We have designed and implemented a platform that enables monitoring and actuation in multiple buildings, that has been utilised in the context of a research project in Greece, focusing on public school buildings. The Green Mindset project has installed IoT devices in 12 Greek public schools to monitor energy consumption, along with indoor and outdoor environmental parameters. We present the architecture and actual deployment of our system, along with a rst set of ndings.
There are basically two routes towards cutting down on our carbon emissions:
use more energy-e cient infrastructure/materials, or promote behavioral change
on people living/working inside buildings, in order to consume less energy. The
second approach is a solution to reach the envisioned environmental goals in
reasonable time. In general, about 75% of buildings in Europe are residential [
Our system provides real-time monitoring and actuation of multiple buildings over an IoT infrastructure. Our current implementation focuses more on energy consumption and environmental parameters monitoring, but can be tailored in order to support other applications and scenarios. In this context, it is currently being used as the basis for a Greek research project targeting energy e ciency in public school buildings. It is based on the installation of custom IoT infrastructure in a number of Greek public school buildings, monitoring energy consumption and environmental parameters, supplemented by a set of software
Copyright 2015 for this paper by its authors. Copying permitted for private and academic purposes tools aiming to help in educating students on energy and environmental matters and also achieve better energy e ciency in buildings. 2
We wanted our system to be able to monitor and manage IoT infrastructure dispersed throughout several buildings in real-time and while using a single interface. The infrastructure is also spread out through several rooms inside the buildings. Moreover, we wanted our system to provide real-time update on what is happening in the real world, which on the one hand translates to storing vast quantities of data (since data are then produced by IoT nodes with a high rate), and on the other hand means e cient implementation practices throughout the whole stack of layers comprising the system.
Moving over to the implementation from an end-user point of view, the system aims to support di erent end-user groups. E.g., in the speci c case of public school buildings, there are several such groups that can be identi ed: the students, the educators, the school building administrators and the Ministry of Education administrative sta . There is thus a need for the interface of the system to provide services and information in a way that suits all of these di erent categories of end-users. Lastly, we wanted our system to be expandable and easy to interface with other systems or components. There are numerous approaches and additional applications that can be implemented on top of the system.
Regarding the overall architecture of the system, at the lower layers we utilize multiple installations at school buildings. Each one of these installations consists of a multitude of IoT nodes (over 10) that communicate with the cloud infrastructure directly or via an IoT gateway device. The IoT gateways help to coordinate the IoT nodes and enable communication with the part of the system that is not capable of direct communication with the Internet, i.e., the IEEE 802.15.4 nodes that monitor the environment inside the buildings.
On the building level, the IoT nodes form an adaptive tree routing path to the gateway device over which they report their measurements to the gateway. On the Internet level, all information and control messages from and to the IoT nodes is passed on to a message bus system that is responsible for distributing the information gathered to the various subsystems responsible for storing the data, processing the data or generating noti cations. On the data storage level, we use a distributed architecture with multiple services that o er us the ability to store data in multiple levels. With respect to interfacing the system with the end-users and the general public, we have implemented 2 web portals.
The overall design for the installation of IoT infrastructure in the school buildings participating in the project follows the same pattern inside each building: a) electricity consumption meters are installed to monitor the consumption of the building as a whole, or speci c oors/sectors, b) a subset of the school's classrooms, or other rooms, have been tted with IoT nodes, c) a set of gateway nodes provide bridging to the Internet, and, d) all IoT nodes communicate wirelessly with each other and the gateways. The system is complemented by the installation of IoT sensing and actuator infrastructure inside 2 o ce buildings, with devices also linked to the HVAC equipment., o ering remote control of both air-conditioning/heating and lights. Apart from the indoor IoT infrastructure, there is a weather station and an air pollution monitoring station installed outdoors in each of the participating buildings. As mentioned, we aimed for a system functioning over a heterogeneous IoT infrastructure. We utilised custom IoT nodes, built on open-source hardware and software. Additional hardware logic interconnects the sensors utilized for each class, i.e., humidity, temperature, luminosity, CH4 and CO sensors, together with a noise level and PIR (motion) sensor. A custom plastic casing houses the hardware to enable easier installation in classrooms and o er additional safety to/from students and educators. The IoT nodes are AC powered, while they use an Xbee for communication. The weather station measurements are transmitted over Ethernet, also used to provide power. The radiation and environmental conditions monitors installed outside use WiFi to communicate with the Internet.
We implemented a web portal that enables a holistic view of the IoT infrastructure. End-users of the portal include educators and administrative sta from each school participating in the project, as well as administrative sta from the Greek Ministry of Educational A airs.The rst category of users have access only to data related to the respective school building, while the second ones have access to the entirety of the buildings. Visualization aspects also include some basic tools to add annotations to the produced data from the end-users, i.e., students, sta , parents, and also the ability to produce alerts from certain extraordinary events. In this way, end-users could provide annotations to the gathered data in order e.g., to classify certain events happening in the realworld and that could not be detected easily by software techniques, such as a spike in electricity consumption happening due to a school event. 4
6 primary, 5 secondary, 1 high school each node has 5 sensors students in all levels teachers in all levels classroom sensors (indoor) classroom sensors (indoor)
Concluding, our system follows a scalable architecture, built on open source technologies, that can be easily adapted and expanded. We have also developed several end-user interfaces that provide tools to both the educators and the administrators of such building facilities. Our rst results are promising, showing that the current incarnation of the system is more than adequate for these tasks and it already helps in discovering trends in energy consumption, as well as ndings in other cases. Regarding our future work, we intend to place an additional focus on the educational aspects.