Blind and Visually Impaired Community
↑
1
↓
Is there a need for robotic guide dogs? (self.Blind)
submitted by secret_asian_man22
My last post contained a survey on guide dog technology. Since, I have looked through the survey/question sticky on the front page and read about this community's annoyances when answering questions. Yes, I am a research student, and I'm sorry for contributing to the desert of students asking barely thought-out questions.
Hopefully, the second go-around is better, so here's my revised question. Do you believe there is a necessity to create robotic guide dogs that are cheaper or last longer than traditional guide dogs? Why or why not?
If this revised question is still annoying (and I wouldn't be surprised, I'm still learning about the etiquette so bear with me) just let me know and I'll just delete this post entirely. Thanks!
Hopefully, the second go-around is better, so here's my revised question. Do you believe there is a necessity to create robotic guide dogs that are cheaper or last longer than traditional guide dogs? Why or why not?
If this revised question is still annoying (and I wouldn't be surprised, I'm still learning about the etiquette so bear with me) just let me know and I'll just delete this post entirely. Thanks!
MostlyBlindGamer 6 points 2y ago
A guide dog can do dogs things and guide things. The best robotic dog (Boston Dynamics' Spot) can't do all the dog things yet.
Are you going to do better?
The scope of a project like that is immense, isn't it?
If you'll allow me to be crudely pragmatic, dogs eat cheap enough food and are cheap enough to replace. A robot would have to be significantly better or significantly cheaper.
Are you going to do better?
The scope of a project like that is immense, isn't it?
If you'll allow me to be crudely pragmatic, dogs eat cheap enough food and are cheap enough to replace. A robot would have to be significantly better or significantly cheaper.
secret_asian_man22 [OP] 1 points 2y ago
Definitely too immense for research students, which is why this isn't something we're seriously prototyping.
As far as I could find, the average guide dog costs $59,000 over its lifespan of around nine years, most of that being training. That's more than a car, which is probably a suitable baseline for the estimated cost of a robotic guide dog, if even.
How many dog things are needed to act as a sufficient guide? Is there a great need for, say, going up sets of stairs if accessibility ramps and elevators are widespread?
As far as I could find, the average guide dog costs $59,000 over its lifespan of around nine years, most of that being training. That's more than a car, which is probably a suitable baseline for the estimated cost of a robotic guide dog, if even.
How many dog things are needed to act as a sufficient guide? Is there a great need for, say, going up sets of stairs if accessibility ramps and elevators are widespread?
MostlyBlindGamer 3 points 2y ago
Spot costs 74k. It also needs to be recharged and maintained. Can you make it cheaper *and* better than that?
How many dog things? All of them. Let me go ahead and get a luxury cars' worth of robot and just make sure there aren't stairs anywhere I need to go. You gotta be kidding me.
How many dog things? All of them. Let me go ahead and get a luxury cars' worth of robot and just make sure there aren't stairs anywhere I need to go. You gotta be kidding me.
secret_asian_man22 [OP] 1 points 2y ago
Does it need to look like a dog though? Spot is so advanced because it’s mimicking an actual dog. Solely for the purposes of guiding, are there more dog things than all terrain navigation (including stairs) that a robot replacement would have to mimic? Besides, recharged and maintained applies to dogs too, even a decent battery would last longer than a dog without rest.
And, just to clarify again, I’m not making anything, certainly not cheaper and better than Boston dynamics
And, just to clarify again, I’m not making anything, certainly not cheaper and better than Boston dynamics
MostlyBlindGamer 3 points 2y ago
What are you looking for them, exactly?
Another user already mentioned sensors and AI instead of the dog form factor.
The biggest thing you get out of an animal is actually its intelligence. A dog won't cross the street with you if there's traffic.
Waymo's cars will actually slow down for traffic humans couldn't see, because they have IR cameras. There's a lot of interesting work out there in autonomous vehicles.
Another user already mentioned sensors and AI instead of the dog form factor.
The biggest thing you get out of an animal is actually its intelligence. A dog won't cross the street with you if there's traffic.
Waymo's cars will actually slow down for traffic humans couldn't see, because they have IR cameras. There's a lot of interesting work out there in autonomous vehicles.
