Quick Update

Just a quick update on what I've been working on over the last few days. I've been setting up my "work environment" on my laptop. Have Linux installed and finally have the wireless working properly. Hooray! If anyone else is doing this i suggest fedora 7 because it has a lot of the wireless card drivers included. Its a bit of a pain installing these manually. I've set up eclipse with subclipse so I can access my svn repository on googlecode directly through eclipse. Having trouble installing a python add on for eclipse so might have a look other editors, any suggestions anyone? Have been learning about the technologies i'll be using. Worked through a Python tutorial. Python is dead handy. I have some very very basic experience with php and sql so im currently going over these again to understand them better.

Design

Ok my drawings are not amazing but hopefully they will explain my idea a little. Im trying to make the whole thing less restrictive without effecting tag reading too much. There will be no glove and the reader will be positioned just above the wrist at an angle from the hand, so that the hand can open and close normally. All thoughts and comments welcome.

System sketch

Heres a rough sketch of my system architecture. Sorry its a little hard to see.

"Reminding Short-Term Memory Sufferers to Complete Routine Tasks"

This is Caities final report for the RFID ODCSSS project. Heres a quick recap on what the system does. Small RFID tags are attached to a number of objects, which can be read by the RFID glove. The glove consists of an RFID reader sewn into the palm. This connects to through a simple circuit to the Gumstix computer. This circuit also supplies power to the reader and Gumstix. The Gumstix reads the data picked up by the reader through one of its serial ports. The Gumstix is connected to an expansion board with wireless capabilities which is used to connect to a remote server on a pc and pass the data to a webpage. The webpage updates a database with the tag information.

The system is designed to recognise individual tasks. Activities are associated with a number of tags. A task is identified based on the percentage of associated tags that have been read, within a given time period. Each task also has a specific "end" object or the last object that will be touched in a given task, which indicates when the task has been completed. Overall a nice simple and effective method.

For testing a number of users were asked to interact with objects involved with spefic tasks, and later asked for feedback on using the glove. Overall many found the glove restrictive mainly because of the RFID reader in the palm, and the glove being uncomfortable. My first thought was to move the reader to the wrist but this was already tested and was not very effective due to the short range of the reader. However I do have another idea that might work, which would also make most of the glove unnecessary. Will post a sketch of this soon... Oooh.

"An Introduction to RFID Technology"

Lorcan sent me this very interesting paper. It explains in detail how the RFID technology actually works, and a number of potential applications for it. RFID can be divided into the two types active and passive. The system I am using is passive where by the RFID tags do not need their own power source but are activated by the reader. There are again two types, magnetic induction and and electromagnetic wave capture. The main difference is the range of the reader. I am using near field RFID based on magnetic induction.



Basically the reader contains a coil and passes an alternating electrical current through which creates a localised alternating magnetic field. The tag contains a smaller coil. The readers magnetic field creates an alternating voltage across the tags coil when it enters this field. A capacitor causes a charge to build up, thus powering the tag. The tag can then generate its own magnetic field which opposes that of the reader. The reader can interpret this as an increasing current. The tag can vary its load over its coil which results in variations in its magnetic field corresponding to its binary ID. The reader interprets these signals through the change in current through its own coil.



The paper describes other interesting techniques but they are not applicable to my project.

"Ubiquitous Memories: a memory externalization system using physical objects"

This is a paper I read about a system called "Ubiquitous Memories" which proposes to augment human memory trough interaction with objects. This sounded very interesting at first but parts of the abstract and introduction were a little confusing. It is proposed that the technology will help perform various daily tasks and gives examples; "reducing wasted time, facilitating human to human communication and recycling our experiences". I didnt seem particularly clear to me what their system would do. What exactly is a "recycled experience"?

In a later section the design concept was discussed. The concept is to naturally augment the human memory. The system design was was broken down into the following proceedures. "1: A user perceives an event via his/her body. 2: The perceived event is stored into his/her brain as a memory. 3:The human body is used as media, i.e, memory travels all over the body like electricity, and memory runs out of the hands." At this point the paper completely lost me, and goes on to say the memory is transferred to an object and can be later recalled through touch. Im not sure if these points had some leaning to the actual biology behind memory or they were trying to describe the concept in a very abstract way.

The whole thing makes more sense when the hardware is described. There is a head mounted display over the left eye which can display video to the user. This eye piece also has a camera to record users activities. There is an RFID reader on one wrist. These are both connected to a remote control for the system which connects to a hip mounted wearable computer connected to wireless LAN. Its all very "we are borg". I find the location of the RFIS reader interesting. I wonder how effective this position is at reading tags, it does'nt seem to be mentioned in the paper, and wether it is worth investigating. Basically the system works by recording a users activities. Objects are tagged with RFID tags and when read the user can see a video or scene of an event related to the object. The given example is a trophy will trigger memory of a sporting event.

