Robot Questions,
Robot Questions,
by Devtron
Sun Jul 30, 2006 12:40 am
Its been around 10+ years sins i last played with servos and conecting stuff to a computer in the same way as you all do it with robots today.
So im more or less starting from step one again.
For the past 6 months i have been looking more and more into robotics again, they have come fare sins the last time i played with it.
What im looking for in a robot, if this is posibal to build is the folowing.
1 It can Run, Turn, bend, do head stands, pick up a cold beer:), pick it self up when it has fallen.
2 it can walk around avoding objects, tell what rooms its in, and if there is movement detect and respond to it, be set to find a object ( within reason ofcource ) detect if it is near an edge of something ( dont want it to fall down from a table )
3 Voice activation and control.
4 have a camara of some sorts, so you can remote view from say your computer.
This might still not be posible at the Tech level the world is at today, but i know some of these things are posible as toys can do some of it.
So im more or less starting from step one again.
For the past 6 months i have been looking more and more into robotics again, they have come fare sins the last time i played with it.
What im looking for in a robot, if this is posibal to build is the folowing.
1 It can Run, Turn, bend, do head stands, pick up a cold beer:), pick it self up when it has fallen.
2 it can walk around avoding objects, tell what rooms its in, and if there is movement detect and respond to it, be set to find a object ( within reason ofcource ) detect if it is near an edge of something ( dont want it to fall down from a table )
3 Voice activation and control.
4 have a camara of some sorts, so you can remote view from say your computer.
This might still not be posible at the Tech level the world is at today, but i know some of these things are posible as toys can do some of it.
by hivemind
Sun Jul 30, 2006 12:49 am
http://www.tribotix.com/Products/Roboti ... _info1.htm
or... were you looking for a reasonable price, not $20,000? because you could buy a robonova, 17 dof stock (i think? - mine was never stock) and give it an ultrasonic sensor and a little programing, a rotating head and your probably set to go. you could add a wireless camera similar to the CycloidII if you wanted, the head is mostly empty space, so make your own around a camera and then your good to go.
Of course if you have an extra 20k just lying around...
or... were you looking for a reasonable price, not $20,000? because you could buy a robonova, 17 dof stock (i think? - mine was never stock) and give it an ultrasonic sensor and a little programing, a rotating head and your probably set to go. you could add a wireless camera similar to the CycloidII if you wanted, the head is mostly empty space, so make your own around a camera and then your good to go.
Of course if you have an extra 20k just lying around...
by Devtron
Sun Jul 30, 2006 12:59 am
I know that computers in the small size for these robots are not powerfull enouth for a lot of these tasks, but i once build a remote control car that folowed me around, sure it hit a wall or 10 
, now and again, but it could avoide them most of the time and folow me around inside as well as ourside,
what i had been planing on doing is get some kind of robot, that can move around more or less like a human ( or just build an R2D2 style one)
that will also folow me around, and respond to questions.
i know that the sony dog that has been discontinued could read emails, fetch a ball and such, so it should not be to complicated to use software already avalible to create a free roaming robot, i can do a bit of c++ programming, but this may still be WAY over my head to do.
I might just be setting to high a goal for what i would like a robot to do
but using some kind of kit robot combined with a powerfull computer as the brain ( wireless interface to the robot ) something like this must be posible, within a resonable price
, now and again, but it could avoide them most of the time and folow me around inside as well as ourside,
what i had been planing on doing is get some kind of robot, that can move around more or less like a human ( or just build an R2D2 style one)
that will also folow me around, and respond to questions.
i know that the sony dog that has been discontinued could read emails, fetch a ball and such, so it should not be to complicated to use software already avalible to create a free roaming robot, i can do a bit of c++ programming, but this may still be WAY over my head to do.
I might just be setting to high a goal for what i would like a robot to do
but using some kind of kit robot combined with a powerfull computer as the brain ( wireless interface to the robot ) something like this must be posible, within a resonable price
by robodude666
Sun Jul 30, 2006 6:44 am
Devtron,
What you are asking for is what tons of people are trying to do. Honda spent billions im guessing on their bot
To build a small humanoid robot such as the size of the robonova or khr series and be able to mount this stuff and the POWER supply for everything will be too hard. the robot will mostlikly be unbalanced.
1) The robonova can pretty much do all that. with a gyro or accelerometer it can be able to scense if it has falling on its back or front and stand back up. with the gyro it can also keep itself from falling down it if gets nudged. as for beer, it would have to be on the same table as the robot is in. the robonova can pretty much do that too with the right programming and grippers(bauermech has some nice ones.)
