Meeting Minutes
MRC-II System
Design Review 1
2:00 – 3:00 PM
Peter Kazanzides made the Design Review presentation and covered all topics up to the Risk Analysis (with a jump to a brief review of the High-Level Architecture Diagram). The following topics will be covered next week:
Risk Analysis
High-Level Architecture (in detail)
The following comments were made during this presentation:
• A Real Time OS would probably be necessary for real-time sensing and control.
• It may not be valid to lump all Windows versions together – i.e., there may be some relevant differences between Windows NT, 2000 and XP.
• It was acknowledged that it is easier to get device drivers for Windows than for Linux. Driver support for Linux is getting better, however.
• There were some questions regarding the applicability of the composite robot; e.g., would it also apply to closed kinematic chains? It was noted that the virtual interface extends this concept to any system for which you can implement the composition operations.
• The relationship between the Kinematics and the Trajectory Planner should be generalized, so that the Kinematics can be part of the Trajectory Planning and/or the Trajectory Generation.
• The External Watchdog is a highly recommended safety feature and perhaps should be required. Peter will look at a hardware power control system (using a PLC) that was designed by an undergraduate student.
• The requirement for a distributed robot object was challenged. We should reconsider whether it is really necessary to support a distributed robot object (e.g., client/server or CORBA) or whether this capability should be handled by the application. In other words, the application would be a distributed system, with one local module on the robot controller.
• It should be possible to generalize the loops (such as the Servo Control Loop) so that they can be used for other purposes, such as only acquiring sensor data.
