reducing CO poisoning accidents. We also believe that our technique not only can be used for home security but also can be used in industrial applications (for example, to monitor leak occurrence in a pipeline). Keywords: Internet of Things; wireless sensor network; batteryless; carbon monoxide; sensor; embedded system 1. Introduction As billions of smart sensors connect to the Internet, one popular Internet of Things (IoT) application features an intelligent home gateway that becomes a central point of connection for smart sensors that monitor the home environment [1–8]. Carbon monoxide (CO) poisoning is a significant cause of illness and death in the United States (US). The internal combustion engine and stoves burning fossil fuels result in most CO poisoning [9]. In contrast to the US, most CO poisoning in Taiwan results from the misuse of gas-powered water heaters [10]. A CO alarm is commonly used for detecting the presence of CO gas in ambient conditions to prevent CO poisoning. A typical CO alarm sounds a warning before the CO level reaches a dangerous level, but it only alerts people nearby. A CO alarm with Wi-Fi function that connects to the Internet to warn people is very rare [11–14]. Nevertheless, the Wi-Fi communication protocol consumes much energy, which results in lower battery life and battery lifespan as compared with that of a typical stand-alone CO alarm. When a low battery level is achieved, the CO alarm might disconnect from the Internet. In order to overcome a battery reaching a low battery level, a rechargeable battery is used to power the wireless CO alarm. To the best of our knowledge, most indoor CO alarms are powered by a long-lifetime (over a few years) sealed battery. In fact, the CO alarm could possibly fail to function due to battery failure. Taiwan sees especially brutal heat and high humidity in the summer. CO alarms installed in the balcony are easily exposed to high temperature and high humidity conditions, reducing the life of the batteries. Sensors 2016, 16, 1568; doi:10.3390/s16101568 www.mdpi.com/journal/sensors Sensors 2016, 16, 1568 2 of 11 Sensors 2016, 1568 exposed to high temperature and high humidity conditions, reducing the life 2 ofof 11 balcony are16,easily the batteries. In been used to to power a In the the past pastdecade, decade,some someenergy energyharvesting harvestingtechnologies technologieshave havealready already been used power wireless sensor node. If we can harvest enough power for an applied wireless sensor, we do not need a wireless sensor node. If we can harvest enough power for an applied wireless sensor, we do not need to toworry worryabout aboutsensor sensorfailure failurefrom fromrunning runningout outof ofbattery batterypower. power.For Forexample, example,some someradio radiofrequency frequency identification (RFID) sensors were driven by using energy harvesting from the RF field [15,16]. A identification (RFID) sensors were driven by using energy harvesting from the RF field [15,16]. passive RFID sensor does not contain a battery, but most of the low-frequency and high-frequency A passive RFID sensor does not contain a battery, but most of the low-frequency and high-frequency passive m. Ultra-high-frequency passive RFIDs RFIDs provide provide short short read read ranges ranges from from 10 10 cm cm to to 11 m. Ultra-high-frequency (UHF) (UHF) passive passive RFIDs with a long read range can be as long as 12 m; nevertheless, a UHF RFID cannot used RFIDs with a long read range can be as long as 12 m; nevertheless, a UHF RFID cannot be usedbe without without in the most of theThe world. The best-known energy-harvesting are solar cells a licenseainlicense most of world. best-known energy-harvesting collectorscollectors are solar cells and wind and wind turbines, which have become major renewable energy sources. Solar cells and wind turbines, which have become major renewable energy sources. Solar cells and wind turbines are also turbines are also used toremote powersensors wireless remote sensors [17,18]. Unfortunately, solar cells used to power wireless [17,18]. Unfortunately, utilizing solar cellsutilizing and wind turbines and wind turbines in an indoor facesas challenges such as low and indoor lighting and source. a lack in an indoor environment facesenvironment challenges such low indoor lighting a lack of wind of wind source. However, a novel self-powered water consumption sensor has been demonstrated However, a novel self-powered water consumption sensor has been demonstrated [19], and its input [19], andisits input energy from is only harvested from through the flow the of water energy only harvested the flow of water microthrough turbine. the micro turbine. Ultimately, Ultimately, aa batteryless batteryless wireless wirelesssensor sensorwill willpossibly possiblybe beaafinal final candidate candidatein inthe thefuture; future;itit does does not adversely impact the environment. To enable batteryless wireless sensors to operate reliably, a not adversely impact the environment. To enable batteryless wireless sensors to operate reliably, power management circuit that monitors energy harvesters and optimizes the use of harvest energy a power management circuit that monitors energy harvesters and optimizes the use of harvest energy isisaakey keytechnology. technology.Traditionally, Traditionally, the the hydroelectric hydroelectric generator generator is is able able to to supply supply power powerto tothe theelectronic electronic system when the flow in the pipe is stable. Otherwise, the energy runs out when the flow system when the flow in the pipe is stable. Otherwise, the energy runs out when the flowcannot cannotpush push the turbine forward to the hydroelectric generator or when the output energy of the hydroelectric the turbine forward to the hydroelectric generator or when the output energy of the hydroelectric generator to supply supply the theelectronic electronicsystem. system.However, However,ininmany manyreality realityconditions, conditions, such generator is is not not sufficient sufficient to such as as in slow or variation flow, the output voltage of the hydroelectric generator is very weak or nonin slow or variation flow, the output voltage of the hydroelectric generator is very weak or non-stable. stable. Weak energy the system, and non-stable might damage the system Weak energy cannotcannot enable enable the system, and non-stable voltagevoltage might damage the system too. too. In this study, we use an intelligent hydroelectric energy harvesting technology to power In this study, we use an intelligent hydroelectric energy harvesting technology to poweraa CO CO sensor without a battery (COSWOB), whether the flow in the pipe is stable or not. Figure 1 illustrates sensor without a battery (COSWOB), whether the flow in the pipe is stable or not. Figure 1 illustrates an anoutput outputpower powercomparison comparisonof ofthe thetraditional traditionaland andproposed proposedmethods, methods,with withfast fastand andslow slowflow flowrates. rates. Figure Figure1.1.Output Outputpower powercomparison comparisonwith withfast fastand andslow slowflow flowrates. rates. Our Ourstrategy strategyto toprevent preventCO CO poisoning poisoningby bymonitoring monitoringthe theCO COlevel levelby byaa COSWOB COSWOBisis shown shownin in Figure 2. The COSWOB is installed in a cold water inlet of a natural gas water heater, which is Figure 2. The COSWOB is installed in a cold water inlet of a natural gas water heater, which islocated located on and the theWi-Fi Wi-Fialarms alarmsare areplaced placedinin another