Massachusetts Institute of Technology 16.412/6.834J Cognitive Robotics PS 4: Final Project Description Due Dates: Project Proposal - April 11thProject Presentations – May 9th and 11th. Final Project Reports – May 11th. Objectives The purpose of the project is for you to develop a deep understanding of how to apply a focused set of methods towards the creation of a simple cognitive robot, to be able to communicate these methods to others in a simple and intuitive manner, and to be able to innovate upon these methods, or to lend novel insight into their behavior. More specifically, you should demonstrate the ability to: • Clearly state and motivate an interesting, focused innovation to cognitive robotics. An innovation may be an innovative application, a novel algorithmic extension and/or the evaluation of an important analytical question (already exercised in PS2). • Extract and evaluate the relevant literature using the web and library resources (already exercised in PS3, your advanced lecture). • Provide a simple, clean tutorial explanation for one or two novel algorithms used in your project, and extract simple pedagogical examples that highlight the key features of the algorithm. • Implement and demonstrate the application of your algorithm in support of your project goals. • Evaluate the approach analytically and/or empirically.Project Grading • A – represents mastery: the ability to analyze, extend and apply existing methods in a way that is novel and insightful; the ability to explain and motivate in a manner that is particularly intuitive. • B – represents solid competence: the ability to articulately motivate, explain, implement and evaluate a focused set (i.e., 1 or 2) of intelligent embedded systems methods. C – represents partial competence of the above. General • The results of your project are to be captured in a written document of roughly 10 to 30 pages (depending on team size - ~ 10 pages per member) and are to be presented orally, with slides. • You may, if negotiated with me in advance, feel free to use your work in this class to extend or complement the work you do within some other project. • You are welcome to work together in teams, just like in the real world. This will allow you to tackle more ambitious projects. Teams of two typically work best. Each team may produce a single written reports are to be done, as everyone is to have that experience. Important Deadlines The project includes a project proposal (due Monday, April 11th), a presentation of results (May 9th and 11th) and a final report (due May 11th). The deadline for each is given below.Project Proposal (due April 11th): The proposal will be up to 6 pages. It will be due in class on Monday, April 11th. The purpose is to for you to jump start the project before it is too late to get anything done, so take it seriously. Use the project proposal time to do all your background work. Plan your project in two spirals, each of roughly equal length. In the first spiral you should implement a complete, but simplified version of your system. Stub out most of the difficult components, but make sure that you have place holders for all modules. In addition, make sure that you have exercised all interfaces. The purpose of the first spiral is to place on the table all key issues. During your second spiral, elaborate the modules to their full capability, in light of the time available. At the end of the first spiral, make sure to carefully assess where you stand, and the difficulty of each module, and then adjust the scope of your project accordingly. Your proposal should include: • Cover page • Introduction, including motivation for the project and assessment of its value to the technical community. • Concise problem statement outlining of the objectives of the project. • Statement of previous work and methods that you build on. This is where you do your background work. • Outline of the proposed technical approach. • Plans for Spiral 1 and Spiral 2. • Planning, including division of labor and time schedule. Also feel free to send me a one page email bouncing your ideas off me. I’ll be on travel for a week, but will look at them when I return. Project Presentations (due May 9th and 11th):You will give a brief presentation of what you’ve learned in your project. These presentations will be May 9th and 11th. The exact length of the presentation depends on the number of projects, but estimate roughly 8 minutes per person. Project Reports (due 11th): See the project objectives at the beginning of this handout for the key elements of your project report. Your final project report will be due at the end of the day during the final day of classes, May 11th. Please submit your final project to the course Secretary, Brian O’Connail. Include a hardcopy of your final project, and and a CD or .zip file, containing your project writeup, project presentation, source code for your project, and any explanation required to run your code. Types of Projects Projects can be one of the following types: • A simulated area exam. This concept is most familiar to Course 6 students. You read a focused set of papers and do a scholarly critique. This should consist of one to two primary papers, or at most three, that describe a couple, closely related methods. Part of the process is to pick a set of dimensions along which to compare and contrast the papers. To demonstrate competence you should clearly explain the approaches and perform a careful analysis, often including an implementation of one or two of the methods and some empirical evaluation. This is an enhancement of what you did in problem set 2. To demonstrate mastery you should move beyond comparison, highlighting novel insights about the work and by proposing interesting ways in which the different research approaches can be synthesized or extended. You conclude with a description of the degree to which you are moved, and why. For example, you could contrast a Bayes net approach to state estimation to estimation using probabilistic hybrid automata; you could compare two recent papers on planning based on propositional encodings, plan graphs and model-checking; or you could contrasttechniques for tracking hidden Markov models both from the POMDP and model-based reasoning literature. When you have a rough idea of what you might want to do, please bounce the idea off me.I
View Full Document
Unlocking...