4

I'm involved in research on psychologically plausible models of reinforcement learning, and as such I thought it'd be nice to try and see how well some to the models out there perform in the real world (i.e. sensory-motor learning on a mobile robot). This is already been done in some robotics labs, such Sutton's implementation of the Horde Architecture on the "Critterbot". However, these implementations involve robots custom-build by robotics experts in order to deal with the trials and tribulations of learning on a long time-scale:

"The robot has been designed to withstand the rigors of reinforcement learning experiments; it can drive into walls for hours without damage or burning out its motors, it can dock autonomously with its charging station, and it can run continuously for twelve hours without recharging."

Unfortunately I'm no expert when it comes to designing robots, and don't have access to a high quality machine shop even if I did; I'm stuck with whatever I can buy off-the-self or assemble by hand. Are these constraints common enough for amateur robotics suppliers to cater to, or should I expect to have to start from scratch?

zergylord
  • 143
  • 6
  • 1
    This question can't be answered in a useful way. Any answer describing the state of the art of robotics, will get out-dated in a couple of months, a year tops. It's also very general. – Shahbaz Jun 11 '13 at 08:45
  • Just a pointer to a line of particularity durable and longlasting off-the-self or kit robots would be fine. If a particular company/supplier caters to such needs, then the answer wouldn't be out-dated rapidly. – zergylord Jun 11 '13 at 10:46
  • Welcome to robotics zergylord, but I'm afraid that there is no practical, answerable question based on actual problems that you face here. Take a look at [ask] and [about] and if you can edit your question to make it answerable then I would be happy to reopen it. – Mark Booth Jun 11 '13 at 11:23
  • How is this too localized? Am I the only person concerned with actuator reliably and the battery life of robots? I read all of those links; which part actually applies to me? – zergylord Jun 12 '13 at 03:28
  • http://robotics.stackexchange.com/q/554/1460 So asking about battery life in general is too localized and quickly outdated, but asking about battery life constraints on a particular robot model is widely applicable and long lasting knowledge? I'm deeply confused about what the guidelines for this S.E. are... – zergylord Jun 12 '13 at 03:31
  • Nevermind, found the meta question you guys seem to be talking about: http://meta.robotics.stackexchange.com/q/91/1460. Any chance of reopening if I get rid of "state of the art" in the title? – zergylord Jun 12 '13 at 03:39
  • @zergylord, now the question is a bit better. Regarding your one-before-last comment, the way you ask the question is something like: "What battery in the market should I use?" which is a search-for-me-and-let-me-know type of question. If you really want to know about how battery life constraints work with respect to how complex the robot is, you could have asked something along the lines of: "Bla bla bla, the more secure, reliable and durable the more battery power I need. How do I decide what is the optimal tradeoff?", and you would get an answer like "It depends on your application" – Shahbaz Jun 12 '13 at 08:23
  • Anyway, try to be more specific in your question and sure it would get reopened if it becomes a good question. – Shahbaz Jun 12 '13 at 08:24
  • 1
    Sorry zergylord, I should have linked to that meta post when I made my pre closure comment. I will reopen this for now so we can see if anyone can answer it as it stands, but I will have to be sensitive to future close votes - at the moment it is very close to a NARQ (not a real question). It would be better if you asked a question about a real problem you face along with what you have tried so far to solve it rather than dealing with a hypothetical situation. – Mark Booth Jun 12 '13 at 11:55
  • It's not hypothetical, but I see your point about being close to a NARQ. The question is basically that I don't know where to start: e.g. should I be reading on energy efficient actuators and plan to build my bot piece by piece or would modifying a preexisting model, like a roomba, result in better sensor/actuator reliability and battery life? – zergylord Jun 13 '13 at 20:00
  • I think I see the question you're trying to ask. You want a robot that's durable enough to withstand long time-scale testing, but you don't know what to look for. Fill in the blanks for us: "What features does a robot need to be able to do _____ for _______ – Ian Jun 27 '13 at 02:10
  • This question is salvageable, but must be edited. Currently it is a combination of a shopping question ("which amateur robotics suppliers can fill my need?") and lacks a definition of a practical problem (a durable, reliable robot is desired, but no information is given on minimum specifications). – Ian Jun 27 '13 at 14:29

2 Answers2

2

In short, yes, there are a number of robotics companies catering to just your needs. Specifically which apply to you depends on what type of robot you desire. Naturally cost varies with the complexity and durability. For example if you would be satisfied with a differential drive robot for indoor use then the iRobot Create may suit your needs. If you need something for outside use then you could look at Clearpath Robotics products. For a full humanoid robot the Aldeberan Nao is a good choice. For quadrotors check out the AR.Drone Parrot. And I understand Liquid Robotics offers a good selection of water based robots.

DaemonMaker
  • 3,921
  • 3
  • 21
  • 32
  • 1
    THanks! Just what I was looking for. Though I've worked with Naos before and I'll vouch for them not being durable lol – zergylord Jun 27 '13 at 06:20
  • I would be interested in hearing more about what your doing because my interests also lie in RL. – DaemonMaker Jun 27 '13 at 15:36
0

Off-the-shelf robot kits are a good place to start for an amateur. If you're not using system-based kits (i.e. Lego), then they can usually be modified. Mount bumpers around the sides to help it withstand crashes and add a larger battery with a higher milli-amp-hour (mAh) rating for longer battery life. It's tough to provide any more specific solutions than that.

Crunchy
  • 161
  • 3