location, such a bathroom, living room, on aa balcony, balcony, and another location, such as aasbathroom, living room, and and bedroom. The COSWOB monitors the CO level in ambient conditions, which is generated from bedroom. The COSWOB monitors the CO level in ambient conditions, which is generated from the gas the gasheater, water and heater, and transmits these accumulated data to an intelligent gatewaythe through the water transmits these accumulated data to an intelligent gateway through Bluetooth Bluetooth (BLE)When protocol. When the CO level reaches alevel, dangerous level, the COSWOB low-powerlow-power (BLE) protocol. the CO level reaches a dangerous the COSWOB alarm sounds alarm sounds loudly. Additionally, an application which is run in the intelligent gateway will send to a loudly. Additionally, an application which is run in the intelligent gateway will send a notification Sensors 2016, 16, 1568 3 of 11 Sensors 2016, 16, 1568 3 of 11 activate Wi-Fi alarms when the COSWOB senses CO over the threshold value in the air. The Wi-Fi Sensors 2016, 16, 1568 of 11 notification to activate Wi-Fi alarms when the COSWOB senses CO over the threshold value in 3the alarms can generate louder sounds to alert users in the bathroom and other people in the living room air. The Wi-Fi alarms can generate louder sounds to alert users in the bathroom and other people in or bedroom. Meanwhile, the intelligent gateway also sends a CO warning message smartphones of notification to activate Wi-Fi alarms when the intelligent COSWOB senses over the value in to the the living room or bedroom. Meanwhile, the gateway also sends athreshold warningtomessage air. The Wi-Fi alarms can generate louder sounds to alert users in the bathroom and other people in their family members outdoors in case users miss this louder alarm. smartphones of their family members outdoors in case users miss this louder alarm. the living room or bedroom. Meanwhile, the intelligent gateway also sends a warning message to smartphones of their family members outdoors in case users miss this louder alarm. 2. Schematic of the key devices in our strategy for avoiding carbon monoxide (CO) poisoning Figure Figure 2. Schematic of the key devices in our strategy for avoiding carbon monoxide (CO) happen. poisoning happen. Figure 2. Schematic of the key devices in our strategy for avoiding carbon monoxide (CO) poisoning The typical CO alarm works 24 h per day to detect CO release from the natural gas water heater. happen. mostly in certain periods a day, and the TheHowever, typical hot COwater alarmconsumption works 24 hisper dayconcentrated to detect CO release from in the natural gasnatural water heater. gas water heater only possibly releases CO during those periods. It is not necessary to detect CO from However, The hot typical water CO consumption is mostly concentrated in certain periods in a day, and the natural alarm works 24 h per day to detect CO release from the natural gas water heater. the natural gas water heater when it is shut down; it will waste battery energy of the typical CO gas water heater only possibly releases CO during those periods. It is not necessary to detect CO However, hot water consumption is mostly concentrated in certain periods in a day, and the natural alarm. In contrast to the typical CO alarm, when the natural gas water heater is shut down, the gas water heater only possibly releases CO during those periods. It is not necessary to detect CO fromtypical from the natural water when shutflowing down; it will waste battery the COSWOB isgas always OFFheater due to the lackitofiswater through the generator. It isenergy only ONofwhen the natural gas water heater when it is shut down; it will waste battery energy of the typical COdown, CO alarm. In contrast to the typical CO alarm, when the natural gas water heater is shut users need hot water to operate the natural gas water heater. When the cold water flows through the alarm. In is contrast toOFF the typical CO when the flowing naturala gas water heater isthe shut down, the ON generator of the COSWOB, the to COSWOB will turn ON within few seconds. COSWOB is the COSWOB always due thealarm, lack of water through theAfter generator. It is only COSWOB is always OFF due to the lack of water flowing through the generator. It is only ON when ON, need it connects to an intelligent gateway until the is OFF.When Moreover, typical wireless COthrough when users hot water to operate the natural gasCOSWOB water heater. the cold water flows users need hotconnect water totooperate the natural gas water is heater. When the cold battery water flows through the alarms only the Internet when an alarm activated to extend life and battery the generator ofofthe COSWOB, theCOSWOB COSWOB will turn ON within aseconds. few seconds. After the COSWOB generator the COSWOB, the will turn ON within a few After the COSWOB lifespan. Therefore, the COSWOB can provide early warnings to people before the CO level reachesis is ON,ON, it connects totoan gateway until COSWOB isMoreover, OFF. Moreover, typicalCO wireless it connects an intelligent intelligent gateway until the the COSWOB is OFF. typical wireless a dangerous level. CO alarms Internetwhen whenananalarm alarm is activated to extend battery lifebattery and battery alarmsonly onlyconnect connect to to the the Internet is activated to extend battery life and lifespan. Therefore, COSWOB canprovide provide early early warnings people before the CO reaches 2. Therefore, System Hardware lifespan. the the COSWOB can warningstoto people before thelevel CO level reaches a dangerous level. a dangerousThe level. COSWOB consists of three modules—a power module, a sensor module, and a microcontroller module, as shown in Figure 3. Each module has a dimension of 25 mm × 25 mm 2. System Hardware 2. System Hardware (width × length), and they are connected through socket connectors with a common power and communication bus. consists of three modules—a power module, a sensor module, and a The COSWOB The COSWOB consists of three modules—a power module, a sensor module, and a microcontroller microcontroller module, as shown in Figure 3. Each module has a dimension of 25 mm × 25 mm module, as shown in Figure 3. Each module has a dimension of 25 mm × 25 mm (width × length), (width × length), and they are connected through socket connectors with a common power and and they are connected communication bus.through socket connectors with a common power and communication bus. Figure 3. Cont. Sensors 2016, 16, 1568 4 of 11 Sensors 2016, 16, 1568 4 of 11 Sensors 2016, 16, 1568 4 of 11 Figure 3. Photographs theprinted printedcircuit circuit board thethe power module; (b) the sensor module;module; (c) Figure 3. Photographs of of the boardofof(a)(a) power module; (b) the sensor the microcontroller module; and (d) the complete assembled stack. (c) the microcontroller module; and (d) the complete assembled stack. 