//////////////////////////////////////////////
// Data collection Specification
//////////////////////////////////////////////
This folder contains datasets of network features, extracted from WSN measurements at an industrial environment during the period 05-06/06/2014. 

Contributors:
Mikel Azkune, Applied Photonics Group, University of the Basque Country, Spain 
Athanasia Panousopoulou, Institute of Computer Science, Foundation for Research and Technology-Hellas.

Contact address: apanouso@ics.forth.gr

If you find any of this dataset useful for your research, please give credit in your publications where it is due.


**********************************************
Section 1: Network Configuration
**********************************************

1. Sensor Nodes Platform: AdvanticSyS XM1000 and CM5000-SMA (IEEE 802.15.4-compliant devices)
2. Network Size: 10 sensor nodes (node id=1,2,..10) and a sink node (running on a pandaboard).
3. Transmission Period: 6sec
4. Protocol Stack: Customized, based on Contiki OS-version 2.6- and incorporating the IETF Standard for Routing over Low-Power and Lossy Networks. For more information (see the acknowledgement section).

**********************************************
Section 2: Type of Traffic
**********************************************
Network monitoring traffic including, metrics from the Physical, MAC, and NWK layers, as well as ambient sensors readings (temperature and humidity) and voltage threshold on each sensor node.
The traffic recorded is related to the links established at the application layer between each node (id = 1,2,..10) and a sink node. 
The traffic is recorded at the sink node.

**********************************************
Section 3: Data Format
**********************************************

Each csv file relates to the network traffic recorded at the sink node, and originated by each individual sensor node.

The naming convention of each .csv file is defined as:
node"i".csv, where
"i" stands for id of the sensor node. 
 
Each .csv file contains 18 columns:
1. mean value of received signal strenght over the multi-hop path between node i and sink node (dBm)
2. standard deviation of received signal strenght over the multi-hop path between node i and sink node
3. mean value of link quality indicator over the multi-hop path between node i and sink node
4. standard deviation of link quality indicator over the multi-hop path between node i and sink node
5. mean value of the noise floor over the multi-hop path between node i and sink node (dBm)
6. standard deviation of the noise floor over the multi-hop path between node i and sink node 
7. transmission rate at the MAC layer (bpm) at the i-th node (unicast traffic only)
8. reception rate at the MAC layer (bpm) at the i-th node (unicast traffic only)
9. mean value of the routing path length between node i and sink node (number of hops)
10. standard deviation of the routing path length between node i and sink node (number of hops)
11. estimated Packet Reception Ratio, using WMEWMA estimator [3]
12. mean value of temperature on the i-th node (Celcius degrees)
13. standard deviation of temperature on the i-th node.
14. mean value of humidity on the i-th node. (%)
15. standard deviation of humidity on the i-th node.
16. mean value of voltage level on the i-th node. (V)
17. standard deviation of voltage level on the i-th node.
18. Recorded Packet Reception Ratio. 

The above data are features extracted by splitting raw network instances into windows of observation with length equals 500 sequencial entries per node. 
**No other pre-processing (e.g., filtering, z-score normalization) applies on this specific dataset**


Acknowledgements:
This work was supported by the FP7 EU-HYDROBIONETS project (ICT-2011-7,  GA-2011-287613).
We are also grateful to Acciona Agua (http://www.acciona-agua.com/) for providing the premises of La Tordera's desalination plant, as well as to Professor Baltasar Beferull-Lozano, Eugenio Celada, and Ioannis Glaropoulos for assisting with the WSN protocol stacks development and deployment at the industrial plant.

The protocol stack deployment was based on: 
[1] D. Alonso-Roman, E. Celada-Funes, C. Asensio-Marco, B. Beferull-Lozano, Improving reliability and efficiency of communications in wsns under high traffic
demand, in: Wireless Communications and Networking Conference (WCNC), 2013 IEEE, 2013, pp. 268-273.
[2] P. Di Marco, C. Fischione, G. Athanasiou, P.-V. Mekikis, Harmonizing mac and routing in low power and lossy networks, in: Global Communications Conference (GLOBECOM), 2013 IEEE, 2013, pp. 231-236.

Additional References:
[3] A. Woo, T. Tong, D. Culler, Taming the underlying challenges of reliable multihop routing in sensor networks, in: Proceedings of the 1st International Conference
on Embedded Networked Sensor Systems, SenSys ’03, ACM, New York, NY, USA, 2003, pp. 14-27.