Loop closure: what it means.
A SLAM system accumulates motion estimates as it explores. Recognizing a previously visited place provides an additional constraint that can help correct accumulated error in the estimated route or map. (MathWorks: SLAM)
This is different from closing a motor-control feedback loop. One is a mapping and localization operation; the other regulates a physical state such as joint position. The shared word loop does not make their algorithms interchangeable. (MathWorks: SLAM)
Put the definition to work.
A robot that recognizes a familiar corridor can use that match to improve its map. The phrase describes a mapping operation, not a motor-control feedback loop. (MathWorks: SLAM)
Questions to ask next.
Use these questions as a reading checklist, not as a device-selection recommendation. Exact manufacturer documentation takes priority over a general definition.
- Place-recognition method
- Consistency of observations
- Effect on the map and trajectory estimate
For the broader explanation, explore the electronics inside robots. This page defines the term; the companion guide connects it to the rest of the system.
Common Loop closure questions.
Does every familiar-looking image prove a loop closure?
No. Treat a candidate revisit as something the mapping system must evaluate. A valid match needs to be consistent with the mapping method and observations. (MathWorks: SLAM)