2.1. Power Module 2.1. Power Module The power module (Figure 3a) is used to harvest energy from the water flow in a pipe of the gas The power (Figure 3a) is used to from theCOSWOB. water flow a pipe of the gas water heatermodule and supply multiple outputs (2.5harvest V, 3.3 V,energy and 5 V) to the Thein power module Figure 3. Photographs of the printed circuit board of (a) the power module; (b) the sensor module; (c) waterisheater and supply multiple outputs (2.5 V, 3.3 V, and 5 V) to the COSWOB. The power module composed of three core units: (1) Energy Harvesting with Boost Charger; (2) Voltage Reference the microcontroller module; and (d) the complete assembled stack. Generator; and (3) Step-Up Converter. To drive a sensor readout circuit, 5 V of power supply and 2.5 is composed of three core units: (1) Energy Harvesting with Boost Charger; (2) Voltage Reference V reference areConverter. required, and functional diagram of circuit, the power is shown in and 2.1.ofPower Module Generator; and (3) voltage Step-Up To the drive a sensor readout 5 Vmodule of power supply 4. In thevoltage harvesting state, the Energy Harvesting Circuit collects the non-stable energy is from 2.5 V Figure of reference are required, and the functional diagram of the power module shown The power module (Figure 3a) is used to harvest energy from the water flow in a pipe of the gas the hydroelectric generator and stores energy in the Energy Storagecollects Element, which is the 0.1-F water4. heater and harvesting supply multiple outputs (2.5 V, 3.3 Harvesting V, and 5 V) to Circuit the COSWOB. The the power module energy in Figure In the state, thethe Energy non-stable supercapacitor in this work. The Battery Threshold Control Circuit monitors the voltage of the is composed of threegenerator core units: and (1) Energy with Boost Charger; (2) Voltage Reference from the hydroelectric storesHarvesting the energy in the Energy Storage Element, which is the supercapacitor. When the Converter. voltage of To thedrive supercapacitor is greater than 4.5 V, the supply enable and signal is set, Generator; and (3) Step-Up a sensor readout circuit, 5 V of power 2.5 0.1-F supercapacitor in this work. The Battery Threshold Control Circuit monitors the voltage of and DC/DC Converters start to work (Working diagram state). On other hand, voltageinof the V of the reference voltage are required, and the functional of the the power modulethe is shown the supercapacitor. When the voltage ofinput the energy supercapacitor is greater than 4.5 V, the enable signal supercapacitor to drop when is not enough to supply the sensing system. Figure 4. In the starts harvesting state, the the Energy Harvesting Circuit collects the non-stable energy from Until is set,the and the DC/DC Converters start to work (Working state). On the other hand, the voltage dropsgenerator to underand 2.3 stores V, thethe enable signal is Energy unset and the system back to harvesting hydroelectric energy in the Storage Element,goes which is the 0.1-Fvoltage of the supercapacitor starts drop when the input energy is not enough to supply the sensing state. A DC/DCin Step-Up Converter and aThreshold low-dropout Voltage Reference Generator are proposed in supercapacitor this to work. The Battery Control Circuit monitors the voltage of the system. supercapacitor. When voltage of isisgreater than theregulation enable signal isto set,harvesting design. A low-side switch regulator ideal for boost and SEPIC DC/DC is intended to Until this the voltage drops tothe under 2.3 V,the thesupercapacitor enable signal unset and4.5 theV,system goes back and the the DC/DC Converters startConverter to and worka(Working On all thethe other hand, the voltage ofare theproposed proposed Step-Up circuit. It state). provides active functions to provide local state. achieve A DC/DC Step-Up Converter low-dropout Voltage Reference Generator supercapacitor starts to dropfast when the input energy isand enough supply the sensing system. Until with transient response accurate regulation. Switching frequency is in thisDC/DC design.conversion A low-side switch regulator ideal fornot boost andtoSEPIC DC/DC regulation is intended the voltage drops to under 2.3 V, the enable signal is unset and the system goes back to harvesting internally set to 1.6 MHz, allowing the use of a tiny surface mount inductor and chip capacitors, while to achieve the proposed Step-Up Converter circuit.Voltage It provides all Generator the active functions in to provide state. A DC/DC Step-Up Converter andDC/DC a low-dropout Reference are proposed providing efficiencies near 90%. regulator used current-mode control and internal local DC/DC conversion with fast transient response and accurate regulation. Switching this design. A low-side regulator ideal for boost and SEPIC DC/DC regulation is intended tofrequency compensation methodsswitch provides a minimum component count and high-performance regulation is internally set to 1.6 MHz, allowing thecircuit. useThe ofIt energy aprovides tiny surface mount inductor and chip capacitors, achieve the Converter all the active functions provide over a wideproposed range ofStep-Up operating conditions. harvesting with a boostto charger islocal designed DC/DC conversion with fast transient response and accurateused regulation. Switching control frequencyand is internal whilewith providing efficiencies near 90%. DC/DC regulator current-mode the flexibility to support a variety of energy storage elements, such as a rechargeable battery, a internally set to 1.6 MHz, allowingathe use of a tinycomponent surface mount inductor chip capacitors, while compensation methods minimum andand high-performance regulation supercapacitor, or a provides conventional capacitor. Therefore, acount Battery Threshold Control circuit is providing efficiencies near 90%. DC/DC regulator used current-mode control and internal over adesigned wide range operating conditions. energy a boost charger designed in theof sub-circuit. The availabilityThe of the sourcesharvesting from whichwith harvesters extract theirisenergy compensation methods provides a minimum component count and high-performance regulation can often be sporadic or time-varying. The energy harvesting with a boost charger typically need with the to of support a variety of The energy storage elements, such as a will rechargeable overflexibility a wide range operating conditions. energy harvesting with a boost charger is designed battery, some energy storage element, such as a supercapacitor in the proposed system. The supercapacitor a supercapacitor, or a conventional capacitor. Therefore, a Batterysuch Threshold Control circuit with the flexibility to support a variety of energy storage elements, as a rechargeable battery,is a designed constant to the COSWOB andTherefore, also allows to handle anyextract peak currents that is cannot supercapacitor, or power a conventional capacitor. a it Battery Threshold Control in theprovides sub-circuit. The availability of the sources from which harvesters theircircuit energy can often directly come the input source. designedor intime-varying. thefrom sub-circuit. The availability of the sourceswith fromawhich harvesters their energy be sporadic The energy harvesting boost chargerextract will typically need some can often be sporadic or time-varying. The energy harvesting with a boost charger will typically need energy storage element, such as a supercapacitor in the proposed system. The supercapacitor provides some energy storage element, such as a supercapacitor in the proposed system. The supercapacitor constant power to the COSWOB and also allows it to handle any peak currents that cannot directly provides constant power to the COSWOB and also allows it to handle any peak currents that cannot come directly from the input come fromsource. the input source. Figure 4. The functional block diagram of power module. Figure 4. The functional block diagram of power module. Figure 4. The functional block diagram of power module. Sensors 2016, 16, 1568 5 of 11 2.2. Sensor Module The sensor module (Figure 3b) supports both types of electrochemical sensors (three leads Sensors 2016, 16, 1568 5 of 11 and two leads gas sensors), but three-lead electrochemical CO sensor (CO-AF, Alphasense, Essex, UK) is 