2)avoid objects, mount a few IR sensors and youl be good. youl need a lot of code to be able to figure out what needs to be done next. especially if you have more than 1 or 2 sensors. tell what room its in, not an easy task. the easiest way to do that is to have an IR emitter around each room with their own freqency and the nova will pick it up and know which room it is in. visually, not gunna happen. you can try popping on a camera and making it take tons of pictures but it wont help. the robobasic language sucks for sensors and stuff that arent like a gyro or a ir sensor. you will need to get a new board and even then it will be hard. objects get moved around in a room, youl also need some crazy image processing stuff to get a best match for picture.
find an object? you should look at POB. they made a camera that can detect closed shapes. amd detect if near an edge? IR sensor would work.
3) um, no comment on that... gunna need to pop out some real $ to get a good system.
4)
you can easily get a wireless camera and mount it to the chest or head of the robot and make it send a wireless signal to your TV (or computer if you have tv tuner card) to view whats happening.
How much you willing to fork out for this robot? I did a lot of research in these fields and know some good parts you can get.
EDIT:
#2 might be possible with that AIBO dog from sony. it has a pretty good processor i hear and it has a camera and can do image processing. But the dog costs about 3,000 and parts for moding it will cost around another 1000 or so i would guess. do some googling on development for AIBO.
-robodude666
What you are asking for is what tons of people are trying to do. Honda spent billions im guessing on their bot
To build a small humanoid robot such as the size of the robonova or khr series and be able to mount this stuff and the POWER supply for everything will be too hard. the robot will mostlikly be unbalanced.
1) The robonova can pretty much do all that. with a gyro or accelerometer it can be able to scense if it has falling on its back or front and stand back up. with the gyro it can also keep itself from falling down it if gets nudged. as for beer, it would have to be on the same table as the robot is in. the robonova can pretty much do that too with the right programming and grippers(bauermech has some nice ones.)
2)avoid objects, mount a few IR sensors and youl be good. youl need a lot of code to be able to figure out what needs to be done next. especially if you have more than 1 or 2 sensors. tell what room its in, not an easy task. the easiest way to do that is to have an IR emitter around each room with their own freqency and the nova will pick it up and know which room it is in. visually, not gunna happen. you can try popping on a camera and making it take tons of pictures but it wont help. the robobasic language sucks for sensors and stuff that arent like a gyro or a ir sensor. you will need to get a new board and even then it will be hard. objects get moved around in a room, youl also need some crazy image processing stuff to get a best match for picture.
find an object? you should look at POB. they made a camera that can detect closed shapes. amd detect if near an edge? IR sensor would work.
3) um, no comment on that... gunna need to pop out some real $ to get a good system.
4)
you can easily get a wireless camera and mount it to the chest or head of the robot and make it send a wireless signal to your TV (or computer if you have tv tuner card) to view whats happening.
How much you willing to fork out for this robot? I did a lot of research in these fields and know some good parts you can get.
EDIT:
#2 might be possible with that AIBO dog from sony. it has a pretty good processor i hear and it has a camera and can do image processing. But the dog costs about 3,000 and parts for moding it will cost around another 1000 or so i would guess. do some googling on development for AIBO.
-robodude666
by Robo1
Sun Jul 30, 2006 12:36 pm
I would agree with the two pevious reply.
I've just got a khr-2 and want to achive every thing in your list. It just depends on how much time/money/experiance.
I came from a programming background so the programing isn't to hard to get a rough robot doing what you want but for me the electronics part is my down fall. I can normally get the right components but getting to work and getting the circut board small so that It fits on the robot is the hard part.
I recon my robot once built will cost around £3000 and that's not including the £2000 I want to spend up grading all the servos. And take about a year to build. But I think I have a small advantage as I'm just about to start a postgraduate masters in robots at bristol UNI.
If you want to have fun and do this as a hobby just get a RN or khr and just play around. this technology has only been around for 4-6 years so very young, so if you start playing around now in a couple of years time when this thing get more advanced you will be able to use them better.
Bren[/i]
I've just got a khr-2 and want to achive every thing in your list. It just depends on how much time/money/experiance.
I came from a programming background so the programing isn't to hard to get a rough robot doing what you want but for me the electronics part is my down fall. I can normally get the right components but getting to work and getting the circut board small so that It fits on the robot is the hard part.
I recon my robot once built will cost around £3000 and that's not including the £2000 I want to spend up grading all the servos. And take about a year to build. But I think I have a small advantage as I'm just about to start a postgraduate masters in robots at bristol UNI.
If you want to have fun and do this as a hobby just get a RN or khr and just play around. this technology has only been around for 4-6 years so very young, so if you start playing around now in a couple of years time when this thing get more advanced you will be able to use them better.
