Physical indication of sensor orientation supports on-site judgement better than documentation
Aliases: sensor direction indicator · field-of-view cue · perception orientation
What it is
A physical sensor-orientation cue uses a lens window, gimbal pose, shell axis, light, or projection to reveal primary sensing direction on site. A person can judge likely visibility from current pose without recalling model-specific fields of view and coordinate frames.
Why it happens
Documentation requires model recall and transformation into current pose. A co-located cue removes that step and moves with the sensor. Decorative eyes off the true axis, or a wide lens that appears narrow, create confident error.
Studying it
Documentation, body marking, dynamic light, and projection can be compared on coverage judgement, time, confidence, and chosen position across poses. Ground truth should be measured detection rather than geometric field alone. Occlusion, bright light, multiple sensors, and expertise matter.
Where it stops holding
A cue approximates a probabilistic region and cannot guarantee recognition inside it. Omnidirectional sensing, fusion, or software cropping may lack one axis; effective coverage or current attention should replace a false cone. Safety must not depend on people placing themselves correctly.
Applying it
- Align sensor, cue, and software-used field, updating with gimbal motion and cropping.
- Audit decorative eyes, windows, and lights for competing false fronts.
- Ask first-time users to mark visibility around several poses and compare with probability maps, prioritising confident misses.
Related
- Same group: X3.06.1 Robot sensing is not omnidirectional, so blind spots must be known · X3.06.2 A robot should behave cautiously when a person is in a blind spot · X3.06.4 Sensor fusion can mask component failure without eliminating system blind spots
- Adjacent: X2.03 Orientation and gaze · X3.01 Visible current state
- Search terms:
sensor orientation cue·field-of-view indicator·perception visibility