A preview must precede rather than coincide with action
Aliases: anticipatory timing · cue lead time · feedforward interval
What it is
Preview lead time is the interval between an intent cue and the corresponding robot action. A cue concurrent with first displacement may explain an event already underway, but it cannot support prediction, decision, and bodily preparation. Feedforward requires a cue that precedes action while remaining clearly associated with it.
Why it happens
Human response includes detection, identification, response selection, and execution, with additional cost from display delay, occlusion, and attention switching. Too little lead time lets motion begin before response completes. Excessive lead allows the cue to decay from working memory or become an unreliable promise as plans change. Timing therefore depends on consequence and required response, not a universal duration.
Studying it
Lead time can be varied while cue content and motion remain constant. Pre-motion prediction, gaze shift, completed yielding, conflict, and unnecessary waiting reveal effectiveness. Threshold estimation provides a task-specific minimum but requires replication under target distance, noise, and crowding. Cancellation and rerouting trials expose the false-commitment cost of long previews.
Where it stops holding
Longer is not always better. Previewing every high-frequency micro-motion creates load, and unstable plans can only signal probability rather than certainty. Large distant motions may require early preview, whereas harmless internal adjustment may not warrant interruption. Accessibility needs also change detection and response time.
Applying it
- Work backwards from the human response, budgeting detection, comprehension, movement, and safety margin rather than triggering cue and control command together.
- Define lead-time ranges by speed, distance, occlusion, and consequence; visibly retract stale cues when plans change.
- Log cue onset, human detection, human movement, and first robot displacement. Validate the proportion of safe responses completed before motion begins.