Daily Updates

8/15/02

Greetings,
    Today we met as a group and further developed the ANTS system.
Now we have a division of high and low level AI.  The underlying hardware,
which will need to be initially modeled for our simulation falls into
neither a high/low level, as the hardware has no intellegence.  Mike's
simulation will be providing the environment for the simulation.  Based on
Deep Space 1, we will have a modified system that has 4 layers built ontop
of the hardware.  These layers will consist of the  1) Planner: stipulates
mission goals and directives 2) Scheduler: given mission goals by the
planner and produces a step-by-step list of actions that the ANTS system
needs to carry out 3) Monitor: 'overlooks' the commands sent by the
Executor and monitors the state of different hardware components to
oversee progress or diagnose problems 4) Executor: receives action list
and carries it out, operating subsystem components as necessary.  The high
level controls will be given to the planner and scheduler, the low level
controls will be given to the executor.  A divison between high and low
level will be made for the monitor.  The system will work in a way that
the problems travel up the chain of command, first registering with the
executor (who can potentially alter its method of fulfilling a listed
command, or merely try again), then the Scheduler (if Executor cannot
resolve the problem, Scheduler can potentially schedule a different action
to be carried out), and finally the Planner (if the Scheduler cannot
resolve the problem, the Planner may need to reassess its goals).
    In addition to our great advances in terms of planning, we updated the
website (and had our webpage listed on the WPI CS MQP webpage).
Additionally, work was done on the summary for this friday, the website
was updated and work was done in Java 3D, as we will be using this to view
the state of our simulation.  Everything is going well!

Sincerely,
ANTS2