Closed GokulEpiphany closed 2 years ago
Benchbot environments are challenges agnostic. They are designed to be used
for other tasks beyond our current challenge. It is the task that defines
the available control options for a given run e.g.
semantic_slam:passive:ground_truth
defines a task with passive control
(move_next).
On Sat, 4 June 2022, 12:20 am GokulEpiphany, @.***> wrote:
Shouldn't 'move_next' not be available in environments where we are expected to do active actuation (e.g company:5)? Please do clarify.
— Reply to this email directly, view it on GitHub https://github.com/qcr/benchbot/issues/71, or unsubscribe https://github.com/notifications/unsubscribe-auth/AK5HZJ7HK2HTGSNVRQRASBTVNIIBRANCNFSM5XZDLXRA . You are receiving this because you are subscribed to this thread.Message ID: @.***>
That makes sense, Thanks for the clarification!
Shouldn't 'move_next' not be available in environments where we are expected to do active actuation (e.g company:5)? cmd used to start: benchbot_run --robot carter_omni --env office:5 --task semantic_slam:passive:ground_truth -f
There are no constraints on which environments can be started for any given task? I understand running the following command, benchbot_run --robot carter_omni --env office:5 --task semantic_slam:active:dead_reckoning -f would start benchbot with action space not consisting of 'move_next'