I was at the AI/GI/CRV/IS set of conferences last week and took part in a panel discussion on the future of artificial intelligence, along with Jonathan Schaeffer and Kate Larson, and moderated by Kevin Leyton-Brown. We had been charged by Kevin to try and be a bit controversial, and managed get a bit of debate going, especially between Jonathan and myself, although I think we really pretty much agree deep down on the big issues.
One on my more provocative comments was that both the research areas of artificial intelligence and computer vision could be regarded as branches of robotics. This was not meant to be a comment on the relative importance or the areas, but to unify them under a common umbrella. It also does not imply that all AI must deal with embodiment, just as all AI need not deal with cognition. In terms of positioning thee areas within computer science though, I think this is a needed (and correct) grouping consistent with their historical relationship and genesis.
2008
2008
I attended the Apple World-Wide Developer's Conference last week (in additional to visiting people at Stanford, Google and PARC). Most of the actual content that was presented at the conference is covered by non-disclosure agreements, but a few generic observations can be made.
iPhone development is going to be huge. A diverse range of topics was covered, and the iPhone development sessions, in particular, were jam-packed and in several cases mandated the use of overflow rooms (and the main rooms were already very large).
West, and they were pretty full
WWDC was very large and was very well executed and professional. I have been attending various professional conferences for 20 years, and this was probably the most impressive I have seen in terms of organization, preparation and logistics. Of course, the fact that it was held in San Francisco didn't hurt either!
Since the iPhone development environment is very similar to the one for regular OS X, one can expect a lot of crossover from developers who originally have plans for only one of the two platforms. The UIKit for the iPhone isn't quite the same as the OS X environment, but additional the development effort required to port an iPhone application to the Mac might be a small as a day or two. This is also an important benefit for Apple.
rehearsed, albeit occasionally a bit dry
2008
We applied for and received funding from Microsoft to develop tools and infrastructure to use Microsoft Robotics Studio (MSRS) with our underwater and walking robot systems. Our work has involved the control of our robot both underwater and on land using visual markers, especially because any other type of communication underwater is so difficult. This funding is part of Microsoft's Human Robot Interaction Awards and was awarded to Ioannis Rekleitis and myself.
To quote our own project information, there are many challenges to be faced here due to the fact that our Aqua vehicle moves in a variety of terrains and is capable of providing very limited sensory feedback in the form of video footage and the state of an inertial measurement unit (IMU). We want to use one video and IMU feedback for control.
Current operations with the robot require a skilled operator who is capable of guiding the robot either in walking or in swimming mode. In our project we proposed to implement a user interface utilizing the strengths of the Microsoft Robotics Studio (MSRS) to provide an interface for controlling the robot as well as a visualization tool for interpreting the visual feedback. This work would also extend a new method for communicating with AQUA when a direct link to a controlling console is not available; this method called RoboChat is based on cue cards that are presented in AQUA’s vision sensor, instructing the vehicle to perform high level actions. While on land operations communication between an operator and a vehicle is easy to implement in a variety of methods, e.g. wireless/wired links; underwater communications are a lot more restrictive in terms of cost, bulk, energy, and bandwidth.
Some of our work with RobotChat was recently presented at the 2008 International Conference on Robotics and Automation. The Microsoft announcement is here
2008
I am going to the Robotics Science and Systems Conference and want to load my Garmin GPS with a map of Europe. The open source gpsbabel program seems suitable, but the macports install (port install gpsbabel) fails with an error as follows:
configure.in:39: error: AC_SUBST: `' is not a valid shell variable name
The solution for me was to visit the source directory on my machine:
cd /opt/local/var/macports/build/_opt_local_var_macports_sources_rsync.macports.org_release_ports_textproc_gpsbabel/work/gpsbabel-1.3.5
and edit the file config.in
to delete the lines mentioning the (obsolete) AC_EXEEXT. Then it installed fine.
Once a suitable map has been created, the Garmin 60CSX, like other Garmin USB devices, can be directly mounted on the desktop as a mass storage device. This requires you to go into the System setup menu, in the interface section, and click on "Mass Storage Device".
It's also possible to export a set of points of interest from Google Earth as a .kml file, and then use gpsbabel to convert this to a .gpx file (which can presumably be uploaded directly using gpsbabel).
2008
Today I visited Wolfram Burgard's lab in Freiburg Germany. There are numerous interesting projects going on there, including work on range estimation from video and range data, and modeling of robot kinematics using Gaussian processes (the latter was the subject of a recent paper at RSS). They also have several interesting implementation projects running, including an automated blimp, and automated helicopter, and several types of vehicles.
I also gave a talk there whose primary focus was our own underwater and amphibious robotics work at McGill. Unfortunately, the display system would not recognize my older Powerbook, and as a result the 1-hour talk started a half -hour late. This was a new experience for me. This got me totally stressed out and led to a very uncomfortable talk for me. That said, the visit was very pleasant, interesting and enriching.
