You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Currently, rmf_task::Request is used to define a bunch of requests (loop, delivery, clean, chargebattery) that are later converted into rmf_fleet_adapter::Task within the fleet adapter. Each rmf_fleet_adapter::Task is composed of a string of phases present in src/rmf_fleet_adapter/phases. The definitions of these requests/tasks are rigid and the process of introducing new tasks is fairly involved.
The proposal is to move some of the Phases into the public API in rmf_task. These would serve as building blocks to describe tasks. Hence requests would be described as a collection of these phases. Each phase/block would also inherit from an error handling class that dictates how errors during the phase should be resolved. Similarly for when the phase is cancelled. Each phase would also have invariant_duration and estimate_finish methods to allow the task planner to estimate the duration of the phase and the state of the agent at the end of it.
The text was updated successfully, but these errors were encountered:
Currently,
rmf_task::Request
is used to define a bunch of requests (loop, delivery, clean, chargebattery) that are later converted intormf_fleet_adapter::Task
within the fleet adapter. Eachrmf_fleet_adapter::Task
is composed of a string of phases present insrc/rmf_fleet_adapter/phases
. The definitions of these requests/tasks are rigid and the process of introducing new tasks is fairly involved.The proposal is to move some of the
Phases
into the public API inrmf_task
. These would serve as building blocks to describe tasks. Hence requests would be described as a collection of these phases. Each phase/block would also inherit from an error handling class that dictates how errors during the phase should be resolved. Similarly for when the phase is cancelled. Each phase would also haveinvariant_duration
andestimate_finish
methods to allow the task planner to estimate the duration of the phase and the state of the agent at the end of it.The text was updated successfully, but these errors were encountered: