Deep Dive
How does a robot vacuum know when to recharge and resume cleaning?
It plans ahead. The robot estimates remaining battery against distance to the dock, returns before it runs out, then recharges. Strong models keep persistent state, the saved map plus progress against it, so they resume from where they stopped rather than restarting and re-cleaning finished areas.
For a small apartment the battery is never the constraint: the robot finishes long before it runs low. Scale the floor up, and the most important behavior in the product becomes one almost nobody asks about before buying. A robot working a large home will drain its battery mid-job, and what it does in that moment, whether it abandons the run or banks its progress and returns to finish, is the difference between a machine that handles the house and one that handles a room and gives up.
The feature has a dull name, recharge and resume, and it hides a real autonomy problem. Pulling it apart shows why two robots with the same battery can deliver completely different results on the same floor.
The decision loop
Quit, dock, charge, and return to the right spot
- The dock does itruns between cleans without anyone present
- You do itstill a manual step on every all-in-one dock sold
The decision to quitKnowing when to stop is the first skill
The first judgment is when to break off. A robot has to predict whether it can finish on the charge remaining, and if not, it must leave with enough reserve to navigate all the way back to the dock. Misjudge this and the robot dies in the far bedroom, stranded out of reach of its charger, which is a worse outcome than stopping early. Good battery management leaves a margin for the return trip. Weak management runs until the warning and then gambles on reaching the dock.
This sounds like a power-management detail. It is actually a planning problem: the robot is reasoning about its own future state, estimating consumption against distance, and committing to a decision with a real cost if it is wrong. That is autonomy in a narrow but genuine sense.
The hard partResuming where you left off, not starting over
Coming back is where products separate. A weak implementation returns from charging and starts the entire job again, re-cleaning everything it already finished and often never reaching the part it missed, because it drains again on the same already-done areas. A strong implementation remembers the boundary between done and not-done, navigates back to that line, and continues from it. This requires persistent state that survives the charging gap: not just the map, but the progress against the map.
Persistent mapping is the enabling capability. A robot that builds its map fresh each run cannot meaningfully resume, because it has no durable record of what it completed before docking. A robot with a saved map and saved progress can treat the charge as a pause rather than a reset. This is why resume quality tracks closely with whether a robot has true persistent maps, and why it is worth checking before buying for a large floor.
What a large floor can require
2 trips
On a big home the robot may dock and resume more than once in a single job. Whether it continues or restarts decides if the floor ever gets finished.
Why it caps the classificationResume is an endurance test for autonomy
Recharge and resume is a quiet ceiling on how independent a robot can really be. A machine that needs a human to restart it after every charge, or that never finishes a large floor because it keeps re-cleaning the start, is not operating autonomously over the timescale that matters. On the Autonomy Ladder, the ability to manage your own energy and continue a job across a charging gap is exactly the kind of unsupervised, multi-cycle operation that separates a higher classification from a lower one.
For a small space, none of this matters and the cheapest robot is fine. For a large or multi-floor home, recharge and resume is one of the highest-value things to verify and one of the least demonstrated in a showroom. The robot that quietly banks its progress and comes back is doing the unglamorous work that autonomy actually consists of.
Finishing a small room is easy. Finishing a large floor across a charging gap is the test that separates a tool from a timer.