AchooCashew 3 points 2y ago
Guide dogs may be expensive, but I can't imagine any adequately safe, durable, sensitive, accurate, precise, teachable, customizable, useful technology that could ever be cheaper than a dog.
Iamheno 3 points 2y ago
Here’s my opinion, it depends.
in my case it might. I do endurance/ultra-endurance events. I could see where a robotic assistant with adequate battery could be useful. I’m prepping for a 36 hour event in September. My guide dog won’t be helpful during it I could see using one of the Boston Dynamics “military mule” type robots with appropriate sensors etc. but I’ll be sticking with good ‘ole reliable seeing eye humans.
For an average person with blindness, probably not.
in my case it might. I do endurance/ultra-endurance events. I could see where a robotic assistant with adequate battery could be useful. I’m prepping for a 36 hour event in September. My guide dog won’t be helpful during it I could see using one of the Boston Dynamics “military mule” type robots with appropriate sensors etc. but I’ll be sticking with good ‘ole reliable seeing eye humans.
For an average person with blindness, probably not.
codeplaysleep 3 points 2y ago
If you mean a literal robotic dog, no that would be silly. If you mean a complex set of sensors/feedback devices and AI to build a proper, safe, responsive, and actually useful suite of navigation tools, sure that'd be nice in theory. However, basically nothing created thus far comes close to being as useful as your typical white cane and existing voice navigation on Google Maps.
DrillInstructorJan 2 points 2y ago
I think if you had the technology to make a robot guide dog you could probably just put it into a bit of technology you could wear on your hat or whatever. You would start hitting the same sort of tech challenges that electric cars have. I think adding the robot dog chassis to the issue is just making it a lot more complicated and expensive and hard to store. Don't build a robot dog if you can fit the same guidance system into a phone app.
bradley22 2 points 2y ago
I would use something that could gide me around and wouldn’t care what it looked like.
EvilChocolateCookie 1 points 2y ago
Sure. The robot wouldn't get you know what all over the place.
Apprehensive_Art3339 1 points 2y ago
I think the better question is asking if we need robotic guides to replace guide dogs. I don’t think the “dog” aspect is important for creating a robot. There are such things as guide miniature horses, though they aren’t very popular.
My question is, can a robot be created that can perform the functions a guide dog provides? Maybe—we have self driving cars, maybe that type of technology could work for guide robots? But there is so much more than just navigation and impact avoidance for guide work. The dog has to find a door, find the stairs. find the elevator, find a chair, find the counter (I.e. in a store or doctor’s office). It has to guide around objects, signal uneven ground or unexpected stairs, and other random occurrences. It has to show intelligent disobedience and not follow commands if it would pot the user in danger (oncoming traffic, platform edges, construction, etc.). They have to be highly maneuverable and fit in small and sometimes oddly shaped places (airplane seat, under table at restaurant, etc.).
I just don’t think we have technology that could replace all the functionality of a guide dog. At least not now. But maybe we’ll never have that, or if we do, it is cost prohibitive.
Source: guide dog user
My question is, can a robot be created that can perform the functions a guide dog provides? Maybe—we have self driving cars, maybe that type of technology could work for guide robots? But there is so much more than just navigation and impact avoidance for guide work. The dog has to find a door, find the stairs. find the elevator, find a chair, find the counter (I.e. in a store or doctor’s office). It has to guide around objects, signal uneven ground or unexpected stairs, and other random occurrences. It has to show intelligent disobedience and not follow commands if it would pot the user in danger (oncoming traffic, platform edges, construction, etc.). They have to be highly maneuverable and fit in small and sometimes oddly shaped places (airplane seat, under table at restaurant, etc.).
I just don’t think we have technology that could replace all the functionality of a guide dog. At least not now. But maybe we’ll never have that, or if we do, it is cost prohibitive.
Source: guide dog user
This nonprofit website is run by volunteers.
Please contribute if you can. Thank you!
Our mission is to provide everyone with access to large-
scale community websites for the good of humanity.
Without ads, without tracking, without greed.
scale community websites for the good of humanity.
Without ads, without tracking, without greed.
©2023 HumbleCat Inc • HumbleCat is a 501(c)3 nonprofit based in Michigan, USA.