Human vision is not binary which the core problem with the most simple vision model the areas where our vision is great and areas where the vision is great, the vision is best directly in front and degrades with distance or toward our periphy can be approximated using serveral vision zones.
if we forget about the boolean vision model then we have to introduce different kind of continuous scale we can clearly use a flotating point scale while it's tempting think of the scale as representing probability that flawed approach.
that is not really what human experience and it has old implications, how often should an agents roll a die to check if it sees an object, if the die rolled often enough eventually the object will be seeneven if has a very low value.
this its much different from the boolean vision model but it has added deriment that its more unpredictable and probably inexplicable the player building a model in their head of what the agent is experiencing thus Artificial Intelligence behavior will appear random and arbitrary.
if the value is a little higher than zero the agent sees the object but just is noy sure what it is essentially the object is too blurryto identify the value climbs toward one then the certainty of knowing identity the onject becomes much higher, this avoids any randomness and its more consistent with what humans experience.