EECS 249B
Spring 2015
Contents
Home
Logistics
Lectures
Discussions
Assignments
Projects
Software
References
Seminar
Wiki
Course Development
SVN
|
EECS 249 Course Project
Project Presentations: Dec. 8, 2009
Proposed Project
| Title |
Synthesis Flow for Energy Efficient Building Control and Automation Systems [2 Projects]
|
| Category |
Green Building, Distributed Embedded Systems
|
| Description |
The design of future generation energy efficient buildings will rely on control algorithms that are capable of fine-tuning energy consumption. To be effective, these algorithms need to be supported by a networked system that provides three services: value reading of physical quantities from sensors in different locations, control action dispatching to actuators, and computation of control laws. The computation service is usually distributed on a set of processing units interacting through the network that spans building complexes. To address the scale and complexity of control systems for large buildings, we have proposed a synthesis flow for optimizing building control and automation systems.
In this project, students will design tools and/or mechanisms for the key steps in the synthesis flow. The students can choose one of the following topics.
1) Automatic translation for MMM-based Interchange Format (IF)
In order to integrate different tools and modeling languages, we proposed to use IF in our synthesis flow. The synthesis flow adopts two abstraction levels. The first abstraction level will read as input a description of the control algorithm in the form of a high level modeling language and will generate an intermediate representation IF. After that, the second abstraction layer will be able to translate a program written in the IF language to any of the existing building control languages. We chose to use MMM-based IF because it support hybrid, heterogeneous, and hierarchical systems, and its definition of semantics is the key to enable unambiguous translations between models.
The students will be asked to design and develop a multi-level automatic translation tool between MMM-based IF and a commonly used modeling language (such as Simulink or Modelica). An optional extension is to leverage Metropolis framework for validating the translation.
2) Communication interface synthesis for distributed real-time systems.
When mapping a function model to an architecture platform, communication interface synthesis is used for preserving semantics. For example, communication interfaces and buffer protocols are designed to prevent data loss on the communications between distributed components.
In building control and automation systems, typical architecture platforms are distributed and asynchronous, where processing units, sensors and actuators are running with their own local clocks. Given a building control algorithm described as a synchronous model, the students are asked to design a communication interface mechanism to preserve semantics of the synchronous model when mapping it to the architecture platforms. The students need to validate their mechanism through case studies or formally prove it.
|
| Mentors |
Alessandro Pinto, Yang Yang, Mehdi Maasoumy
|
| Required knowledge: |
Basic understanding of relevant modeling languages for the first project. Basic knowledge of synchronous models and distributed systems for the second project.
|
| Readings: |
For the first project:
[1] Alessandro Pinto, Alberto L. Sangiovanni-Vincentelli, Luca P. Carloni, Roberto Passerone, "Interchange Formats for Hybrid Systems: Review and Proposal", Hybrid Systems: Computation and Control, 2005.
[2] Alessandro Pinto, Luca P. Carloni, Roberto Passerone, Alberto Sangiovanni-Vincentelli, "Interchange Format for Hybrid Systems: Abstract Semantics", Hybrid Systems: Computation and Control, 2006.
For the second project:
[3] Stavros Tripakis, Claudio Pinello, Albert Benveniste, Alberto Sangiovanni-Vincent, Paul Caspi, Marco Di Natale, "Implementing Synchronous Models on Loosely Time Triggered Architectures", IEEE Transactions on Computers, Oct. 2008.
|
| Title |
Composition of analog components in feedback systems
|
| Category |
Analog |
| Description |
Design reuse is an important part of increasing the design productivity. However, for analog systems, this is difficult due to the dependence of the component behaviors on the external environment. Thus, component performance becomes very hard to predict, inhibiting design reuse. Recently, we introduced contracts into the analog platform-based design (APBD) methodology as a way of facilitating compositionality.
In this project, the students will evaluate and improve upon the contract-based analog composition methodology for feedback systems.
The students must first characterize a basic analog block (e.g. op-amp or gm-C cell) using the APBD flow, then compose it in a feedback system (e.g. low-pass filter) using the contract-based design methodology.
|
| Mentors |
Xuening Sun, Pierluigi Nuzzo, James Wu |
| Required knowledge: |
Understanding of analog circuit design. Design experiences using Cadence. Matlab and Ocean scripting experiences a plus. |
| Readings: |
[1] De Bernardinis, F., Gambini, S., Vincis, R., and Svelto, F. 2004. Design space exploration for a UMTS front-end exploiting analog platforms. In Proceedings of the 2004 IEEE/ACM international Conference on Computer-Aided Design.
[2] Xuening Sun, Pierluigi Nuzzo, Chang-Ching Wu, Alberto Sangiovanni-Vincentelli; "Contract-Based System-Level Composition of Analog Circuits." DAC 2009 Proceedings. July, 2009.
|
| Title |
Metro2 & Ptolemy Integration
|
| Category |
Design Tools |
| Description |
The purpose of this project is to link Ptolemy and Metro2 in the way that designers are able to graphically build their function models in Ptolemy and map them onto virtual architecture models in Metro2. This flow will allow people to do design space exploration very efficiently in early stages. The student may need to use the code generation of Ptolemy and develop a 'wrapper' to automatically wrap the generated code into Metro2 components. We may start from the some simple designs, and try to see how far we can go.
|
| Mentors |
Liangpeng (Leo) Guo, Alessandro Pinto |
| Required knowledge: |
Strong programming and software engineering experience. |
| Readings: |
SystemC, Metro2 and Ptolemy documents. |
|
