Development of a small-scale and low cost sensing device towards mobile urban air quality monitoring

Abstract

With air pollution escalating worldwide new approaches are required to enable individual air quality monitoring. In this work the Smart Bike Assistant, a mobile device capable of measuring the most common urban pollutants, namely carbon monoxide, nitrogen dioxide and particulate matter, has been developed with focus on urban bicycle commuters as target users. A customized printed circuit board has been designed that houses the microcontroller and sensors on a small base area of 48x34mm2. By installing the Arduino Micro’s bootloader onto the embedded microcontroller, the convenience of the Arduino environment could be combined with the advantages of a lean customized system. Selecting only low-cost and standard components the parts costs total roughly 30USD at a device weight of less than 40 grams. The installed carbon monoxide, nitrogen dioxide sensors have been calibrated using professional laboratory instruments to pollutant concentration in the expected range. Due to unavailability of adequate equipment, temperature and humidity dependent calibration has not been carried out thus far. The employed dust sensor has been calibrated against the data of a related research paper. [25] De-noising steps and temperature and drift coefficients applied ensure a more reliable output signal. The board furthermore comprises a temperature and humidity sensor and uncalibrated ozone sensor. Additionally a casing concept has been developed that allows convenient mounting of the device on a bicycle and protection against environmental influences. Therefore special consideration has been given to allowing a proper air flow through the device. This is necessary to reduce response time and distortion of the gas sensor readings. To provide a comprehensible feedback the raw sensor readings are converted into values of the standardized Air Quality Index and forwarded to an Android device. The data transfer is established through a Serial USB connection from the device to an Android smartphone. Accordingly the SBA does not require a battery, since it is powered by the hosting smartphone. Besides a simple readout the customized Android application allows uploading of the sensor data and derivation of a rudimentary pollution map. At a later stage the application will be equipped with further features, such as gathering pollution data from nearby public measurement stations or implementation of an incentive scheme for active users. In a field test the general feasibility of the concept has been proven. Constraints have been discovered considering the conducted electrochemical gas sensor calibration. Low ambient temperatures led to a large offset with the CO sensors. This issue can be addressed by more sophisticated multi-point calibration in the future. Due to the project’s substantial extent, this work has the nature of a groundwork and leaves several challenges open for further development in the field of hardware as well as software. A larger field study with voluntary cyclists at a later stage will allow identifying the SBA’s performance at realistic conditions.