Monthly Archives: October 2008

Week 9: more reading and reference collecting

One of the types of sensors I’m looking at are mobile GPS sensors that attempt to determine what mode of transportation one is using, and from that and the GPS trajectory estimate one’s environmental impact (carbon footprint in particular). I’ve found 3 systems doing this so far:

  • Carbon Diem (previously known as Carbon Hero). Carbon Diem is designed to run on GPS-enabled mobile phones (Blackberry and Nokia N-series now, but “platform and provider agnostic”). As this article indicates they have been working on the system since 2006. They are trying to raise money, and focusing on the corporate market initially. The app claims to “tell if you drive, fly, take the train or walk”, and if they can sign a deal with a carrier or handset maker they could potentially launch to consumers in Spring 2009. According to this Guardian article, “the software was almost 100% accurate in working out when people were on airplanes or trains; it was between 65-75% accurate at guessing when people travelled on buses”.
  • Ecorio is a similar application for Google’s Android platform. It also uses GPS to detect the mode of transportation, and estimates carbon output from that. There is apparently support for detecting how efficiently you are driving, which is an interesting twist (though not sure how you provide the feedback when the user is driving). Ecorio also provides suggestions for reductions, such as links to Google Transit, and carpooling info. There appears to be some “what if” functionality built in as well, such as how much carbon will I emit if I start taking public transit half the time (screenshot in this article). Users can also purchase carbon offsets from the phone. There are plans to port the application to other platforms (iPhone is mentioned, but would be difficult given the restrictions on background processing). I wonder if this can be run in the Android SDK simulator, with Bluetooth GPS as input?
  • UCLA’s Personal Environmental Impact Report is another phone-based system. Currently in closed beta, runs on Nokia GPS-enabled phones. They are including other environmental factors, like smog.

It would be interesting to get a hold of some of these sensors and see how accurate they are compared to just recording the odometer & fuel usage information, or marking commutes on a map from memory.

I’m really liking BibDesk, the Mac OS X BibTeX manager. One really nice feature is being able to link URLs or local files to references. So you can save a PDF of a paper to your hard drive and link the BibTeX record to the file for easy access. It can also show you a nearly instant preview of how the currently selected item will be rendered in LaTeX. There are a bunch more features I haven’t dipped into yet, such as auto generation of cite-keys and automatic managment of your PDFs.

I’m still not quite sure how to cite web sites. For now I’m using the misc BibTeX type, and putting a \url{http://...} in the Howpublished field, which seems to produce reasonable output.

Planned items from last week:

  • Read 4 papers from literature review list
    • Understanding mobility based on GPS data [DONE]
    • Learning Transportation Mode from Raw GPS Data for Geographic Applications on the Web [NOT DONE]
    • Displaying dynamic carbon footprints of products on mobile phones [DONE]
    • Energy efficiency: rebounding to a sound analytical perspective [NOT DONE]
  • Start building table of possible sensor inputs [STARTED]
  • Add publications from PET workshop paper to BibTeX [DONE]

Other accomplishments this week:

  • Added a bunch of papers to lit review on Khazzoom-Brookes Postulate
  • Added several more papers while tracking down references from PET paper

Hours worked: 15 (target: 15 hr)

Plans for coming week:

  • Read 4 new papers from literature review list
    • Learning Transportation Mode from Raw GPS Data for Geographic Applications on the Web
    • Energy efficiency: rebounding to a sound analytical perspective
    • New Ways to Promote Proenvironmental Behavior: Expanding and Evaluating Motives for Environmentally Responsible Behavior
    • A Definition of ‘Carbon Footprint’
  • Add more sensors to list
  • Write up notes from papers read during PET paper lit review
    • EcoIsland: A System For Persuading Users To Reduce CO2 Emissions
    • Taking the Guesswork out of Environmentally Sustainable Lifestyles

Pointers to work products:

Cool links:

  • Nothing I can think of
Advertisements

Week 8: BibTeX & BibDesk for annotated bibliographies

I made substantially less progress in the past week than expected for two reasons. First, a crisis came up in LILT that required me to urgently work on SocialSense. Second, as part of the Sustainable Saunders Energy Group, I was notifying the occupants of the sixth floor of the Saunders building about upcoming nighttime air conditioning shutdown that should save at least $100K annually. Hopefully both situations have subsided for now and I can focus on the sustainability research this week.

