Pervasive 2009 had a really exciting program and provided a good overview of current research in pervasive and ubiquitous computing. Have a look at the proceedings of the pervasive 2009 conference. The Noh theater in Nara was a very special and enjoyable venue and it was organized perfectly - as one would expect when travelling to Japan.
The idea of having short and long papers together in the main track worked very well in my view. The number of demos and posters was much higher than in the years before - and that was great and very inspiring. Have a look at the photos for some of the posters and demos.
The program consisted of 20 full papers (18 pages) and 7 notes (8 pages) which were selected in a peer review process out of 147 submissions (113 full papers, 34 notes) which is a acceptance rate of 18%.
John Krumm presented his paper Realistic driving tips for location privacy - again having a good idea making the presentation interesting beyond its content (having review snippets in the footer of the slides - including a fake review). The paper explores the difficulties that arise when creating fake GPS tracks. He motivated that the probabilities need to be taken into account (e.g. you are usually on a road). I liked the approach and the paper is worthwhile to read. I think it could be interesting to compare the approach is not create the tracks but just share them between users (e.g. other people can use parts of my track as fake track and in return I get some tracks that I can use as fake tracks). http://dx.doi.org/10.1007/978-3-642-01516-8_4
If you phone knows where you are you can use this information to control your heating system. This was the basic idea of the research presented by Stephen Intille. They explored using GPS location of the users to automate control of the heating / air condition control in a house. It seems there is quite some potential for saving energy with technology typically used in the US (one temperature control for the whole house). In Europe where heating systems typically offer finer control (e.g. room level) the potential is probably larger.
James Scott presented a paper that showed how you can use force gestures to interact with a device. In contrast to previous research (e.g. GUMMI) the approach works with a ridged device and could be used with current screen technologies.
What do you need to figure out who is holding and using the remote control? This question is addressed in the paper "Inferring Identity Using Accelerometers in Television Remote Controls" that was presented by Jeff Hightower. They looked at how well button press sequences and accelerometer data give you information about which person is using the device.
Geo-fencing: confining Wi-Fi Coverage to Physical Boundaries is an example of how to create technological solutions to fit a user's conceptual model of the world. As people have experience with the physical world and they have mechanisms to negotiate and use space and hence linking technologies that have typically other characteristics (e.g. wireless radio coverage) to the known concept is really interesting.
Situvis, a tool for visualizing sensor data, was presented by Adrian Clear from Aaron's group in Dublin. The software, papers and a video is available at: http://situvis.com/. The basic idea is to have a parallel coordinate visualization of the different sensor information and to provide interaction mechanisms with the data.
Nathan Eagle presented the paper "Methodologies for continuous cellular tower data analysis". He talked about the opportunities that arise when we have massive amounts of information from users - e.g. tracks from 200 million mobile phone user. It really is interesting that based on such methods we may get completely new insights into human behavior and social processes.
If you have seen a further interesting paper in the conference (and there are surely some) that I have missed feel free to give a link to them in the comments to this post.