2.2. Sensor Module only used in this study. A potentiostat is required for maintaining appropriate bias potential between the reference themodule counter(Figure electrodes of the electrochemical gas sensor. We adapt a programmable Theand sensor 3b) supports both types of electrochemical sensors (three leads and analog front end LMP91000 (Texas Instruments,CO Dallas, USA),Alphasense, to provideEssex, a biasUK) between two leads gas(AFE), sensors), but three-lead electrochemical sensorTX, (CO-AF, is only used in this study. A potentiostat is required for maintaining appropriate bias potential between reference and counter electrodes and convert into a voltage from the output current of work electrodes. the reference andconverter the counter(ADC161S626, electrodes of the electrochemical gasDallas, sensor.TX, We USA) adapt asamples programmable An analog-to-digital Texa Instruments, the output analog front end (AFE), LMP91000 Instruments, TX, USA), to for provide a bias between voltage of the LMP91000 and converts(Texas this analog voltageDallas, to a digital value further data gathering reference and counter electrodes and convert into a voltage from the output current of work by a microcontroller. electrodes. An analog-to-digital converter (ADC161S626, Texa Instruments, Dallas, TX, USA) samples the output voltage of the LMP91000 and converts this analog voltage to a digital value for further 2.3. Microcontroller Module data gathering by a microcontroller. The microcontroller (nRF51822, Nordic Semiconductor, Skøyen, Norway) used here is from Nordic, Microcontroller which2.3. is built around aModule 32-bit ARM CortexTM M0 CPU (Central Processing Unit) with an embedded 2.4-GHz transceiver which supports BLE Semiconductor, protocol. TheSkøyen, nRF51822 configures the is LMP91000 The microcontroller (nRF51822,the Nordic Norway) used here from 2 TM through Inter-Integrated interface and reads output voltage of theUnit) LMP91000 Nordic, which is builtCircuit around(Ia C) 32-bit ARM Cortex M0the CPU (Central Processing with anfrom embedded 2.4-GHz transceiver which supports the BLE protocol. TheThe nRF51822 configures the the ADC161S626 through a Serial Peripheral Interface (SPI) interface. COSWOB will transmit LMP91000 Inter-Integrated (I2C) and reads thethe output voltageisof the collected data tothrough the intelligent gatewayCircuit through theinterface BLE protocol. When COSWOB activated, LMP91000 from initializes the ADC161S626 a Serial Peripheral The COSWOB a series of sequences itself. through These sequences initializeInterface not only(SPI) the interface. frequently used peripheral will transmit collected data to the intelligent gateway through the BLE protocol. When the COSWOB devices such as GPIOs, timers, I2C, and SPI, but also the configurations of the LMP91000 and the BLE is activated, a series of sequences initializes itself. These sequences initialize not only the frequently services. After initialization, the nRF51822 periodically reads the output voltage of LMP91000 by the used peripheral devices such as GPIOs, timers, I2C, and SPI, but also the configurations of the ADC data value from the ADC161S626 through the I2C bus. A timer is set to notice the nRF51822 by LMP91000 and the BLE services. After initialization, the nRF51822 periodically reads the output a specific interrupt every by 20 the ms.ADC When anvalue interrupt the accumulation of data voltage of LMP91000 data from happens, the ADC161S626 through the I2C bus. is A operated timer is to calculate thenotice meanthe (64nRF51822 sensor data), the averaged result20 is ms. sentWhen to thean intelligent set to by a and specific interrupt every interrupt gateway happens,through the the BLE protocol. of data is operated to calculate the mean (64 sensor data), and the averaged result is accumulation to the intelligent gateway through the BLEisprotocol. Asent micro-hydropower generator (F50-5V) coupled to the turbine and used to convert the A micro-hydropower generator (F50-5V) is coupled to theinturbine usedmicro-hydropower to convert the mechanical energy from the water into electrical energy, as shown Figureand 5a. The mechanical energy from the water into electrical energy, as shown in Figure 5a. The natural micro- gas generator produces the electric power from the movement of tap water in the pipe of the hydropower generator produces the electric power from the movement of tap water in the pipe of water heater, and the output power directly supplies the power module. The maximum output voltage the natural gas water heater, and the output power directly supplies the power module. The and current are 5 V and >100 mA, respectively. The water pressure of the water outlet closed and open maximum output voltage and current are 5 V and >100 mA, respectively. The water pressure of the at the water maximum voltage are 0.6 Mpa and 1.2 Mpa, respectively. The micro-hydropower generator has outlet closed and open at the maximum voltage are 0.6 Mpa and 1.2 Mpa, respectively. The a dimension of 88 mm ×generator 58 mm × 39amm (lengthof×88width input× and output connect micro-hydropower has dimension mm × × 58height), mm × 39and mmits (length width × height), 1 threadand gauge are G ”. We could easily install the COSWOB on a natural gas water heater by its input and a 2 output connect thread gauge are G ½’’. We could easily install the COSWOB on using a standard flexible pipe. Theby complete assembled stackpipe. wasThe fixed by screws on thestack top surface of the natural gas water heater using a standard flexible complete assembled was fixed by screws on the topthe surface of the micro-hydropower, andthen the electrochemical CO on sensor then micro-hydropower, and electrochemical CO sensor was directly mounted the was socket of the directly mounted on the socket the PCB (Printed Circuit Board) of the sensor module. PCB (Printed Circuit Board) of the of sensor module. A micro-hydropower generator;and and(b) (b) cross-sectional cross-sectional view (c)(c) toptop view of the FigureFigure 5. (a)5.A(a) micro-hydropower generator; viewand and view of CO the CO without battery (COSWOB). sensorsensor without battery (COSWOB). Sensors 2016, 16, 1568 Sensors 2016, 16, 1568 6 of 11 6 of 11 3. System Software 3. System Softwareis run in the intelligent gateway (SZ87R6, Shuttle) for collecting and storing CO An application sensorAn data. A NordicisnRF51 dongle is plugged into the USB port the intelligent gateway. application run inBluetooth the intelligent gateway (SZ87R6, Shuttle) forof collecting and storing CO Once the application is launched, it starts to discover numbers of the Bluetooth dongles on it and then sensor data. A Nordic nRF51 Bluetooth dongle is plugged into the USB port of the intelligent gateway. generates corresponding control daemons for detecting Bluetooth dongles. The control daemons Once the application is launched, it starts to discover numbers of the Bluetooth dongles on it and initialize the corresponding Bluetooth dongles; at the same time, the application generates a data pool then generates corresponding control daemons for detecting Bluetooth dongles. The control daemons for saving data, which is received from the activated COSWOB. The status of the COSWOB is initialize the corresponding Bluetooth dongles; at the same time, the application generates a data updated instantly on the main page of the server application. While receiving data from the activated pool for saving data, which is received from the activated COSWOB. The status of the COSWOB is COSWOB, the application also transmits the CO level detected by