I thought the testing of the system was done quite well. A number of people were asked to remember specific things and to later recall them. Their memory was helped in different ways one of which being the "Ubiquitous Memories" system, and the best human memory aids could be determined.

Overall the system seemed very innovative with plenty of potential, however much of the paper was extremely abstract and unclear.

Project meeting 2

I met with Lorcan and Graeme yesterday for a project meeting. We discussed what I had done so far and what my next milestones should be. This week im going to work on my research document and hopefully get it finished for Friday. This weekend i'll take a look at some more Python tutorials. I've never programmed with this before but it seems fine. Caitie found it worked very well with the Gumstix. I also need to research more on connecting the Gumstix to wireless as im still a little unclear with some aspects of this.

RFID ODCSSS Project

Caitie Lustig took this project for ODCSSS during the summer. I could'nt find the final project report in the CVS repository, but Caitie sent me a rar file with some useful notes and what seems to be a near completed draft of the final report. The paper is titled "Reminding Short-Term Memory Sufferers to Complete Routine Tasks" The main focus of the project was to use RFID technology to asist sufferers of short term memory in completing everyday tasks. The system was designed to determine which activities were being carried out based on what objects were touched. The system can then alert users of forgotten tasks and when an activity has not been completed. I was interested in how this was achieved. Each activity had specific objects associated with it, and a final step ie, the last object that will be touched in that activity. A task is identified by the percentage of associated objects that have been touched. Time limits were also given for each task to be completed. This ensures to notify users in case they had forgotten what they were doing. I like this system. Its simple but effective. Im not sure if multi task recognition was included but I think the system would be easily adapted. The glove was tested with users carrying out a number of specific tasks. The results of these were not in the notes I have, so I will ask Lorcan and Caitie about them.

RSS problem

Heres what I found out about the problem with the RSS feeds: "News feeds (Atom, RSS) don't work with private blogs, because there's no way to provide an account for the readers to login with. It's a feature of the design. http://bloggerstatusforreal.blogspot.com/2006/12/blogger-blogs-and-permissions.html " I guess we will just have to keep the blogs public??

Pics

Heres two pictures of the Gumstix in its jazzy new box connected to my pc with Hyper Terminal on windows.

Heres the circuit I was working on. It just needs to be cut to make it as small as possible so it will fit inside the Gumstix box.

Will keep posting pictures as I go.

What I've been up to

Im a little behind with my posts at the moment, its been a busy week. I met with Lorcan on Thursday and I started work on the circuit which will connect the rfid reader to the gumstix. Turns out soldering is not so straightforward plus I ended up soldering the regulator in the wrong way. Lucky it came back out easy enough and the circuit is now working correctly. The components were flattened as much as possible so it will fit inside the Gumstix box.

Todays Work

I met with Lorcan today and he went through the hardware I'll be using. This was definately a big help and I've gotten my head around how it all connects together and how it works. Also met with Lucy Dunne today, an expert in wearable computing, whoi will a great help with the circuitry involved. Will be meeting with Lorcan again next week to do a bit of soldering. Will post more detail on this later...!

I took home a Gumstix today with a serial lead and power plug, so can connect it to my pc and have some fun learning about it. Its basically a fully functioning mini computer running Linux. Very cool. This one is connected to a wireless expansion board. Also took an RFID card reader and some RFID tags, but wont be working with these just yet.

Had another look into thr RSS problem but could'nt find anything related to it. Submitted a query / bug report so hopefully we will get some more on this soon.

RFID Wiki

Spent some time going through the Gumstix RFID wiki today to familiarise myself a bit more with the harware Im going to use and how it all fits together. A Gumstix is basically a functioning mini computer loaded with a Linux based operating system. A flash memory card serves as a hard disk. A basic breadboard circuit seems to be used to connect various components, and a serial connection is made between the RFID reader and the gumstix. I understand how each of the components work but am still a little uncertain as to how it all fits together. I'm meeting with Lorcan tomorrow and he is going to explain the hardware to me which should clear everything up.