I was looking for a way to get my literature review notes into my proposal document (per Philip’s request) in a sane and maintainable way. As I read each paper, I make notes (usually just bullet points) and once I have reached a stopping point (like the end of a semester) I assemble them into something conceptually coherent. I decided the best way to do this is to enter all the papers I read into my proposal BibTeX database, and place my notes in the annote field. BibDesk is a sweet Mac OS X GUI for maintaining BibTeX databases, and has easy access to the annote field. Then the key was getting the notes into the document. There are BibTeX styles (.bst files) that provide different bibliography styles, and some will print the annote fields of each referenced item, producing an annotated bibliography. I ended up using the annotation style, which ships as a standard part of TeX Live 2008. There are other annotated bibliography styles that could be installed, but I decided to go with the built-in one for now. Obviously I will switch back to the normal bibliography style when the literature review section is done.

Planned items from last week:

  • Read 2 papers from literature review list
    • Understanding mobility based on GPS data
      • Not done
    • Learning Transportation Mode from Raw GPS Data for Geographic Applications on the Web
      • Not done
  • Start building table of possible sensor inputs
    • Not done
  • Add publications from PET workshop paper to BibTeX
    • Not done

Other accomplishments this week:

  • Fixed formatting on my research portfolio per Philip’s suggestions
  • Moved literature review notes into proposal document per Philip’s request via annotated BibTeX entries

Hours worked: 6 (target: 15 hr)

Plans for coming week:

  • Read 4 papers from literature review list
    • Understanding mobility based on GPS data
    • Learning Transportation Mode from Raw GPS Data for Geographic Applications on the Web
  • Start building table of possible sensor inputs
  • Add publications from PET workshop paper to BibTeX

Pointers to work products:

Cool links:

  • Caffeine, a Mac OS X application that allows disabling sleep, screen dimming, and the screen saver via a menu bar item. This is really handy for presentations, so you don’t have to keep moving the mouse to keep your system awake. There’s even a timer functionality so it only stays “awake” for a fixed amount of time, so you don’t have to worry about forgetting to turn it off. Freeware.

Week 7: LaTeX style and outline

I wasn’t as productive this week as I had hoped to be. I was feeling under the weather on Thursday and Friday, which cut into my research time. I didn’t make any progress on reading more papers, and I spent more time than I had expected integrating the latest (if you can call something from 2000 latest) LaTeX thesis style changes from the bowels of bertha into the public Google Code repository. The good news is that I don’t expect to have to spend much more time on the thesis style, though it would be good for someone to go through it once in comparison with the latest style guide from grad division. Mark Stillwell has indicated a willingness to do that update.

I’m hoping to have at least 20 new items for my literature review by the end of the semester, so given the number of weeks left, I should be reading papers at a steady rate of at least 2 per week. That’s going to be a recurring item on my weekly todo list.

The other area I need to work on is collecting all the various potential sensors into one list so I can start concentrating on which ones I will be using for my research.

Planned items from last week:

  • Read 4 papers from literature review list
    • No progress
  • Look at websites with related functionality from list
    • No progress
  • Add publications from PET workshop paper to BibTeX
    • No progress
  • Write up outline of final tech report
    • My final tech report is my proposal, so filled in the chapters with an outline
  • Fill in more parts of research portfolio

Other accomplishments this week:

Hours worked: 7 (target: 15 hr)

Plans for coming week:

  • Read 2 papers from literature review list
    • Understanding mobility based on GPS data
    • Learning Transportation Mode from Raw GPS Data for Geographic Applications on the Web
  • Start building table of possible sensor inputs
  • Add publications from PET workshop paper to BibTeX

Pointers to work products:

Cool links:

  • None

ICS 699 Fall 2008 week 6: TeX and literature review

This week was spent finishing off loose ends and preparing for my literature review. I finished all the blogging related to Ubicomp, and made the UH thesis LaTeX style publicly available in a more accessible format. I started the literature review (read and commented on 1 measly paper so far), but now have a central location for my literature review notes.

