5
$\begingroup$

I am just wondering about this. I am a computer science student, and I have been looking into the SLAM (Simultaneous Localization and Mapping) problem. The problem is that you want to both create a map of the environment, while at the same time localizing your position in this map. You can use any suitable sensor, most commonly camera or LiDAR. This problem seems to be solved (though there are still technical improvements being made and depending on the types of sensors and requirements, there are different algorithms being developed that can trade-off different factors).

I am assuming the human eyes are a bit similar to a stereo camera and so they should be able to internally quite accurately estimate the depth of what they look at.

Now the question is, why wouldn't evolution have given us the quite useful ability to accurately build a map of the environment? I know, we do kind of have that ability, but it seems to be still very bad compared to what robots using SLAM algorithms are capable of.

Maybe my question is a bit naïve, as I don't know much about the capabilities of the human brain and eyes. Though as far as I know, there are low power devices like FPGAs that are capable of performing SLAM algorithm (src).

$\endgroup$
1
  • 1
    $\begingroup$ There are "map memorisation systems" for want a better term, like The Knowledge. Plus, coach drivers and lorry drivers have been known to be able to accurately detail routes around the country. Is this something you are referring to? $\endgroup$ Jun 16, 2023 at 21:21

2 Answers 2

4
$\begingroup$

SHORT ANSWER

why wouldn't evolution have given us the quite useful ability to accurately build a map of the environment?

Because of the energetic cost of building and maintaining such a map! Some bird species (see "The genius of birds" by Jennifer Ackerman) build and memorize detailed maps of their environment, along with temporal information about where and when they hid which seeds, because it is key to their survival. Humans evolved more limited mapping and memorization abilities because they were nomads in a changing environment, hunting moving targets, and such skills were not as essential to their survival.

LONGER ANSWER

Ed Young, in his book "An immense world", describes the concept of Umwelt (from the German word for "environment"), defined as the "part of an animal's surroundings that an animal can sense and experience -- its perceptual world". He gives various examples of senses (e.g. "magnetic field" of robins and sea turtles, etc.) and of sense organs (e.g. "There are animals with eyes on their genitals"), I recommend the reading in general if you are interested in bio mimetism and robotics.

More particularly, Young mentions that "an Umwelt cannot expand indefinitely", because "Senses always come at a cost":

Animals have to keep the neurons of their sensory systems in a perpetual state of readiness so that they can fire when necessary. This is tiring work, like drawing a bow and holding it in place so that when the moment comes, an arrow can be shot. Even when your eyelids are closed, your visual system is a monumental drain on your reserves. For that reason, no animal can sense everything well.

The same argument applies to the organization and memorization of the information given by such senses: it does not come for free, and each species evolve to organize and memorize exactly the information that its individuals need to survive. My guess is that in the future, robots with stringer access to energy (such as nano robots, or even "normal" robots in a world less dependent on fossil energies?) will go through a similar selection process and draw less accurate maps, with a level of precision more tuned to the task they are programmed for.

I hope it helps! (and, from a fellow Computer Scientist interested in Psychology, a warm welcome to the Psychology Stack Exchange!)

$\endgroup$
4
$\begingroup$

Short answer
The brain is quite able to build internal spatial maps of the environment.

Background
If you've ever walked alongside a blind person in an environment they are familiar with, you will have noticed how capable the brain is of generating internal spatial maps. As an anecdote: I just started a postdoc at Hopkins Hospital, working with folks with end stage retinitis pigmentosa; they were legally blind. I was asked to bring one of them across the street to one of the other buildings to continue their testing. I had no clue, no clue as to where we actually were, where the exit of the building was we were in, let alone how to get to the designated location in the other building. The blind person noticed my anxiety, and being a long time patient there, they took my arm and requested permission to guide me back. They brought us back, across corridors, down to the basement, taking elevators and what not. And they were fast I'll tell you, happily chit chatting along the way.

The reason normally sighted folks don't have such detailed mental maps available is, likely, simply because they don't need them that much; visual cues and landmarks are 'computationally' less demanding and do the job well in most sighted people.

Here are some references that provide less anecdotal backgrounds to the mental spatial mapping observed in blind folks:

References
- Dodds et al., J Vis Impairment & Blindness (1982); 76(1)
- Hersch et al, ACM Transactions on Accessible Computing (2020); 13(2:6); 1–32

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.