Developed countries such as our European Community su er from the problem of air pollution: from the quality of air in public and private spaces, to the concentration of "sentinel" gas and the risk of res. The major sources of pollution can be easily observed and managed, such as heavy industries, power plants with non-renewable sources and pollution from road transport vehicles. On the other hand the rest of the air pollution sources are more rooted, speedily distributed and not easily reachable by technical checks from authorized public authorities: just think of domestic boilers and ues, which are present in most of apartments and o ces, need constant maintenance carried out by the owners; this causes, in most cases, a degradation of the systems, due to their negligence, which produces a decrease of e ciency and an increase in polluted gases production, up to the worst cases, of accidents and explosions. The recent incidents in Italy[1] and the interest in preserving the environment have led us to seek a solution to the problem, based on the qualitative monitoring of air quality, therefore of toxic and explosive agents, in small areas through a simple use, cost-e ective and batterypowered tools, able to be used by the majority of private citizens and institutions, in order to create a geolocalized network through which map air quality, even in areas with historical-cultural constraints, wooded and rural areas or otherwise di cult to reach environments (such as, for example, the exhaust ues of heating systems). Through the recent tencologies of IOT, Cloud Computing and Edge Computing, the aim of our research is to create a monitoring network based on individual cost-e cient sensor devices, capable of reaching and therefore mapping even the smallest and unreachable risk zones. The capillary network is also able to analyze and evaluate densely populated areas and is able to o er a monitoring service for private citizens, local and national authorities.
Copyright c by the paper's authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0). In: E. Calvanese Strinati, D. Charitos, I. Chatzigiannakis, P. Ciampolini, F. Cuomo, P. Di Lorenzo, D. Gavalas, S. Hanke, A. Komninos, G. Mylonas: Proceeding of the Poster and Workshop Sessions of AmI-2019, the 2019 European Conference on Ambient Intelligence, Rome, Italy, November 13-15, 2019.
According to the Federchimica last research on "Comparative research on gas emissions"[
UE Parliament council of 21 May 2008: "On ambient air quality and cleaner air for Europe" "In order to protect human health and the environment as a whole, it is particularly important to combat emissions of pollutants at source and to identify and implement the most e ective emission reduction measures at local, national and Community level. Therefore, emissions of harmful air pollutants should be avoided, prevented or reduced and appropriate objectives set for ambient air quality taking into account relevant World Health Organisation standards, guidelines and programmes." [10] 1.2
There is also another big problem associated with gas utilization: the growth of domestic accidents. According a recent \Statistic from gas accidents " elaborated by CIG [4] in 2017, there has been an increase of 5:5% of natural gas put in the Italian network and so an increase also in the accidents happened: a growth from 123 to 140. In particular accidents with explosions and res have growth from 83 to 106. The same trend is for LGP. Analyzing this survey we nd that in the 29% of accidents there was absence or non-suitability of smoke evacuation system and in the 27% the causes were an incorrect installation. The majority of accidents could be avoided with a more accurate and immediate alert system positioning near the boiler. 1.3
This type of problem can be faced by the recent growth and spread of IOT, Cloud and Edge Computing technologies. Our work aims therefore to the creation of cost-e cient tools for data acquisition and creation of a network can provide to citizens, private organizations and public authorities a comprehensive and extensive mapping and a broadcast warning system.
The realization of simple battery-powered qualitative sensors, with low consumption and minimum overall footprints, is the key to managing the problem of ue smoke control. We have ignored technical issues such as the accuracy and sensitivity of sensors for greater simplicity of implementation and use (following the paradigm of "easier, less expensive, less subject to breakage "1) but above all we intend to create a network information accessible to all for the di usion of the concept "smart cities" in conjunction with the energy saving, pollution control and monitoring of res and explosions. 2
1"Frustra t per plura quod eri potest per pauciora." Occam's razor. 2.1
In the context of low-power consumption, long life and battery-powered devices (for the realization of extremely portable devices), the choice of the transmission system fell on LoRa Network [5]. The technical details are given in the manuals available on the manufacturer's website[6]. Here, a brief list of the main reasons why our choice fell on this system:
Emergency signals must be sent and received without power supply.
Enclosed spaces, such as chimneys, have little space available for the construction of complex systems with spacious antennas and high power supply.
The data to be sent is made up of small information packets, so a low channel capacity is su cient. The data are not sensitive or highly accurate and transmission timing is not a speci c problem. The transmission must be extremely low power.
The transmission range must be long range and can vary greatly among the various environments of use. The network created must have future expansion capabilities.
Data should be geolocalizated easily.
LoRa Network is a long range transmission technology (up to 10 km in open space area) particularly well suited for IOT devices; its noise pro le is very low and is also characterized by a strong resistance to noise, working in fact below the noise threshold characteristic of the non-shielded environments. There are many lowcost solutions for "embedded " implementation of this transmission system, all with the same power consumption characteristics, range and cost compatible with our project. 2.2
There are many low-cost solutions for the "embedded" implementation of this transmission system, all with the same power consumption characteristics, range and cost compatible with our project. The M Q x series of sensors for the detection of various gases in the air has as main features:
Wide detecting scope Fast response High sensitivity stable and long life Simple pilot circuit Low energy consumption 2.2.1
They are used in air quality control equipment for buildings/o ces and are suitable for detection of N H3, N Ox, alcohol, benzene, smoke, CO, CO2 gasses and many more. Connection with the external components is very simple, but most importantly it is standard: this implies that the di erent devices implementation is possible only through the change of the sensor.This is a considerable advantage in terms of ease of use, implementation of management and low cost.
