Engin 103

Project 2

Good Design: Data Modeling and the Predictability a System

Predictability and reliability are certainly two necessary features of a well-designed system. Now that you know how to do data modeling, interpolation (smoothening) and prediction on any set of data, you can apply it to check the predictability of your system. Any system can be modeled as a box to which some input quantity X is fed and from which some output quantity Y come out. If you do data modeling on (X,Y), the system can be described with an equation.

Think of a task then design and build a system to perform that task. The simplest design that can perform that task is usually best. Common machines require humans as operators, and humans as complex systems work under the influence of many factors that easily lead to unpredictable behavior. Minimizing human intervention, ideally to an intelligent ‘digital” response of yes/no, helps increase predictability. For example, I used to drive a long distance on a regular basis, by using highways, very-low-traffic schedules, and cruise control, my total travel time from point A to point B could be predicted within 1% of error.

In this project, you are required to work in teams to design and build a system, then demonstrate its predictability. That is, you will need to model the some data X and Y by relating them with the best equation Y’=f(X), where f could be a polynomial, exponential, or any other elementary function in X. Once the best equation is obtained, it will be possible to predict Y’ for any X. If you take into consideration engineering principles and how to eliminate unpredictable behaviors as discussed above, your system should be very predictable, that is, the predicted value Y’ should be very close to the actual value Y. First day you will present your system, second day the equation when you will be required to make a prediction Y’ for some value X, given by the audience, with your model. Then show the actual result from your system. Performance grade will be based on how close the predicted and actual values are from each other. Grades from both presentation days will be averaged.

Total materials cost should be less than $20, copies of receipts to be submitted with project reports.

Each team will do a 5 minute presentation on their device in each of the two presentation days. In the first day, the teams will give an introduction (what they did, how they achieved predictability, etc.) followed by the demonstrations, in the second day the team will present the model and prediction. The webpage on the project, along with the project report will be due the class after the second day of the presentations (please check the e-syllabus for exact dates). The team leader will meet with the instructor to discuss team progress on the project on a weekly basis. The project report is expected to be a good written document (see Good Writing Practices), and graded under three categories: correct grammar and neat presentation; logical arguments and structure; accurate report of the team project, completeness, and no plagiarism. Project report will be submitted in hardcopies with member signatures and also in electronic form (see Computer Files: Names and Electronic Submissions). A complete report should include the following sections:

-Introduction: brief description of project objectives in your own words, background information needed for the design with emphasis on predictability of your mechanism, work distribution among the team members, and timelines for the different parts of the project: research, design, building, analysis/calibration.

-Design and building: this section should include sketches and diagrams: how the different design elements and hardware components were selected enhance the predictability of the mechanism, the list of components with specification and prices.

-Analysis/Calibration: should include data obtained from your device: tables of X and Y, insert of Excel worksheet with discussions of different types of function relating those two variables and pinpointing which function is the best (with the smallest “standard deviation”, see CW 3, 4, 5). It should also include results from testing the prediction made by this model and an assessment of the predictabily of your device. It should also include a brief manual of operation, troubleshouting list, and appropriate recognition of other author’s materials if used in your project.

-Conclusion: overview of the team achivement and lections learned for the future.

Grades will be computed as follows:

Items			Points
Project completed and presented			70
Project performance (perform tasks specified)			50
Good design			30
Project presentation and webpage			50
Written reports	Report submitted		70
	Project report: will not accepted without all member’s signatures on percentage of participation	Grammar and presentation	10
		Logical arguments; structure	10
		Accurate and complete	10
Total project grade			300