the COSWOB to a smartphone updated instantly on the main page of the server application. While receiving data from the activated through the Wi-Fi protocol. Moreover, the application also sends a notification to trigger the Wi-Fi COSWOB, the application also transmits the CO level detected by the COSWOB to a smartphone alarm if necessary. through the Wi-Fi protocol. Moreover, the application also sends a notification to trigger the Wi-Fi We also developed a mobile application for monitoring the COSWOB. The mobile application is alarm if necessary. available on the Android operating system. Users or their family members can monitor the status of We also developed a mobile application for monitoring the COSWOB. The mobile application the COSWOB in real-time through mobile devices, such as a smartphone and a tablet. The mobile is available on the Android operating system. Users or their family members can monitor the application listens for the data transmitted from the intelligent gateway. For visual displays, we status of the COSWOB in real-time through mobile devices, such as a smartphone and a tablet. typically use a numerical indicator and a chart. Moreover, when the detected CO level reaches a The mobile application listens for the data transmitted from the intelligent gateway. For visual dangerous level, the mobile application will alert the user with siren sounds, and a warning sign and displays, we typically use a numerical indicator and a chart. Moreover, when the detected CO level the occurrence time will be displayed on the screen to notify users. The warning sign lasts until users reaches a dangerous level, the mobile application will alert the user with siren sounds, and a warning dismiss it. sign and the occurrence time will be displayed on the screen to notify users. The warning sign lasts until users dismiss it. 4. Calibration Method of the COSWOB 4. Calibration Method of the COSWOB The COSWOB is calibrated and tested using a homemade gas sensor calibration system where known concentrations of CO gas aretested introduced the cylinders, as shown in Figure This The COSWOB is calibrated and using afrom homemade gas sensor calibration system6.where method works by introducing gasare of known concentrations (25, 50, 100 ppm) then performing check known concentrations of CO gas introduced from the cylinders, as shown in Figure 6. This amethod with a gas reference sensor which can be used to calibrate the COSWOB. In order to calibrate the works by introducing gas of known concentrations (25, 50, 100 ppm) then performing a check with COSWOB, its output signals were also compared to the output signals of a reference sensor (LPT-Aa gas reference sensor which can be used to calibrate the COSWOB. In order to calibrate the COSWOB, COB, Critical Environment Technologies Canada Delta, Canada), which provides Critical a high its output signals were also compared to the outputInc., signals of aBC, reference sensor (LPT-A-COB, level of accuracy for the measurement of 0–100 ppm CO with 4–20 mA of output. A precision 150Environment Technologies Canada Inc., Delta, BC, Canada), which provides a high level of accuracy for ohm resistor with a 0.1% tolerance is connected to the output terminal of the LPT-A-COB reference the measurement of 0–100 ppm CO with 4–20 mA of output. A precision 150-ohm resistor with a 0.1% sensor. Theisvoltage drop the precision is measuredreference by data acquisition tolerance connected toacross the output terminalresistor of the LPT-A-COB sensor. The(DAQ) voltage(USBdrop 6281, National Instrument, Austin, TX, USA), and NI SingnalExpress is used to analyze across the precision resistor is measured by data acquisition (DAQ) (USB-6281, National acquisition Instrument, data. Austin, TX, USA), and NI SingnalExpress is used to analyze acquisition data. Figure6.6.Schematic Schematicdiagram diagramof ofaagas gassensor sensorcalibration calibrationsystem systemfor forthe theCOSWOB. COSWOB. Figure Sensors 2016, 16, 1568 7 of 11 5. Results and Discussion The micro-hydropower generator consists of a rotor with a circular permanent magnet, Sensors 2016, 16, 1568 7 ofa11stator, and a three-phase AC-DC converter. The stator used the delta connection method to connect windings. 5. Results and Discussion The rotor is placed in a flow chamber of the micro-hydropower generator, and blades of the rotor are perpendicular to the flow direction the tap Blades on thepermanent rotor are magnet, angled to transform The micro-hydropower generator of consists of awater. rotor with a circular a stator, energy the tap flow stream into rotational energy. generator is designed andfrom a three-phase AC-DC converter. The stator usedThe themicro-hydropower delta connection method to connect windings. TheCOSWOB rotor is placed a flowinchamber of the micro-hydropower generator, andin blades of to operate in the thatinflows only one direction. Attempting operation the reverse the rotor are perpendicular to the flow direction of the tap water. Blades on the rotor are angled direction does not damage the micro-hydropower or other components of the COSWOB. No to output transform from the tap flow stream into energy. micro-hydropower generator voltage of the energy micro-hydropower generator wasrotational observed as an The operation in the opposite direction is designed to operate in the COSWOB that flows in only one direction. Attempting operation in the because the rotation efficiency of angled blades in the reverse direction is very small. In order to add reverse direction does not damage the micro-hydropower or other components of the COSWOB. No the maximum energy to the tap water, the inlet and outlet sizes of the flow chamber are different to output voltage of the micro-hydropower generator was observed as an operation in the opposite improve rotorbecause rotationthe efficiency. Moreover,ofthe O-ring between cap and flow is chamber keepInwater direction rotation efficiency angled blades in thethe reverse direction very small. fromorder leaking out form the flow chamber of the micro-hydropower generator. A circular permanent to add the maximum energy to the tap water, the inlet and outlet sizes of the flow chamber are magnet is adhered to therotor inside of theefficiency. rotor blade. The circular permanent different to improve rotation Moreover, the O-ring betweenmagnet the capwill andalso flowmove, producing a rotating magnetic field as revolves around rotor. A sinusoidal waveform is chamber keep water from leaking outwater form flow the flow chamber of thethe micro-hydropower generator. A circular permanent is adhered to the inside of the rotoras blade. The in circular permanent magnet observed between anymagnet two winding terminals of the stator, shown Figure 7a, and the phase of ◦ . revolves will also move, producing rotating as water flow around of thethe rotor. A the output waveform from thea stator aremagnetic equally field divided into 120 The frequency sinusoidal sinusoidal waveform is observed between any two winding terminals of the stator, as shown in waveform as a function of the water flow rate, measured in the flow rate, is in the range 1.7–5.2 L/min, Figure 7a, and the phase of the output waveform from the stator are equally divided into 120°. The as shown in Figure 7b. Clearly, a linear dependence of the frequency of the sinusoidal waveform on the frequency of the sinusoidal waveform as a function of the water flow rate, measured in the flow rate, water flow rate is observed. The three-phase full-wave rectifier consists of a three-phase diode bridge, is in the range 1.7–5.2 L/min, as