As noted in the sensor data, the output response (analog signal) strongly depends on the temperature and humidity conditions. This is an important problem but for our purposes (the acquisition of quantitative but not qualitative data) it can be solved by simple software algorithms, even if they are not extremely optimal. 2.2.2
Resistance value of M Q x series is di erence to various kinds and various concentration gases. So, sensitivity adjustment is very necessary. In order to calibrate the detector for 100 ppm N H3 or 50 ppm Alcohol concentration in air and use value of Load resistance about 20 K (10 K to 40 K ). When accurately measuring, the proper alarm point for the gas detector should be determined after considering the temperature and humidity in uence. 2.3
First, we have designed and built a prototype for generic use of the sensor, to which we added the transceiving system of low consumption and long-range LoRa[5] Network. The collected data is distributed in three "end-user services" in order to provide the same number of overall system capabilities:
Private telemetry City administrative certi cation and monitoring
Big data analysis.
We have con gured a dragino gateway and send data from our device to The Things Network [7]. In The Things Network servers, data are not collected because is only a hub for distribute it over the web. TTN has several plugins to redistribute data, depending which services you would use; in our case we used two di erent plugins always based on REST and MQTT protocols. 3
The use of devices with the characteristics described above (in short: small footprint, ease of construction, ease of use, low energy consumption, long-range connection and low noise pro le, geo-localization capability) for the acquisition of data, allows the creation of a useful database with the following features:
Local and global air pollution maps Statistical analysis of the presence of various polluting and non-polluting gases Fire risk monitoring in rural, wooded or sparsely populated areas Monitoring of densely populated urban areas with historical and cultural constraints Exhaust gas monitoring and analysis of boilers in private and public environments Boiler operation and e ciency monitoring for the user Alarm network located in large urban and non-urban areas
Quality Mapping urban areas and places of public and private workplaces 3.1
We have understood that low-cost IOT technology can meet the three great needs of a modern smart city: A robust and widespread alarm system, surveillance and security tools for local administrations Quality control tools for the certi cation of domestic systems
A database accessible to all, to manage all this kind of data
The solution to these three major problems inevitably leads to the growth of eco-sustainability, informalization and security of the entire urban environment. For this reason we focused on the choice of as many three services realized in an "ad hoc" way to respond to these three major problems in order to emphasize the qualities of our idea. 3.1.1
We decided to use a structure as much as possible common to all three requirements described above. The basic operation is simple: the LoRaWAN device connects with TheThingsNetwork [7] service that collects various information and increases with data such as geo-location, user, etc. The message is then routed to the database DynamoDB [8]. Data collected and stored in this database can be made available to the public Both private and public organizations for analyzing pollution purpose, through the Restfull API: a common API easy to use and very often implemented in various web services.
Amazon DynamoDB is a key-value and document database that delivers single-digit millisecond performance at any scale. It's a fully managed and durable database with built-in security, backup and restore, and in-memory caching for internet-scale applications. It can handle more than 10 trillion requests per day and can support peaks of more than 20 million requests per second. 3.1.2
For the use of the private user, a simple smartphone application accesses the database, thus creating a simple analysis and control tool, low-budget and easy-to-use. 3.1.3
We decided to use Elastic Search[9]: a search engine developed in Java, accessible via http request (or web interface) and a schema-free JSON documents. In case of public dataset with low-sensitive data collection, we provided a very powerful solution in search capability with near real-time search. Overall this service has tools like Beats (a very useful tool for transferring IOT data to the engine) and an analytics and visualization tool named Kibana[9]. Kibana is an extreme powerful tool for viewing data in real time, implementing machine learning algorithms evaluations and send noti cation on a trend. 4
We tested our prototype, made with various development tools for everyone, achieving remarkable results in terms of:
Design cost Design, analysis and testing times Cost of realization Robustness of signals and sensors Quality of data sent and received Speed of transmission protocols Ability to use the end user both as a data analysis, as a smartphone application, and as a map of the sensor network.
We then collected a lot of data from a single home chimney in real time: the connection between TheThingsNetwork Amazon DynamoDB and Elastic Search allowed us to analyze the data in three di erent ways. Through Kibana and the use of numerous devices it will also be possible to implement Machine Learning systems that can prevent the risk of re and alarm in a preventive way: something that we have moved into "future developments". [4] http://www.expoclima.net/focus/eventi/gli_incidenti_da_gas_in_italia_sono_in_aumento_ _sia_quelli_da_gas_naturale_che_da_gpl.htm [10] DIRECTIVE 2008/50/EC OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of 21 May 2008 on ambient air quality and cleaner air for Europe