I’m exploring how to make robots reliably recover when something unexpected happens during task execution. I’m especially interested in practical engineering approaches, existing solutions, and the experiences of people building autonomous robotic systems.
I began exploring this autonomous robotics problem on 2026-11-09 at 15:27 UTC.
- One unexpected event I’m considering is: An unexpected obstacle blocks its planned path.
- Another unexpected event I’m considering is: A sensor starts providing incorrect readings.
- A further failure case is: Its navigation or manipulation strategy fails.
- I’m also considering what happens when repeated recovery attempts don't work.
The recovery choices I’m weighing are to retry, replan, use an alternative strategy, or request human intervention.
I’m asking: What do you think is the hardest part of solving this problem?
I have not settled on a recovery design or a specific engineering method; I’m seeking practical approaches and experiences from people building autonomous robotic systems.