Literature Review on Current Research


Pick a topic appropriate to the class, such as "wavelet usage in compression", "controlling an assembly line with a Z80 CPU", "using pass transistors in an ALU", etc. Find at least 10 references on it.

In this assignment, you will read and summarize current research papers. The papers should come from established conferences and journals, such as the IEEE Workshop on Signal Processing Systems (SiPS) and the IEEE Transactions on Computers. By "current", I mean that the majority of the papers should have been published in the last 4 years. If there is an established reference (i.e. all the papers you read also refer to it) then you should use it, too. Examples include Ingrid Daubechies' 1990 paper "The wavelet transform, time-frequency localization and signal analysis" from the IEEE Transactions on Information Theory, and Stephane Mallat's 1989 paper, "A Theory for Multiresolution Signal Decomposition: The Wavelet Representation," from IEEE Pattern Analysis and Machine Intelligence. You can use books too, but books will not count towards the 10 minimum.

Web pages do not count, and should not be used! One exception to this is if you include a graph, picture, or other figure that you found on-line that supports what you say. In this case, you must include where the figure came from. If you find a figure in electronic format that is included in one of the papers, cite the paper as the source. The reason to cite the paper is that web-pages come and go, but papers are archived. That is, in 10 years, there is a good chance that the on-line version will not be in the same location, but a good library will have the published paper. Another way to think of this is that you want to document where someone else can find the source (i.e., author, title, journal name, volume, number, and page), not how you specifically got it (e.g., I saw my roommate reading it and he gave me a copy).

Your goal is to write a literature survey about what has been done already, and what open problems remain in the topic. Discuss the different papers in relation to each other, in your own words. Do they agree with each other? Do they disagree? Does one disprove another? Does one improve upon another? Do they make conflicting claims? Which one is correct? Do not omit a reference because it disagrees with the others, however you can dismiss one if you have a good reason, such as a flaw in methodology. You must clearly state what the open problems are. Your introduction should give a brief overview of what has been done, then the following paragraphs should go into more detail. Be sure to use citations, and as many as are appropriate (if 3 papers talk about the same thing, cite all 3).

The text you write should flow, meaning that you will have to link different sources together in your explanation. That is, do not write: "Topic A is important because is explains wavelets. The FFT has many uses as paper B mentions." (Are you talking about wavelets, or the FFT?) Here is a better example of a proper transition: "Topic A is important because is explains wavelets. Both the wavelet transform and the FFT are used for signal analysis. Paper B discusses several other uses for the FFT." Your paragraphs should be focussed on one sub-topic, and the paragraphs should be connected to each other.

Include a copy of your papers, and cite the source. (The sources will be returned to you, so feel free to turn in your only copy. They do not have to be in perfect condition; I would prefer to see your underlining, high-lighting, notes in margins, etc.) For long sources, you do not have to turn in the whole thing (i.e. don't photocopy a book), just the page(s) that is(are) relevant. Points will be deducted for relying on other sources (like webpages). You will be asked to present your summary to the class (if time permits).

Your literature review should be at least 5 pages, single column, double spaced, 12-point times font, 1 inch margins. Include any figures on additional pages. The papers must be cited in your review. You may do this by placing the reference number in square brackets. Be sure to have the references included in a "reference" section. If your summary exceeds 7 pages (of text), you may be asked to re-write it. Your review should not rely too heavily on the abstracts, introductions, and conclusions.

If you include a graph/figure, cite its source. Even if you made the graph yourself, be sure to say so, i.e. "figure 1 - spectrum [made by MCW]".

The following pages contain information needed for this assignment. Keep in mind that some of the information from the linked pages below may not apply to this assignment. If there is a contradiction, this assignment's specifics have a higher priority. If you are not sure, highlight the text in question and show it to the professor.

See how to quote a reference.

See project write-up for information about the style of references.

    You must turn in:
  1. A copy of the articles
  2. Your review

See comments from previous classes for examples of comments that I wrote on previous summaries and other technical writing assignments. If I have to write a comment, I also deduct points.

As of Summer 2003, I do not want you to use ANY acronyms in your paper summary, no matter how obvious they are. For example, if you want to mention HyperText Markup Language in your paper, spell out HyperText Markup Language, do not abbreviate it as HTML.

Both IEEE and ACM publish good conference proceedings and journals. Use your best judgement on other sources (i.e. an encyclopedia would not be appropriate, a webpage is not appropriate, most newspapers and popular magazines are not likely to contain good information for this assignment, etc.).