shown in Figure 7b. Clearly, a linear dependence of the frequency of comprising six diodes. It converts three-phase ACis voltage from stator to the DC voltage the sinusoidal waveform on theawater flow rate observed. Thethe three-phase full-wave rectifierthat is periodic over one-sixth of the input AC voltage cycle. To get ripple-free DC voltage, it is to consists of a three-phase diode bridge, comprising six diodes. It converts a three-phase AC regulated voltage a ripple-free voltage using an LDO regulator. The output voltage the LDO as from the DC stator to the DC voltage that islinear periodic over one-sixth of the input AC of voltage cycle.regulator To get ripple-free DCwater voltage, it israte regulated to ain ripple-free DCThe voltage using an LDO a function of the flow is shown Figure 7c. output voltage oflinear LDOregulator. regulatorThe reaches output voltage of 5the as a its function ofpower the water flowW. rate shown voltage in Figurebecomes 7c. The tiny a saturation value of V LDO at 4.2regulator L/min, and output is ~2.5 Itsisoutput output voltage of LDO regulator reaches a saturation value of 5 V at 4.2 L/min, and its output under a flow rate of 1.7 L/min. We also test flow rate of a few domestic water pipelinespower by a flow is ~2.5 W.(LFE1A3F1, Its output voltage tiny under a flow rate of 1.7 3.3 L/min. We also test flow rate of means a switch meter SMC,becomes Tokyo, Japan), which ranges from L/min to 8.2 L/min. This few domestic water pipelines by a flow switch meter (LFE1A3F1, SMC, Tokyo, Japan), which ranges that normal domestic water can directly drive the micro-hydropower generator to power the COSWOB. from 3.3 L/min to 8.2 L/min. This means that normal domestic water can directly drive the microAdditionally, we can also measure the water flow rate by monitoring the frequency of the sinusoidal hydropower generator to power the COSWOB. Additionally, we can also measure the water flow waveform one loop of stator or voltagewaveform of the LDO regulator. the linear rate byofmonitoring thethe frequency of output the sinusoidal of one loop of Nevertheless, the stator or output rangevoltage of theofrelationship between output voltage of the LDOofregulator and the flow rate is from the LDO regulator. Nevertheless, the linear range the relationship between output 1.7 L/min The linear ofrate theisrelationship between the frequency the of sinusoidal voltagetoof3.6 theL/min. LDO regulator and range the flow from 1.7 L/min to 3.6 L/min. The linearof range the relationship of the any loop stator and flow that waveform of anybetween loop ofthe thefrequency stator and the sinusoidal flow rate waveform is from 1.7ofL/min toof 5.3the L/min. Thisthe means rateestimate is from the 1.7 L/min 5.3 L/min. means that we can estimateofthe usage based on theof any we can water to usage basedThis on the monitoring frequency thewater sinusoidal waveform frequency of the sinusoidal waveform of any loop of the stator. loop monitoring of the stator. Figure 7. (a)7.The output signals ofofthree statoratataaflow flowrate rate L/min; (b) Frequency Figure (a) The output signals threeloops loops of of the the stator of of 1.71.7 L/min; (b) Frequency of theofsinusoidal waveform ofof the loopsof ofthe thestator stator a function of water the water flow rate; the sinusoidal waveform theone oneof ofthree three loops asas a function of the flow rate; (c) output The output voltage of the LDO regulatoras as aa function function of flow rate. (c) The voltage of the LDO regulator ofthe thewater water flow rate. Sensors 2016, 16, 1568 8 of 11 The flow rate in a real domestic water supply system is not constant. The variable flow rate results Sensors 2016, 16, 1568 8 of 11 in unstable voltage generated by the micro-hydropower generator. It will degrade the performance The flow rate in a real domestic water supply system is not constant. The variablethis flowissue. rate results of the COSWOB. Therefore, we designed a novel power module to overcome At first, in unstable voltage generated by the micro-hydropower generator. It will degrade the performance of we evaluated that the power consumption of the COSWOB, which includes a three-lead electrochemical COSWOB. Therefore, weand designed a novel power module overcome issue.inAttotal. first, In weour CO the sensor, a readout circuit, a microcontroller with BLE, to was aroundthis 12 mW evaluated that the power consumption of the COSWOB, which includes a three-lead electrochemical power module, the capacitance of a supercapacitor in the energy harvesting with a boost charger can CO sensor, a readout circuit, and a microcontroller with BLE, was around 12 mW in total. In our power dominate the operating time of the COSWOB when the energy harvesting from the micro-hydropower module, the capacitance of a supercapacitor in the energy harvesting with a boost charger can dominate generator is tiny; even domestic water is off in the short duration time. We used a 0.1-F supercapacitor the operating time of the COSWOB when the energy harvesting from the micro-hydropower as energy storage to store the unstable input energy from the micro-hydropower generator. Five volts generator is tiny; even domestic water is off in the short duration time. We used a 0.1-F supercapacitor is the and meanttoto remain stored ininput the supercapacitor the COSWOB. Therefore, maximum as limit energy storage store the unstable energy from thein micro-hydropower generator.the Five volts ideal energy stored in the supercomputer is 1.25 J, based on Equation (1). Therefore, 1.25 J equals is the limit and meant to remain stored in the supercapacitor in the COSWOB. Therefore, the to 1.25maximum W/s. When theenergy input stored powerinisthe zero, the energy ofisthe supercapacitor the COSWOB ideal supercomputer 1.250.1-F J, based on Equationcan (1). offer Therefore, 1.25 J an operating time of at least 1.5 min. equals to 1.25 W/s. When the input power is zero, the energy of the 0.1-F supercapacitor can offer the COSWOB an operating time of at least 1.5 min. 1 2 2 Estored 1.25 J.J. E = 2=CVCV ==0.5 0.5××0.1 0.1× × 55 = = 1.25 (1) (1) intelligent hydroelectricenergy energy harvesting harvesting technology composes an an energy TheThe intelligent hydroelectric technologyininthis thiswork work composes energy harvesting circuit, an energy storage element, a battery threshold control circuit, and DC/DC harvesting circuit, an energy storage element, a battery threshold control circuit, and DC/DC converters. Theinput inputcan can be be any any non-stable but should be greater thanthan 330 mV. chart converters. The non-stablevoltage voltage but should be greater 330 A mV. A of chart charges of the supercapacitor with different input voltages is shown in Figure 8a. We experimented of charges of the supercapacitor with different input voltages is shown in Figure 8a. We experimented with six different input voltages: 3 V, 2.5 V, 2 V, 1.5 V, 1 V, and 0.5 V. The charging times of the output with six different input voltages: 3 V, 2.5 V, 2 V, 1.5 V, 1 V, and 0.5 V. The charging times of the output voltage that can reach 5 V are around 7 s, 9 s, 12 s, 23 s, 56 s, and 275 s, respectively. The charge curve voltage that can reach 5 V are around 7 s, 9 s, 12 s, 23 s, 56 s, and 275 s, respectively. The charge curve has has two slopes of charging rate because the energy harvesting circuit has two kinds of charging twomethods: slopes ofacharging rate because the energy circuit has two charging methods: trickle current charge and a main harvesting boost charge. Morever, the kinds outputofvoltage of microa trickle current charge and