Present: Rich, Rob, Frederic, Conor and Kanna

Rich and Conor have managed to work out substantial detail in the boot up handshake (see attachment to this note) on startup of AMC and VCS. Hardware for AMC stack operations and tests is quite delayed. 

• Behavior addition/deletion working fine; this follows some design revisions based on transitions between behaviors. The AMC now sends across a number of behaviors based on its lookahead window.
  • Consequently an "arbitrary" length of behavior instantiated on the stack in VCS defined by this lookahead
• The implication to startup are that the AMC now has an initialization sequence
  • AMC will initialize the VCS once its all its reactors are fired up, especially the critical adaptor which communicates with the VCS.
  • Then the AMC will wait till it gets an init msg. from AMC Agent on the VCS side (see attached image) with a STARTED:0 (id ==0) coming from the VCS
  • An init.cfg startup file on the AMC side is equivalent to what the LC uses currently and will have* MissionTimer* DepthEnvelope* SetPoint:ID=0* Each behavior invocation comes with a unique and specific ID which is common to both AMC and VCS. This allows the ID to be used as a "handle" to manipulate the behavior running (or dormant) on the stack on the VCS side. In addition, it provides a unique identifier to the AMC when the state of the behavior changes on the VCS side. So when the Layered Control starts a behavior, it sends back a STARTED:ID message back to the AMC with this unique ID.
• One issue to resolve is read/write timing dealing with how quickly the VCS and AMC adaptors are able to send/receive messages. Conor and Rob are looking at timing issues on both ends.
• Conor will be building up the infrastructure for regression tests of the AMC allowing predefined mission scripts to be run with the simulator* We are now quite delayed with respect to getting an AMC stack up and running; the initial target date of end of Feb has slipped substantially and concerns were expressed about the implication of this slip to getting the actual board on the AUV itself. Kanna will be coordinating with Hans to work out what is doable realistically for the latter.
  -

AMC-VCS-handshake.jpg (image/jpeg)