NOTE: This assignment, like others in this class, is due at the beginning of the class period. This means that if you are even a minute late, you lose 20%. If you are worried about potentially being late, turn in your reviews ahead of time. Do this by submitting them to me during office hours or by sliding it under my office door. Do not send assignments to me through email or leave them in my departmental mail box.
As discussed in the syllabus and in class, those students taking 5973-002, as opposed to 4970-001, are required to do an additional assignment. That assignment involves doing a literature search, writing summaries of appropriate technical papers found, making comparisons between these papers, and evaluating the appropriateness of the methods discussed for a particular task. Students not taking this version of the course do not need to complete this assignment.
The topic for your technical papers reviews is technical approaches to trust for autonomous robotic systems. Naturally, trust includes many psychological, sociological, and cultural issues. While these are interesting, we are concerned here with technical approaches to making robotic systems trustworthy. This includes such issues as reliability, robustness, and redundancy. Trust may be characterized by confidence values and may be related to proofs or guarantees of performance. The application task to consider is a lunar mission to explore former Apollo landing sites for the purpose of material testing.
You are to find publications describing five different technical approaches to trust. These may include approaches discussed in the textbook (e.g., sensor fusion) but should not be limited to such. You will need to find at least one publication describing each approach.
The exact number of publications you use is up to you to determine. If you find a single publication that describes an approach in sufficient detail for you to get a good understanding of how it works, that one publication is sufficient. If you discover that the one publication you have found on an approach is insufficient for you to understand that approach, however, you'll need to find more publications on that approach or move on to another approach.
The publications you use must be refereed technical publications. These include conference papers and journal articles (whether published in print or on the web) but not popular sources such as magazines (e.g., Discover Magazine) or un-refereed sources (such as most web sites, even for departments or labs). If you are in doubt about a possible source, you should check with me before using it.
For each of the five approaches for which you have found publications, you are to write a summary that includes the main idea behind the approach and discusses the components of that approach and how they fit together. You should relate the approaches to concepts from the textbook as appropriate.
Each approach summary should run from 1.5 to 2 pages in length (roughly 80 characters per line, 50 lines per page).
Please note: Taking the first line or two from each paragraph in a paper, stringing them together, and changing around a few words here or there to make things read better, is NOT a summary. It is plagiarism -- a form of academic misconduct. Any time you quote a source, you must include the quotation in quotation marks and clearly indicate the source of the quotation. If you find yourself with more than a couple of brief quotes in each summary, then you are quoting too much. To summarize a paper, you need to (1) read it, (2) understand it, and (3) briefly relate its main points in your own words. If you don't have your own words to describe the approach, that probably means that you don't understand the paper -- you'll need to go back to steps 1 and 2 and visit me during office hours as needed to help you with step 2. (I don't expect most students to have problems understanding the difference between a summary and plagiarism. This message is for those few who do.)
You are to write a comparison the five approaches, pointing out both similarities and differences between them. Again, you should use concepts from the textbook as appropriate in making these comparisons.
The comparisons of all five approaches to one another should run a total of 1.5 to 2 pages in length (roughly 80 characters per line, 50 lines per page).
You are to write an evaluation of each approach with respect to the task of exploring former Apollo landing sites for the purpose of material testing. Think of the kinds of activities the robot(s) should carry out and discuss which approaches would increase our trust in a robot system engaging in these activities. You should include both your reasoning and your conclusion as to how suitable each approach is for the task.
Please note: You are not being asked to say which approach is "best" for this task. Rather, you are being asked to say how suitable each approach is for the task. Approaches may be complementary, rather than exclusive. If so, you should indicate this in your evaluation.
The evaluation of all five approaches on this task should run a total of 1.5 to 2 pages in length (roughly 80 characters per line, 50 lines per page).
As can be computed from the above, your completed reviews should be roughly 10.5 to 14 pages in length. If you can complete this assignment in 10 pages or less, that is fine -- there is nothing wrong with being concise as long as you are being complete. If, however, you run over 15 pages, you are being too verbose and will be penalized. This limit does not include any figures, which are encouraged and may take up any amount of space, nor citations.
Turn in complete citations of these five papers. Here are guidelines for what a complete citation includes.