Minutes from "Senzations '07" on 10.09.2007 - 00:00

Summer School on Applications of WSN
September 10-14, 2007, Warsaw, Poland

In the beginnig of september three members of the GKMM (Arthur Herzog, Daniel Jacobi and Pablo Guerrero) and also Khalid Nawaz from DVS group participated in this years summer school on Applications of WSN (Senzations'07). This was the second session of this kind and it was located this year in Poland at the Warsaw University of Technology. It was organised by Srdjan Krco from Ericsson Ireland and Konrad Wrona from SAP France. Senzations'07 was co-located with IEEE Eurocon 2007 and included a shared session on WSN.

Some fundamental talks on WSN protocols, operating systems and security were given. Konrad Wrona gave some advanced insights to security in WSNs. From Philips one of their lead developers in the fieds of batteries and body sensors, Martin Ouwerkerk, presented ongoing research. One of the organizers Srdjan Krco presented Ericssons ideas for WSN in mobile networks and healthcare. His collegue Mattias Johansson envisioned ubiquituous services for home, car and personal environment. Different talks were related to the EU project WASP, which focuses on three main areas: road transport, elderly care, and herd control. Here elderly care, and herd control were addressed. Some participating students also had the opportunity to present their ongoing research. Next we comment on the highlights of the summer school.

Concerning security Konrad pointed out that most WSN projects don't have any security mechanisms built in at the moment. He discussed some common attacks on network security and took a closer look on some of them related to WSN's. When cryptography is used in projects then it is either based on symmetric cryptogrphy or on elyptic curve cryptography (ECC). The symmetric cryptography is quite fast and the ECC has the advantage that to achieve the same degree of security the keys are much smaller than keys used in common RSA. But one of the main problems is the distribution of the keys over the wireless channel and management of keys after deployment. The usage of one global key is not appropriate here, since by compromising a single node whole networks security is compromised. There are different reseach directions like centralized approach, pair-wise keys, pre-distribution. For all approaches an important factor is the overhead cost in computation and communication to gain security.

The talk of Philips research started with a short introducion to the area of interest of Martin Ouwerkerk and some details about batteries in sensors and electronic. He presented the sensor node platform SAND and later on their current research on detecting emotions. For the development of SAND they defined three phases of sensor node development. For the first phase the sensors needed to have the shape of a cylinder, in the second phase the sensor had to be like a button 0.5cm thick and about 2cm in diameter and for the third phase the sensor should have the thickness of only 0.5 mm. These shapes are meant to hold the sensors, microcontroller, radio chip and battery. At the moment they finished to produce phase two prototypes. For the 3rd phase the main problem is the battery, but the recent developement in the area of battery research, seems to make it possible. For SAND Philips used one of their own DSP's that originaly was ment for head phones and so had a very small power consumption with high computing power. As sensors there is a set of more than 8 boards with different sensors available and the power supply is a standard Li-ion button cell.
Philips' current research is looking into emotions of people. Therefore the SAND sensor node was equipped with different sensors like ECG, heartbeat detection, skin electrical resistance and some more. The combination of those measurements make it possible to distinguish the emotions of a person. The emotions are classified in a two-dimensinoal matrix with activity level and positive/negative weighting on their axis.

The two researchers from Ericsson Ireland gave a broad overview of their visions in the areass of connected WSN's and ubiquitous infrastructures.
To develop an infrastructure to connect sensor networks some aspects in integrating these networks were stated, like heterogeneity of devices, or privacy, security and trust. With the new incentives for moblie phone bussiness there are also new communication traffic patterns that arise from the difference of continuose sensor data and sequential phone call data. The new mobile network infrastructure has to support this and can also support sensor data from mobile phones or peer-to-peer mobile sensor networks.
The talk about ubiquitous infrastructures was far more looking into the future and so has not very much detailed information. The vision was, that via the mobile phone network many sensor networks are connected, e.g. the sensor network at home, in the car, in public places or even at your own body. These can then be accessed by everyone to get in touch to their home networks. One of the major problems of this approach is how to trust people using your networks.

We also heard two talks that were related to the WASP project. As stated above it is based on three different scenarios and two of them were covered in the summer school: elderly care and herd control. The talk about herd control from Kees Lokhorst, Animal Sciences Group, Wageningen, UR, was rather detailed and gave a good insight to the problems that were encountered in a real world scenario. The task of the herd control scenario is to monitor the health, position, orientation, etc of the animals like cows, pigs and even hens. The gathered information then can be used to decide about the quality of the corresponding food produced depending on the health and activity of the individual animal. The same information can also be used to early detect animal diseases and avoid their spreading. The food inspectors can use this information to verify the feeding and transportation regulations. The main drawback why this technology is not used in practice yet is the relative high price of the sensors per animal.

On the last day many participants gave short and interesting talks about their work and at the end the best talks got rewarded. As a result Pablo Guerrero won the 2nd place.