Bren[/i]
New Parts R on the way ...
by CaptKill4Fun
Sun Jul 30, 2006 5:43 pm
...
Last edited by CaptKill4Fun on Wed Dec 13, 2006 6:21 pm, edited 2 times in total.
Playing w/Dolls ...
by CaptKill4Fun
Sun Jul 30, 2006 7:53 pm
...
Last edited by CaptKill4Fun on Wed Dec 13, 2006 6:21 pm, edited 2 times in total.
Re: Playing w/Dolls ...
by Devtron
Sun Jul 30, 2006 8:40 pm
CaptKill4Fun wrote:Devtron wrote:The Robots we have today, are they more than toy dolls, that you just program to do a flip, and then that is it ?
...Hi,
Yeah, they are much more than toy dolls ... they are a learning experience ...
You learn how to macro program the bots movements ...
You learn how to program using BASIC or C etc ...
If you get really ambitious like Bauer, hivemind, myself and many others ... you start to imagine new parts and how to use existing parts to make the bot even more functional ...
DARPA has challanges for college kids to come up with solutions to very complex autonomous robotic problems ... everything we're doing here is helping taking those steps closer to solving those challanges and many more challanges of our future ...
Wouldn't it be nice to have an overall handy robot around the house? A robot that is a security guard, it picks up your crap and puts it away, it cleans the house, it helps prevent other disasters while you're not home like putting out a fire or at lest detecting the fire and caring for your pets while alerting the local fire department ...
e.g., Putting stuff away ... say you're late for work, you leave the milk and ceral on the (removed by spam filter) table ... your personal bot puts those two things away where they belong ... the cabinet automatically counts how many times you've taken the ceral out and weighs the box to determine when to order the next box ... the refrigerator does the same and knows the expiration date of your foods so it keeps fresh foods available while recycling the old foods all automatically for you ...
There's myriad challanges for all of us to meet these and many more challanges for our future ... and so playing with robotics is the learning solutions ...
What you just explained is fare still fare into the future, i dont forsee anything close to that for at least 20-30 years,
Atm as to what i can see from all the videos i have seen, you have to preprogram everything, it dose not move on its own, or do anything other than its preprogramed taskes.
Re: Playing w/Dolls ...
by CaptKill4Fun
Sun Jul 30, 2006 9:07 pm
...
Last edited by CaptKill4Fun on Wed Dec 13, 2006 6:22 pm, edited 2 times in total.
Re: Playing w/Dolls ...
by Devtron
Mon Jul 31, 2006 11:35 am
CaptKill4Fun wrote:
That's the point!
As we play; we design, build and program and we get new ideas that take us closer to the future ...
If we don't play today, we can't have the tools for tomorrow!
As much as i can locate information on robotis, it seams that microsoft ( yeah the evil guys) are the first with a universal program, to control them with, now i see that as a first step in the right direction.
as for the current state of the robot world, it seams to almost stand still, i hope that some of the toys that will come out in the remainder of this decade will change that. but at the moment, everything seams to stand still
Re: Playing w/Dolls
by sippy
Wed Dec 13, 2006 6:12 pm
Devtron you appear to be judging the entire world of robotics based upon the current state of the "toys".
Saying that the current state of robotics is standing still is amazingly naive. In academia and R&D, where most of robotic developments occur there have been massive developments. The sony QRIO can achieve almost everything you stated and is the first robot to be able to run, DARPA have recently succeeded in their first challenge and are now moving on to bigger and better challenges. The standard of competition at events such as Robocup and FIRA have increased massively in only a few years.
Compared to allot of robotics yes the Robonova is relatively a toy(sorry guys). However its is very sophisticated and capable for £700ish, and with modification is capable of much more. More importantly it will ultimately pave the way for even better robots.
Don’t believe that just because you cannot buy Terminator off the shelf for £700 that robotics is failing.
Might I suggest a subscription to Robotics and Automation(http://www.ieee-ras.org/)
Saying that the current state of robotics is standing still is amazingly naive. In academia and R&D, where most of robotic developments occur there have been massive developments. The sony QRIO can achieve almost everything you stated and is the first robot to be able to run, DARPA have recently succeeded in their first challenge and are now moving on to bigger and better challenges. The standard of competition at events such as Robocup and FIRA have increased massively in only a few years.
Compared to allot of robotics yes the Robonova is relatively a toy(sorry guys). However its is very sophisticated and capable for £700ish, and with modification is capable of much more. More importantly it will ultimately pave the way for even better robots.
Don’t believe that just because you cannot buy Terminator off the shelf for £700 that robotics is failing.
Might I suggest a subscription to Robotics and Automation(http://www.ieee-ras.org/)