References

Will use this post to list any reports I write about. Will probably be handy for later referencing:

A. Schmidt and C. M. Hans-W. Gellerson, "Enabling implicit human computer interaction- a wearable rfid-tag reader", Proc. 4th Int'l Symp. Wearable Computers (ISWC2000), pp. 193-194

Matthai Philipose, Kenneth P. Fishkin, Mike Perkowitz, Donald J. Patterson, Dieter Fox, Henry Kautz, and Dirk Hahnel. "Inferring Activities from Interactions with Objects", (2004)

?? Suppakri Chatchayanuson, Charles Christopher Oneyama, Nachiket Shelgikar, Saravana Sivasankaran "Ketchen Tracker" ??

Some more Research Papers

Another report I looked at was a CMU student project. Again the basic hardware is the same, however data is transferred to a phone or pda and synchronised in real time. The main concept is a system which keeps track of kitchen items and saves time with grocery shopping. The most important points i got from this report is the emphasis it puts on potential problems, such as tags positioned too close together, and error handling. I had'nt yet thought much about this and its something I will have to bear in mind later on in the project. It also mentions that code may not run correctly on the gumstix even if its ok on a Linux machine. Fingers crossed this is not very common... :-(

Some more Research Papers

Heres a few more notes on the papers I read yesterday. The first one is titled "Inferring Activities from Interactions with Objects" its the Intel research Seattle project. It begins with discussing the difficulty of using computers and sensors for interpreting and recognising the activities of individuals as they indertake everyday tasks, and more specifically in the care of elderly people. The main problem seems to be getting the recognition software to adapt to "variety", for example the same task could be completed in different ways. Their solution to this is to represents tasks in a sequence or probable sequence of the objects used.

The basic system is the same as my own project where objects are RFID tagged and read from a mobile RFID reader, in this case an antenna embedded in the palm of a glove. The software implementation is based on probabilistically estimating activities from the sensor data, taking into account different variables such as time. For example in the given scenario, in the frist step of making a cup of tea there will be a high probability of encountering a kettle. This seems like a logical approach to the problem. The system was tested with people undertakning a number of everyday tasks. My own project will need to be tested in a similar way. The results varied depending on the task but were promising overall. One thing thats important to note from the results is that water and metal are bad! Water absorbs the radio waves and metal could short circuit the antenna.

Research Papers

Started some of my reading and research today. The first paper was titled: "Enabling Implicit Human Computer Interaction, A wearable RFID-Tag Reader". The paper deals with the concept of implicit human computer interaction, or in other words using a persons actions as inputs that can be interpretted by a computer. The basic concept of this case study is identical to my own project, where a mobile RFID reader is integrated into clothing, or something carried by the user, which records interactions with various stationary RFID- tags.

The paper gives a brief explanation on how the system works. The RFID-tags are small electrical circuits which become active through electromagnetic induction, or are energised by the magnetic field produced by the reader. While it supplies energy it is able to communicate with the tag, sending a unique identification in a continuous string to an attached computer, which can then in turn be mapped to URLs. This system can be applied to a number of scenarios such as stock control.

The main outcome of this is to develop web based applications which use the RFID data in some useful way. This is again applicable to my project but I think I would like to explore other ways of communicating with the user such as the built in display that Lorcan suggested.

Which version control system?

Finished looking over some version control systems today. Have used GForge CDE and GoogleCode before for previous projects plus some testing with others such as Mercurial. Spent a lot of time today with TRAC, a web basd managment tool. It was mentioned at the meeting so thought I'd check it out. Turned out to be a bit of a pain setting up on windows, I miss my mac, but it was fine in the end. After playing with it for a while it seems to be a nice tool. Although I still like google code which we used last year in our software development project, so decided to go with this. Set up the project quickly, will need to go back and do my description properly, but at least thats one thing out of the way...

First project meeting

I had my first project meeting with Christine today and met with Lorcan, Steve, and Graeme. They talked to us about our first milestones such as setting up a blog, picking a version control system, and our initial reading and research. They told us about a few useful points to bear in mind during the project, such as collaborating with each other and setting up a sort of support group between the six of us through our blogs. Collaboration between different projects can look good when it comes to grades :-) The blog will be used for keeping track of our meetings, our thoughts and ideas, and generally everything we do for the project. A lot of this will be very handy when it comes to writing our reports. One thing that we will not be forgiven for is not backing up properly. Something I learned well when we lost most of our first year project. It is also worth trying to get as much as possible done before the second semester which sounds like its going to be a bit of a nightmare...

Lorcan went through my project with me and talked a little bit more about it. This was good and definately gave me some ideas with where I can go with the project. He will be giving me the hardware I need with in the next weeks or so. I can also email Caitlin who took this project during ODCSSS which is cool. We'll be having our next meeting probably on Monday. For the moment im gonna go pick a version control system. Ooooh