Signal meanings must be learned
Aliases: cue learning · robot signal semantics · symbol comprehension
What it is
Learned robot signal semantics describes how instruction, examples, and stable repetition establish mappings between a light, sound, or pattern and robot state. Salience is not self-explanation. Public bystanders may encounter a robot once and have no opportunity for extended learning.
Why it happens
Arbitrary symbols acquire meaning through association and feedback, shaped by traffic, device, and cultural priors. Congruent mappings learn faster; opposite uses of the same colour across devices create negative transfer. Reliable consequences reinforce a mapping, whereas occasional inconsistency teaches observers to ignore it.
Studying it
First guesses, feedback-based acquisition, delayed retention, and cross-device transfer can be compared through accuracy, learning trials, confidence, and erroneous action. Detection, categorisation, and interpretation are separate stages. Trained-operator success does not establish bystander comprehension, and cultural or industry experience needs stratified sampling.
Where it stops holding
Some icons or motion analogies are relatively transparent but not universal. Training suits repeat professional users, not every passer-by; emergency meaning cannot depend on a long learning sequence. Demanding zero learning can also collapse complex state into an ambiguous symbol.
Applying it
- Reuse stable local mappings and inspect whether nearby equipment assigns a conflicting meaning to the same colour or tone.
- Give professionals action–signal practice with feedback; for bystanders, combine wording, direction, and the robot's subsequent behaviour.
- Test first-use comprehension, trained accuracy, and delayed retention separately, treating confident errors as more severe than uncertainty.