DARPA Embedded Software Working Group September 27-28, 2001 Ford Research Laboratory Dearborn, MI

Introduction Agenda Hotel Information Register Online

Thursday, September 27, 2001

08:00-08:10 Welcome (Milam) - Participant Self-Introductions - Agenda 08:10-08:30 Introduction (Bay) 08:30-09:30 Phase II/OEP Progress - Overall: Goals, Schedules, Progress (Misener) - Baseline Solutions to Challenge Problems and Indications of Phase I Involvement to Date (Varaiya, Girard, Griffiths) 09:30-09:40 BREAK 09:40-12:20 Experimental Results II (for each, 30 minutes per note 1 below) - 09:40-10:00 CMU (Clark) - 10:00-10:20 CMU (Krogh) - 10:20-10:40 UC Berkeley (Ptolemy) - 10:40-11:00 Kestrel - 11:00-11:20 SRI (SAL) - 11:20-11:40 University of Michigan (AIRES) - 11:40-12:00 University of Pennsylvania (DIVES) - 12:00-12:20 Vanderbilt (GME) 12:20-01:20 LUNCH 01:20-03:20 Mid-Term Experiments I: Vision & Measures of Performance - Discussion of Phase I Involvement and Plans (Varaiya/All) - Allocation of Roles and Responsibilties (Varaiya) 03:20-03:40 BREAK 03:40-05:00 Tool Interfacing and Integration Discussion (led by Gabor Karsai) 05:00-05:30 Recapitulation and Preliminary Action Items for Day 2 (Varaiya)

Friday, September 28, 2001

08:00-08:15 Summary of Prior Day (Misener) 08:15-09:00 Automotive OEM Partners Present Evaluation of State of Phase I work - Ford - GM - Motorola - UC Berkeley - Government 09:00-10:00 Continuation of Interface discussions (Karsai/UCB) 10:00-10:15 BREAK 10:15-12:00 Mid-Term Experiments II: Interaction and Planning (Round-table) - Discussion/Plans/Actions - Recapitulation and Interim Action Items (Misener) 12:00-01:00 LUNCH 01:00-01:45 Tours (Milam) 01:45-02:45 Continuation of Prior Discussions & Unfinished Business/Interactions 02:45-02:55 Meeting Recapitulation and Final Action Items (Misener) 02:55-03:00 Closing Remarks (Bay)

¹Experimental Results should contain technology demonstration(s) and address the following:

1. Tool benefits to development process
2. Tool interfaces (inputs, outputs) and integration with other tools
1. In nearly every case it seems necessary to manually translate OEP models (written in Simulink/Stateflow or Teja) into "tool-native" input language - solutions?
2. In the cases where the different analysis tools have common objective (eg. schedulability, code generation, verification) how do you suggest we compare outputs of different tools?
We hope your thoughts will seed the subsequent "joint discussion" segment of the meeting.

Direct comments concerning this meeting to: Nikki Ford