a main boost charge. Morever, the output voltage of micro-hydropower hydropower generator is around 1 V (flow rate = 1.9 L/min) and the voltage of the energy storage generator around 1 V (flow = 1.9 and condition, the voltagea of the energy storage element can be elementis can be charge to rate almost 4.6L/min) V. In this microcontroller module (wireless charge to almost 4.6function) V. In this condition, a microcontroller module (wireless communication function) communication and sensor module (sensor function) cannot work under a supply voltage andof sensor module (sensor function) cannot work under a generator supply voltage of around 1 V when we only around 1 V when we only used the micro-hydropower to supply the COSWOB without a power module. Figure 8 generator shows the to sensing system operation process 1 V ofmodule. chargingFigure and 8 used the micro-hydropower supply the COSWOB withoutwith a power discharging. Thesystem sensing system turns to awith working state when the of the supercapacitor is shows the sensing operation process 1 V of charging andvoltage discharging. The sensing system greater than 4.5 V, andwhen it turns offvoltage when the of the supercapacitor lower4.5 than 2.3 V.itIn this off turns to a working state the of voltage the supercapacitor is greateris than V, and turns work, we use of the supercapacitor that than can still around when when the voltage the0.1-F supercapacitor is lower 2.3 V.work In this work, 75 wesuse the the 0.1-Fhydroelectric supercapacitor generator is turned off. that can still work around 75 s when the hydroelectric generator is turned off. Figure 8. (a) The charge curve differentinput inputvoltages; voltages; The sensing system Figure 8. (a) The charge curveofofsensing sensingsystem system with with different (b)(b) The sensing system operation process. operation process. Sensors Sensors 2016, 2016, 16, 16, 1568 1568 99 of of 11 11 For a calibration of the COSWOB, it is compared with a gas reference sensor. The output voltage a calibration of the COSWOB,reference it is compared with aupon gas reference Thewas output voltage of theFor COSWOB and the LPT-A-COB gas sensor exposuresensor. to CO gas monitored of the COSWOB and the LPT-A-COB reference gas sensor upon exposure to CO gas was monitored and recorded as the output voltage. In Figure 9a, the dynamic output voltage versus time for both and recorded as the output voltage. Inconcentration Figure 9a, thevarying dynamicfrom output voltage versus for both sensors is demonstrated, with the CO 25 to 100 ppm. Notime signification sensors is was demonstrated, CO concentration varying fromto25estimate to 100 ppm. No signification hysteresis observed inwith boththe sensors. We used linear regression the parameters of that hysteresis was observed in both sensors. We used linear regression to estimate the parameters relationship between the COSWOB and the LPT-A-COB and obtained formulas for linear regression. of that relationship the COSWOB theasLPT-A-COB and obtained formulas linear The R-squared valuebetween (R2 = 0.9973) gets closer and to one, shown in Figure 9b; it indicates thatfor a perfect 2 = 0.9973) gets closer to one, as shown in Figure 9b; it indicates regression. The R-squared value (R linear relationship between the COSWOB and the LPT-A-COB. We know the sensitivity of the LPTthat a perfect linear relationship between COSWOB between and the LPT-A-COB. We know sensitivity A-COB is ~11.9 mV/ppm. After getting thethe relationship the two sensors, we canthe calibrate the of the LPT-A-COB is ~11.9 After getting the relationship between the twoair; sensors, wewas can COSWOB. The first, a biasmV/ppm. of the COSWOB, was performed by introducing free the bias calibrate COSWOB. Thewas first,then a bias of the COSWOB, wasthe performed byofintroducing freewas air; ~0.509 V. the A zero calibration made. After calculation, sensitivity the COSWOB the bias was ~0.509 A zero calibration calculation, the sensitivity of the ~1.03 ppm/mV. AfterV.calibrating, Equationwas (2) then can made. be usedAfter to calculate the CO level in ambient COSWOB was ~1.03 ppm/mV. After calibrating, Equation (2) can be used to calculate the CO level in conditions: ambient conditions: CO (ppm) = 1.03 × V − 509, (2) COLevel (ppm) = 1.03 × VCOSWOB − 509, (2) where COLevel is CO concentration in ambient conditions, and VCOSWOB is the output voltage of the where CO is CO concentration in ambient conditions, and VCOSWOB is the output voltage of the COSWOB Level and is in mV. COSWOB and is in mV. (a) Dynamic Dynamic response response curves curves of of the the output output voltage voltage of of the the COSWOB COSWOB and and the the LPT-A-COB Figure 9. (a) concentrations ranging fromfrom 25 to 1000 Linear(b) regression of the relationship versus time timefor forCO CO concentrations ranging 25 toppm; 1000(b)ppm; Linear regression of the between the COSWOB and the LPT-A-COB. relationship between the COSWOB and the LPT-A-COB. Once Once the the application application on on the the intelligent intelligent gateway gateway is is launched, launched, the the application application automatically automatically started started to discover COSWOBs nearby. After finding the COSWOBs, its status is updated instantly to discover COSWOBs nearby. After finding the COSWOBs, its status is updated instantly on on the the main main page of the server application, as shown in Figures 10a–c. If the COSWOB is OFF, the indicator icon page of the server application, as shown in Figure 10a–c. If the COSWOB is OFF, the indicator icon on on the the main main page page of of the the application application displays displays OFF OFF status, status, as as shown shown at at row row == 11 and and column column ==11in inFigure Figure10a. 10a. The The application applicationrepeats repeatsto toscan scanany anywake-up wake-upCOSWOB COSWOBevery every3030s s(not (notoptimal) optimal) until until the the activated activated COSWOB is connected to the intelligent gateway. After the COSWOB is connected to the intelligent COSWOB is connected to the intelligent gateway. After the COSWOB is connected to the intelligent gateway, gateway, the the indicator indicator icon icon on on the the main main page page quickly quickly changes changes to to the the ON ON status status from from the the OFF OFF status, status, and the application application starts startstotocontinuously continuouslyreceive receivethe the data from COSWOB every second. When and the data from thethe COSWOB every second. When the the CO level exceeds a dangerous level (50 ppm), the indicator icon is quickly changed to the alarm CO level exceeds a dangerous level (50 ppm), the indicator icon is quickly changed to the alarm icon, icon, as shown in Figure 10c. Meanwhile, an information of COSWOB the COSWOB automatically pops as shown in Figure 10c. Meanwhile, an information pagepage of the automatically pops up, up, the CO level around the natural gas water heater is real-time displayed on a chart, and the user the CO level around the natural gas water heater is real-time displayed on a chart, and the user can can overview the history thelevel, CO level, as shown in Figure 10d. Thealarms Wi-Fi (TuneBox, alarms (TuneBox, overview the history of theofCO as shown in Figure 10d. The Wi-Fi Nexum, Nexum, Taipei, in Taiwan) in location the other location will generate a loud sound alert other people. Taipei, Taiwan) the other will generate a loud sound to alert othertopeople. Additionally, Additionally, the value of the CO level is also transmitted to smartphones that belong to other family the value of the CO