Open Problems in Robotics
scottlocklin.wordpress.com
Open Problems in Robotics
1–10 of 232 posts
Re: Open Problems in Robotics
#2> I’ll point out that the humble housefly has no problem understanding the concept of “shit in front of you; avoid,”
Surely in the case of the house fly, it would be more like 'dinner is served' than 'avoid'!
But in all seriousness, this is a good list to remind us how vastly far away from human-level AI we are.
Re: Open Problems in Robotics
#3I know that there's a lot of successful work in specific controlled environments (company X's factory floor), and that a lot of environment understanding/SLAM is broadly unsolved in arbitrary uncontrolled environments, but what about specific uncontrolled environments? What if i want a beer serving robot to learn a consistent, high-accuracy model of just my house, the way it is, and I'm willing to put in some technical work? Can i do any better than completely hand crafting a 3D object model or similar?
Re: Open Problems in Robotics
#4One problem I encountered while working on robotics is that many commonly used algorithms yield approximate solutions with no error bounds. They work 99.99‰ of the time. This is fine from a computer science or math point of view, but very scary from an engineering perspective, specifically when there are humans nearby. A big part of me struggles to accept the suitability of algorithms coming from gaming engines or machine learning etc for real world heavy duty robots. The lack of rigour in the field is astonishing.
Re: Open Problems in Robotics
#5Honestly, late yesterday evening after work I was looking at floor full of toys that my little kids were playing with in yet another lockdown day thinking I wish there was a robot that I could build or buy to tidy this up.
Did some research found this research project, promising but from 2018 and looks like it didn't go anywhere.
So now I'm thinking what about the build option.
Re: Open Problems in Robotics
#6In the area of Motion Planning (my own area of research), the most that can be said is that practical solutions exist for a tiny subset of cases, workable methods exist for a larger subset, expensive methods exist for a still larger subset, and everthing else might as well be impossible.
- If you've got a low-dimensional problem, say 2D or 3D, without uncertainty (or at least bounded enough to pad obstacles and ignore it), search-based planners like A* and its derivatives work. Add uncertainty, complex non-holonomic constraints, limited horizons, etc and it becomes much harder.
- If you've got a higher dimensional problem, say a 6/7-DoF arm, even multi-armed or humanoid robots, and you don't have uncertainty and dynamics (or can ignore them), sampling-based planners like RRTs and PRMs and their derivatives will often practically work. Actually useful guarantees of finding a solution or trying to find an optimal solution in useful time are still very much unsolved.
- If your problem is basically open, and something approximating the "straight-line" from A to B is in the same local minima as a solution, trajectory optimization will work for a lot of problems. Motion planning is very much non-convex, though, so it's very easy to go from a problem solvable with optimization to a problem that isn't.
- Planning with non-rigid objects, significant uncertainty (effectively continuous MDPs or POMDPs), and/or complex dynamics are all very unsolved problems.
For motion planning problems in the gray area of "practically solvable", the art is figuring out how to simplify the problem as much as possible to make it tractable - highly optimized collision checking (generally speaking the performance bottelneck), combining search/sampling-based + optimization methods to get an initial solution from a global method and then refining it towards a local minima with optimization, or using special hardware or sensors to bound dynamics and uncertainty so it can be ignored.
Re: Open Problems in Robotics
#7One problem I encountered while working on robotics is that many commonly used algorithms yield approximate solutions with no error bounds. They work 99.99‰ of the time. This is fine from a computer science or math point of view, but very scary from an engineering perspective, specifically when there are humans nearby. A big part of me struggles to accept the suitability of algorithms coming from gaming engines or ma…
This is the scariest part of using machine learning as an engineer on any practical application as well.
Without an error bound, ML can’t be in charge of anything that could put human lives at risk.
This is also why I don’t understand all the hype about FSD / L5 autonomous driving. We don’t even know yet if such error bounds even exist, so we don’t even know if machine learning is even the right tool for FSD yet. All certification entities for control systems that put human lives at risk in aviation, automotive, etc. require those error bounds. So it actually doesn’t really matter if Tesla comes up with a “maybe L5” system, without right error bounds, their cars won’t be certified as L5 and drivers will need to keep hands on the steering wheel.
Re: Open Problems in Robotics
#8One problem I encountered while working on robotics is that many commonly used algorithms yield approximate solutions with no error bounds. They work 99.99‰ of the time. This is fine from a computer science or math point of view, but very scary from an engineering perspective, specifically when there are humans nearby. A big part of me struggles to accept the suitability of algorithms coming from gaming engines or ma…
Though, probably humans fumble or bump things >1/10k times. I certainly do, anyways.
Re: Open Problems in Robotics
#9I know that there's a lot of successful work in specific controlled environments (company X's factory floor), and that a lot of environment understanding/SLAM is broadly unsolved in arbitrary uncontrolled environments, but what about specific uncontrolled environments? What if i want a beer serving robot to learn a consistent, high-accuracy model of just my house, the way it is, and I'm willing to put in some technic…
If you have high-quality 3D sensors (or in some cases, good-enough 2D cameras) and usable IMU/odometry, you can assemble quite reasonable 3D models for many small environments. The biggest challenges in a home environment are that many items we want in our houses (stainless steel appliances, chairs with thin legs, glass anything, reflective floor materials) are almost pathologically bad from a computer vision standpoint and are very hard to see and avoid in an uncontrolled household setting. Vision-based navigation against a known environment is doable with reasonable reliability, but adding fixed markers or positioning systems goes a long way towards robustness.
Re: Open Problems in Robotics
#10One problem I encountered while working on robotics is that many commonly used algorithms yield approximate solutions with no error bounds. They work 99.99‰ of the time. This is fine from a computer science or math point of view, but very scary from an engineering perspective, specifically when there are humans nearby. A big part of me struggles to accept the suitability of algorithms coming from gaming engines or ma…
This is the scariest part of using machine learning as an engineer on any practical application as well. Without an error bound, ML can’t be in charge of anything that could put human lives at risk. This is also why I don’t understand all the hype about FSD / L5 autonomous driving. We don’t even know yet if such error bounds even exist, so we don’t even know if machine learning is even the right tool for FSD yet. All…
What do you think the error bounds are for a human?
I know it sounds like a flippant question, but for certain applications, if we can get a model that's better than human, then it doesn't need to be perfect.
And they way we currently do this in all sorts of ways is to pair a human with a computer so that they each do what they're best at. It doesn't have to be about full automation.