CMPE259, Spring 2013, Section 01: CMPE259 Class Projects

As an advanced graduate class, CMPE 259 has a significant project component. Here, you will find some ideas for class projects. You are welcome to propose your own project idea which you will need to discuss with the instructor.

Note that these are project ideas and NOT necessarily project descriptions. In other words, these ideas need to be fleshed out. Feel free to discuss them with the instructor.

IMPORTANT: You must submit a project proposal by the end of the 3rd. week of the quarter (by April 19th, 2013). Project proposals must contain: the topic of your project, the motivation for the project (i.e., why did you pick this project? why is it important/timely?, why hasn't it been solved before?), related work, the approach/design you plan to use, experimental methodology (how you plan to evaulate your solution), a demo plan (note that demos will depend on the type of project), and a timeline you plan to follow to carry out the project during the quarter.Project updates will be reported during the quarter.

Project deliverables include: a project report (in the style of a technical paper you would submit to a conference), your source code, and a project presentation. Project presentations will take place on the day of the final exam.

Some project ideas: 

1. Temperature and humidity monitoring

Using TinyOS 2 and a number of temperature and humidity sensing nodes (e.g., TelosB motes) implement a temperature/humidity monitoring sensor network. The network should have 3 "tiers", namely: sensing nodes, data fowarders, and data sinks. The system should include:

      1. Communication between sensing nodes and data forwarders.
      2. Communication between data forwarders and sinks.
      3. Data provenance, i.e., identifying where the data is coming from.
      4. Accurate temperature and humidity information.
      5. Storage mechanism for collected data.
      6. Simple Web interface to access data.

 

2. Smart home for remotely-assisted living

Create a home-area network (HAN) that connects sensor feeds, alarms, etc. and includes a Web-based user interface to access information provided by the HAN that will enable remote monitoring of the home. In this current form, the project is quite general and open-ended. This means it can be tailored based on specific interests, skill set, etc. If you are interested in the project, you should refine the project proposal and run it by the isntructor.

 

3. Implement epidemic routing [1] in a mote-based sensor network.

Use link quality [2] as a utility function to control the epidemic "spread". Evaluate the performance-power consumption trade-off of using link quality estimation. 

[1] Amin Vahdat and David Becker, Epidemic routing for partially-connected ad hoc networks, Duke University Technical Report CS-200006, April 2000. 

[2] The Link Estimation Exchange Protocol (LEEP) TEP 124. 

 

4. Implement the Directed Diffusion protocol.

Implement the Directed Diffusion protocol (see the Directed Diffusion paper in the Routing reading list) on top of the Collection Tree Protocol [3]. In the original implementation of CTP uses link quality (as given by LEEP [2]) as the routing metric. Compare the perfomance of the resulting scheme if hop count is used as routing metric.

[3] The Collection Tree Protocol, TEP 123. 

 

5. Power analysis of routing.

This project aims at characterizing the power consumption of routing in a Mote network. The characterization should be done by direct measurements as well as instrumentation of the operating system. Selection of the routing protocol is part of the project. The basic functions executed by routing should be identified and their power consumption analyzed. 

 

6. Instrument a network simulator, e.g., QualNet, to account for energy consumption.

 

7. Implement a localization mechanism using motes.

Study the impact of factors such as placement of the motes, as well as the number of motes used as beacons. 

 

8. Energy consumption characterization of a sensor network platform.

 

9. A Dynamic Scheme to Boost QoS of 802.11n (and e) Based on Traffic Forecast (a tentative title)

Description: This project aims at investigating a scheme for improving the QoS service of IEEE 802.11n (and e) even when 802.11 networks are heavily congested by adjusting the speed of waking up frozen classes in proportion to network congestion. Currently 802.11n (e) maintains four classes or outgoing queues at each node to prioritize traffic by assigning different values of three contention-related parameters on the classes, namely IFS (InterFrame Space), the initial CW (Contention Window) size, and finally the maximum CW size. This Class-of-Service approach, however, suffers the severe degradation of QoS as the network becomes crowded since more frames are likely to be transmitted simultaneously regardless of their urgency, leading to deterring high-priority traffic from being promptly sent.

With the goal of alleviating this rapid deterioration of this QoS service, this project consists of two tasks in detail. The first one is to explore a traffic forecaster which can accurately predict the 802.11 network congestion state in near future. Especially, it will evaluate the applicability of the traffic forecaster proposed in a paper titled as “Collision-Free Medium Access Based on Traffic Forecasting” to 802.11 networks rather than applications for which this technique had been designed. In detail, this forecaster will be compared with other ones, especially EWMA (Exponential Weighted Moving Average) and its advanced one which dynamically adapts the weight factor in EWMA equation according to the rate of changes. This task will also figure out some way to generate some network traffic in simulators such as NS-3 behaving in the same way as real ones.

The second task is to evaluate the effect of a new parameter such as the rate of running back-off timers on differentiating traffic. Up to now all the researches about the QoS have focused on adjusting the three differentiation-related parameters of IEEE 802.11 according to the measured channel state. All the conventional techniques are about how to set these three parameters when the underlying channel is free and new frames are ready to be delivered. Any research has not paid attention to postponed frames of low classes which may have higher priority in the current completion since their timeout has been lessened in the last competition. This project will come up with a technique to dynamically slow down back-off timers associated with deferred frames of low classes as collisions occur more frequently with an aim to give more chance to high-priority classes without sacrificing the total network utilization. Finally, it will also evaluate some trade-off between QoS and fairness.