Wearable Activity Recognition – Talk by Kristof van Laerhoven

admin Schreibe einen Kommentar

How many sensors do we need? That was one point in the discussion after Kristof’s talk. His approach, in contrast to many others, is to use a large number of sensors for activity recognition. This offers more freedom with regard to placement of sensors, variations of sensors, and also provides redundancy but makes the overall system more complex. His argument is that in the long term (when sensors will be an integral part of garments) the multi-sensor approach is superior – let’s wait some 10 years and then discuss it again 😉

From a scientific perspective and in particular for machine learning (where he sees the greatest challenge) the larger number of distributed sensors is more interesting. I find his porcupine2 (http://porcupine2.sourceforge.net/) stuff quite interesting.

Kristof’s web page: www.mis.informatik.tu-darmstadt.de/People/kristof

Guests @ UIE

admin Schreibe einen Kommentar

We are delighted to have a group of people visiting from Lancaster University in the UK. Prof. Nigel Davies and 2 of his PhD students, Oliver Storz and André Hesse, came last week to Bonn and will stay with us for the next 3 months. It is great that Nigel decided to have his Sabbatical at Fraunhofer IAIS and B-IT, University of Bonn.

Earlier this evening we had already a brain storming session at the Bier garden of the Bahnhöfchen in Beuel (www.bahnhoefchen.de). I am really looking forward to interesting joined projects in the next weeks and months.

GPS statistics – or 25 hours in the car not driving?

admin Schreibe einen Kommentar

My navigation system records simple statistics. I was surprised to see that I have been sitting in my car for about 25 hours in the last 4 months – not driving. Overall it means that 30% of the time I am actually not driving (usually waiting at a traffic light or a railroad crossing – not sure if the 30% are a Bonn-phenomenon).

This makes me wonder if it would be useful to design technologies that provide entertainment or education during these forced waiting times. Could I have used these 25 hours to learn or improve a language? Or could I have watched some funny clips from youtube? Or is listening to the radio all we need while driving (and not driving)?

Paul Holleis joined UIE in Bonn

admin Schreibe einen Kommentar

The Embedded Interaction Research Group (www.hcilab.org) moved in the beginning of April from the University of Munich to the University of Bonn. Paul Holleis (www.paul-holleis.de) , who worked on the project for the last 3 years joined us now in Bonn. It is really great to have him and his experience here!

Matthias Kranz (www.ibr.cs.tu-bs.de/users/kranz/) who also worked on the project is now in Braunschweig working with Michael Beigl. Braunschweig is too close to not work together – this term we run a seminar on context-aware and ambient systems in parallel at TU Braunschweig and University of Bonn.

User tests and final presentations of the lab course

admin Schreibe einen Kommentar



This week the final presentations of the lab course were due. It was really interesting to see what motivated students can achieve in just 4 weeks! The projects explored the idea of contextual ECG and students implemented the data acquisition, transmission over the network, and visualisation. The user studies showed that there is quite some potential in the idea (even though there is also still some way to go before the system is perfect 😉 We plan to publish a paper on the results of the course.

Having a digital presence after life?

admin Schreibe einen Kommentar

An event this week reminded me that life has an end. Getting a link to a Google Map page (satellite image) where someone found his last resting place shows how far reaching new technologies have penetrated our life. This made me think about a demo I saw at Ubicomp last year (http://mastaba.digital-shrine.com/). The digital family shrine did not really relate to my cultural experience and felt somehow strange, but still very interesting and intriguing. I wonder in what form of digital presence after life will become common in Germany…

Panel on Sensor Networks – Applications are the Key

admin Schreibe einen Kommentar

Debora Estrin made an interesting statement. The “early challenges” (the thousands or millions of randomly scattered sensor notes) do not have much applicability outside the battlefield. The new challenges are heterogeneity (specific sensors with specific capabilities) and interactivity (basically sense-making is a process where humans are involved). She made the point that the logical consequence is that dealing with data is the essential issue and statistics have an increasing role. Furthermore these new research directions make a stromg call for application driven research. With these very insightful comments she criticised a lot of the current work in sensor networks. Especially the observation that there is no such thing as a „general sensor“ – it points out that concrete applications are required to make meaningful contributions, even to basic research in sensor networks research.

Talks and Demos at PerCom 2007 in White Plains, NY

admin Schreibe einen Kommentar

This year we (my previous group from LMU Munich) have a significant presence at PerCom. Form the 20 full papers the Embedded Interaction Research Group (www.hcilab.org) has 3, and additionally 1 of the 7 concise papers is from us. With a total of 207 submissions and an acceptance rate of around 10% this is quite an achievement for the team – and a good high point for the project before moving it to University of Bonn.

Gregor(y) Broll had the demo developed in the Perci-Project yesterday. Lucia Terrenghi and I had our talks today. And Gregor Broll, Sebastian Boring, and Raphael Wimmer have their talks tomorrow. Overall the conference has quite a diverse and interesting program, which seem to be more technical and network oriented than Ubicomp or Pervasive. The publications will be online available at the IEEE digital library or on our new publication webpage.

What can you do with a Wii controller? Use tape and connect a toothbrush and program a nice UI (fish tank) in Flash. Quite an interesting demo from Waseda University in Japan – at least the person who did the demo had really clean teeth in the evening.


PerTec Workshop in NY, Ripping off the Antenna

admin Schreibe einen Kommentar

At PerCom 2007 (www.percom.org) Florian Michahelles (Auto-ID labs, ETH Zurich), Frédéric Thiesse (University of St. Gallen), John R. Williams (MIT Cambridge) and I are running the the PerTec workshop (www.autoidlabs.org/events/pertec2007). There is quite some interest in the topic and the range of topics is from technical to user interface and security.

In contrast to the workshop 1 year ago at Pervasive 2006 it seems that item level RFID-tagging is undisputed and that the only discussion point is when it is coming – in 6 month or 10 years. There is also still some discussion about what types of products are the first ones that are tagged – is it pharmaceuticals or cloth?

There was an interesting contribution by Paul Moskowitz from IBM, the clipped tag. It is a tag where you can physically rip off some part the antenna to reduce the read range from several meters to centimetres (see the pictures). It is really interesting that people can do a very clear and visible action to change the characteristics of a tag. The only questions that remains for me – will people trust that this really rips of the antenna? Probably yes…

The topics we discussed included security, privacy, location and RFID, end user issues, and connection of sensors to RFID, we hope to write it up in an overview article.

In the break out groups one discussion centered on the question what would we need to enable end-users to create novel applications using RFID? Further results of the discussion will be available on the workshop webpage soon.

The overall theme that emerged again and again is impact of real world constraints in RFID systems. Questions like: Can you achieve anonymity with a certain protocol? Can not be answered without knowing how it is used in the real world. Especially having recently learned a lot about data-mining (being at Fraunhofer IAIS) for me the questions of exploiting data collected in RFID systems looks really challenging.