ECE8990 Sensor Networks (Spring 2007)
From Ece
|
ECE8990 Sensor Networks is a graduate-level course in the Department of Electrical and Computer Engineering at Mississippi State University. In spring 2007, the course was taught by J.W. Bruce.
Contents |
Course description
The convergence of (low-power) hardware, software, and (wireless) networking has enabled a whole new class of computing network in which the computing nodes are able to sense their environment, collaborate together, and respond to acheive local-, wide-, or planetary-area results. The sensors themselves can range from small passive microsensors (e.g, smart dust) to large-scale, controllable weather-sensing platforms. Their computation and communication infrastructure are radically different from that found in today's Internet-based systems, reflecting the device- and application-driven nature of these systems. This course will survey the current sensor network literature, focusing on the hardware, software and architectural design challenges posed by these systems. Throughout the semester, we'll look to identify open research challenges and directions.
Grading
| Class presentation of papers | 15% |
|---|---|
| Paper summaries and class room participation | 15% |
| Simulation assignments (2-3 are planned) | 30% |
| Final project deliverables and report | 25% |
| Final project class presentation | 15% |
Paper reviews
Students are required to read, and review graduate-level papers throughout the semester. An average of 3-6 papers per week will be read. Written reviews of each paper assigned are due prior to the start of that class, and should be created in the appropriate location in this wiki. You must have a reasonable and complete review in the wiki before class begins, or you will be penalized/executed.
Papers will be reviewed and presented according to the following paper reading schedule.
For each paper, students should write a review answering each of the following questions:
- What problems (with prior work or the lack thereof) were addressed or surveyed by the authors?
- What solutions were proposed or surveyed by the authors?
- What are the technical strengths and main contributions of the paper's proposed solutions?
- What are the technical weaknesses of the paper's proposed solutions?
- What suggestions do you have to improve upon the paper's ideas?
Your reviews should be no more than a single page long. Ideally, you should be striving to review the paper thoroughly but succinctly. A half page summary is probably ideal for most but the longest and seminal papers.
Paper presentations
There will be papers to be discussed in class assigned to a student to present. Assignments will rotate thoughout the class. Papers will be assigned approximately one week in advance of the presentation date. The presenter of a given paper place their Powerpoint slides in the appropriate place in this wiki before class begins.
The in-class presenter of a particular paper does not have to submit written reviews for any of the papers reviewed that same day in class. The paper schedule may vary over the course of the semester, e.g. as new papers become published at the most recent conferences.
You should estimate spending 2-4 minutes per slide in your presentation. Your paper presentation should roughly parallel the written reviews that you have been writing. However, you cannot assume that the audience has read the paper that you are presenting. Your goal is to break the paper down into a digestable, manageable presentation to get the key points across to your audience. Here is a suggested outline for your talk:
- Big picture (2 minutes; 1 slide)
- What is the ultimate goal of this research? Motivate the topic.
- Problem Description (4 minutes; 1 slide)
- What exactly are the authors solving? How does it fit in the big picture?
- Technical Content and Results (6-9 minutes; 2 slides max)
- You need to prove that you understand the main points in the paper.
- Conclusions and Related Work (3 minutes; 1 slide)
- Relate the results to the big picture, and compare to other approaches
- Critique (5 minutes; 1 slide)
- What do you think about the paper (pros/cons)? What future research would be interesting to do in this area?
- Discussion, Questions and Answers
- Bring some questions for the audience, in case the audience does not have questions for you!
Simulation projects
Our simulation projects will be based on the sensor network simulator that we are developing here at Mississippi State. You will need to install the latest stable version of the Python language on your laptop. Python is very easy and quick to learn, especially if you already know an OO language like C++ or Java. The main Python website has links to many tutorials at several different difficulty levels. Please read/work through several to get Python-doctrinated.
Dr. Norman Matloff @ UC-Davis has some very good tutorials (for technical types like us) covering Python, SimPy, and other Python uses. I would encourage you to read these tutorials. They are quick and easy.
Then, when you have verified that Python works properly, you will need to install the SimPy discrete-event simulator module. The MSU sensor net simulator is built upon SimPy. If you want to make sure that your SimPy is working properly, you can try this supermarket simulation file.
The MSU sensor net simulator package (which is likely to change with each assignment) is available with the simulation assignments listed below:
- Simulation assignment #1 is due on 19 February 2007
- Simulation assignment #2 is due on 26 February 2007
- Simulation assignment #3 is due on 26 March 2007
- Project assignment is due at 11:59PM on 22 April 2007
The MSU sensor net simulator has available on-line documentation.
Sensor Networks Spring '07 Submissions
Research projects
More information will be forthcoming.