level is also transmitted to smartphones that belong to other family members; members; the mobile application will alert family members to take action when the CO level reaches the mobile application will alert family members to take action when the CO level reaches a dangerous alevel, dangerous level, as shown as shown in Figure 11. in Figure 11. Sensors 2016, 16, 1568 Sensors 2016, 16, 1568 Sensors 2016, 16, 1568 10 of 11 10 of 11 10 of 11 Figure Screenshotsofofthe themain main pages pages of of the application gateway, various Figure 10. runinin the intelligent gateway, various Figure 10.10.Screenshots pages of the application applicationrun inthe theintelligent intelligent gateway, various states: (a) COSWOB OFF; (b) COSWOB ON; and (c) alarm activated; (d) Screenshot of the information states: (a) COSWOB OFF; (b) COSWOB ON; and (c) alarm activated; (d) Screenshot of the information states: (a) COSWOB OFF; (b) COSWOB ON; and activated; Screenshot of the information COSWOB. of the COSWOB. of of thethe COSWOB. Figure 11. forthe theCOSWOB COSWOBdata datavisualization. visualization. Figure 11.Screenshots Screenshotsof ofthe themobile mobile application application for Figure 11. Screenshots of the mobile application for the COSWOB data visualization. 6. 6.Conclusions Conclusions WeWe have demonstrated thatthat our our application will be for people to avoidtoCO poisoning 6. Conclusions have demonstrated application willa great be a benefit great benefit for people avoid CO from gas water heaters. Our application not only alerts users and their family members inside poisoning from gas water heaters. Our application not only alerts users and their family membersthe We have demonstrated that our application will be a great benefit for people to avoid CO house, but also notifies members outside the house, when levels reach inside the house, but family also notifies family members outside the CO house, when COa dangerous levels reachlevel. a poisoning from gas water heaters. Our application not only alerts users and their family members dangerous Moreover, our COSWOB is only powered by thegenerator. micro-hydropower Moreover, ourlevel. COSWOB is only powered by the micro-hydropower There is nogenerator. danger that inside the house, but also notifies family members outside the house, when CO levels reach a Theremight is no danger that people might be living withor a dangerous CO level, might or thatnot the COSWOB people be living with a dangerous CO level, that the COSWOB work due might to a loss dangerous level. Moreover, our COSWOB isdomestic only powered bypetroleum the micro-hydropower generator. not work due to a loss of battery life. The liquefied gas leakage is also an of battery life. The domestic liquefied petroleum gas leakage is also an important issue; the sensor There is no danger that people might be living with a dangerous CO level, or that the COSWOB might important the sensor module of thewith COSWOB canelectrochemical easily integratenatural with gas three-lead module of theissue; COSWOB can easily integrate three-lead (NG) or notelectrochemical work due to natural a loss gas of battery life. The domestic liquefied petroleum gas leakage is atalso (NG) sensor. next step to monitor the an LPG sensor. The next step will beor toLPG monitor COThe and LPG atwill the be same time inCO theand nextLPG generation important issue; the sensor module of the COSWOB can easily integrate with three-lead time in the next generation of COSWOBs. of same COSWOBs. electrochemical natural gas (NG) or LPG sensor. The next step will be to monitor CO and LPG at the same time in the next generation of COSWOBs. Sensors 2016, 16, 1568 11 of 11 Acknowledgments: This work was supported by National Science Council, Taiwan. Author Contributions: All authors have contributed to the presented work. Chen-Chia Chen, Gang-Neng Sung, Wen-Ching Chen, Chieh-Ming Wu, and Chun-Ming Huang designed the circuit. Jin-Ju Chue designed the form factor of the COSWOB. The applications of the server and mobile devices are designs by Chih-Ting Kuo. All authors have read, written, and revised the manuscript. Conflicts of Interest: The authors declare no conflict of interest. References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. Galinina, O.; Mikhaylov, K.; Andreev, S.; Turlikov, A.; Koucheryavy, Y. Smart home gateway system over Bluetooth low energy with wireless energy transfer capability. EURASIP J. Wirel. Comm. 2015, 2015, 178–196. [CrossRef] Bourobou, S.T.M.; Yoo, Y. User Activity Recognition in Smart Homes Using Pattern Clustering Applied to Temporal ANN Algorithm. Sensors 2015, 15, 11953–11971. [CrossRef] [PubMed] Ghayvat, H.; Mukhopadhyay, S.; Gui, X.; Suryadevara, N. WSN- and IOT-Based Smart Homes and Their Extension to Smart Buildings. Sensors 2015, 15, 10350–10379. [CrossRef] [PubMed] Yang, J.; Zhou, J.; Lv, Z.; Wei, W.; Song, H. A Real-Time Monitoring System of Industry Carbon Monoxide Based on Wireless Sensor Networks. Sensors 2015, 15, 29535–29546. [CrossRef] [PubMed] Wang, Z.; Watkins, D.W.; Ong, K.G.; Shi, X. Cyber-physical systems for water sustainability: Challenges and opportunities. IEEE Commun. Mag. 2015, 53, 216–222. [CrossRef] Chaiwatpongsakorn, C.; Lu, M.; Keener, T.C.; Khang, S.J. The Deployment of Carbon Monoxide Wireless Sensor Network (CO-WSN) for Ambient Air Monitoring. Int. J. Environ. Res. Public Health 2014, 11, 6246–6264. [CrossRef] [PubMed] Abraham, S.; Li, X. Cost-Effective Wireless Sensor Network System for Indoor Air Quality Monitoring Applications. Procedia Comput. Sci. 2014, 34, 165–171. [CrossRef] Spachos, P.; Hatzinakos, D. Real-Time Indoor Carbon Dioxide Monitoring through Cognitive Wireless Sensor Networks. IEEE Sens. J. 2016, 16, 506–514. [CrossRef] Ilano, A.L.; Raffin, T.A. Management of carbon monoxide poisoning. Chest 2015, 97, 165–169. Accident Carbon Monoxide Poisoning and Its Prevention; Taiwan Epidemiology Bulletin; Centers for Disease Control: Taipei, Taiwan, 1995; pp. 1–5. Nest Protect. Available online: https://nest.com/smoke-co-alarm/meet-nest-protect/ (accessed on 28 May 2016). Kidde Smoke and CO Alarm. Available online: http://www.wink.com/products/kidde-smoke-and-coalarm/ (accessed on 28 May 2016). Honeywell 5800CO. Available online: http://www.security.honeywell.com/hsc/products/intruderdetection-systems/life-safety/wireless-carbon-monoxide-detector/242073.html (accessed on 24 July 2016). Wireless CO Detector (WS4913). Available online: http://www.dsc.com/index.php?n=products&o=view& id=129 (accessed on 24 July 2016). Liu, D.H.; Wang, R.; Yao, K. Design and Implementation of a RF Powering Circuit for RFID Tags or Other Batteryless Embedded Devices. Sensors 2014, 14, 14839–14857. [CrossRef] [PubMed] Torres, R.L.S.; Visvanathan, R.; Hoskins, S.; Hengel, A.V.D.; Ranasinghe, D.C. Effectiveness of a Batteryless and Wireless Wearable Sensor System for Identifying Bed and Chair Exits in Healthy Older People. Sensors 2016, 16, 546–563. [CrossRef] [PubMed] Zhang, C.; He, X.F.; Li, S.Y.; Cheng, Y.Q.; Rao, Y. A Wind Energy Powered Wireless Temperature Sensor Node. Sensors 2015, 15, 5020–5031. [CrossRef] [PubMed] Hande, A.; Polk, T.; Walker, W.; Bhatia, D. Self-Powered Wireless Sensor Networks for Remote Patient Monitoring in Hospitals. Sensors 2006, 6, 1102–1117. [CrossRef] Tasic, V.; Staake, T.; Stiefmeier, T.; Tiefenbeck, V.; Fleisch, E.; Troster, G. Self-powered Water Meter for Direct Feedback. In Proceedings of the 2012 3rd International Conference on the Internet of Things, Wuxi, China, 24–26 October 2012. © 2016 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).