Having different reconfiguration time delay for planning

[pooyanjamshidi]

We may want to have a different time delay for each reconfiguration of the robot. I think this makes sense both for CP1 and CP3. These delays are adaptation costs. We need to discuss and think through these further in a meeting @schmerl @javier-camara. We may want to think whether this may change the interface with LL or if not how that may change the quality of adaptation.

[schmerl]

I'm not sure what this is about, @pooyanjamshidi

[pooyanjamshidi]

This is the idea we discussed at ABLE, so the reconfiguration may have different enaction latency, in CP1 we can simulate this easily and in general, I think this is a nice research idea to investigate. I think this is one of the ideas we would like to consider in phase 3 regarding online learning so we learn these latencies and we consider this in adaptation.

[schmerl]

OK, so not intended for Phase 2.

On Thu, Jun 21, 2018 at 9:03 AM Pooyan Jamshidi notifications@github.com wrote:

This is the idea we discussed at ABLE, so the reconfiguration may have different enaction latency, in CP1 we can simulate this easily and in general, I think this is a nice research idea to investigate. I think this is one of the ideas we would like to consider in phase 3 regarding online learning so we learn these latencies and we consider this in adaptation.

— You are receiving this because you were assigned. Reply to this email directly, view it on GitHub https://github.com/cmu-mars/issues/issues/78#issuecomment-399095748, or mute the thread https://github.com/notifications/unsubscribe-auth/AD9_AOaIXGWJOjXsGM_KboesJuVANPTsks5t-5mEgaJpZM4TzH8F .

[ivoysey]

I'll move it to the Phase 3 milestone so we don't think it's something we need to do by next month.