At this point I feel I really have to read as many papers in the area as I can to get a better idea what has already been done, and what is likely to be done in the near future. Naturally, both those things are areas I would like to avoid.

Past weeks accomplishments:

  • Finished blogging last 2 days of conference
  • Wrote up items relevant for PET from conference
  • Wrote up lessons learned for future conference attendance
  • Installed TeX
  • Put UHM LaTeX thesis style up on Google Code for broader adoption, and Google Group for discussion
  • Created CSDL tech report for proposal draft

Hours worked: 15 (target: 15 hr)

Plans for coming week:

  • Read 4 papers from literature review list
  • Look at websites with related functionality from list
  • Add publications from PET workshop paper to BibTeX
  • Write up outline of final tech report
  • Fill in more parts of research portfolio

Pointers to work products:

Cool links:

  • Instapaper is a bookmarking site with a twist. It’s designed to collect URLs for things you want to read later. You use the provided bookmarklet to add things to your reading list, and then as you read things from the list they are recorded as read. That’s nice, but the real power comes from the associated iPhone application that can download web pages from your list for offline reading. It also has a text mode that does a good job of rendering web pages as rich text, which displays faster and with less clutter.

Lessons learned from Ubicomp

These are meta-issues I learned from attending the conference, to remember for the next conference I attend.

  • Don’t leave the night the conference ends if possible. I didn’t know the daily schedule for Ubicomp until long after I had made my travel reservations. Turns out the conference schedule went past 6 PM, so I had to leave before the end of the conference to catch my flight. In future I’ll just plan to stay that extra night so I can be sure to attend the conference until the end, and not have to deal with my luggage the last day of the conference.
  • I think it’s time to update the canonical CSDL presentation template. I think the style and font choice seem somewhat dated now. While I don’t believe in presentations with fancy backgrounds and other fripperies, I think it could be updated. In a conference environment, it was quite helpful to see information in the header & footer of presentations: slide N of M, presenter name & organization.
  • If I had more time in South Korea, I would have visited the DMZ. However, while giving blood last week I learned that there are three different strains of malaria in the DMZ area, and if I had visited I would not have been allowed to give blood for 2 years!

Action items from Ubicomp

This is a list of specific things to follow up on from Ubicomp:

Ubicomp 2008 blogging, day 3: location aware applications

The most interesting part of day 3 for me was the session on location-aware applications. The first paper was The potential for location-aware power management from Robert Harle and Andy Hopper of the University of Cambridge. They analyze a set of highly-accurate location data from a commercial office to see what energy savings could be enabled by knowing where the user is located in the building. They found the potential for savings for devices that can be switched on and off quickly (like lights or waking a computer from sleep), but the problem is much harder for slower systems like HVAC. For the slow systems, only outside office hours and lunchtime could be predicted with the accuracy required to have rooms ready when the user returned.

The other big paper for me was Understanding mobility based on GPS data, which covered using GPS data to determine what transportation method was being used (sound familiar? 🙂 ) They found that just using velocity data from GPS doesn’t work well (< 50% accuracy). They follow the standard machine learning process: segmentation of the data, extracting features that are representative of those segments, and feeding that into a generative model. They had 70 people using their system for a year, which is a huge amount of GPS data. This is definitely a paper I need to fully understand.

I had to leave to catch my flight before the closing panel session finished, but what I heard was interesting. There was spirited debate on where ubiquitous computing was going, where the conference was going, complaints about the many related conferences (Pervasive, MobileCHI). Too bad I had to leave early.

Ubicomp 2009 is in Orlando FL. Not sure at this point if I’ll be going or not.