Purchase & Select Accessories?

[treborz]

The Robonova-1 manual discusses tilt snd gyro sensors. Where can I piuchase the recommended sensors in USA(to simplify shipping)? Also, with gyros taking up 2 ports each, of 8 available, how many gyiros does it take to add a reasonal amount of additional stability to the robot? Since I am interested in autonomous operation I would like to save a couple of ports for the use of other sensors.

Would appreciate any comments.

Thank you,

Robert

[JavaRN]

If you want my feedback on this, my robonova has no gyros, like you I preferred using the ports for sensors rather than for gyros, in fact my robot has 4 IR sensors, a sonar sensor (MaxSonar), and a tilt sensor, and many other things attached to it, surprisingly enough, he rarely falls, and whenever he does there is the tilt sensor which signals him to wake up!

May be I am using RN for tasks which are different from yours, (I use it for educational research not for fighting or acrobatics), but in my opinion and in my application gyros are not needed (even though I can understand that they DO improve the robot's stability)

Hope this gives you some help.

[treborz]

Thank you for your thoughts JavaRN. I have completed a few Parrallex Basic Stamp projects and would like to apply the sensor knowledge that I picked up on the RN. However, since I use gyros on a large scale RC Electric powered helicopter with interesting effects, I would at least like to use one port pair as a stability gyro on the RN, that is if one gyro would even be noticed. Would it? It might come in handy when demonstrating my RN on a carpet.

One gyro controlling the tail rotor on my helicopter is awesome. Placed on a reasonably smooth surface you can manually move the tail of the idling helicopter to any part of a circle and the fuselage snaps right back to its original position-- useful when your hands are full with the other controls. The gyro allows turns only by RC rudder control and not wind gusts. I too have very little interest in RN competition; I am mainly interested in autonomous navigation. BTW, I am retired and spending my kids inheritance on Robot projects.

postscriptum: Does the tilt sensor tell you whether the RN is on its back or front?

[davidalecmcinnes]

I read with interest this topic as my main interest with humanoid robotics is in autonomy, making him independent and interactive. I evolved onto Robonova from lego bots, which had the tendency to fall to pieces every time they hit something.

I have just ordered a gyro to stabilise his walking a bit as his balanced has been affected with the mods I have done, (sensors, pan and tilt, gripper, 3 axis accelerometer, pir). My experience is that the Robonova design is very finely balanced to achieve such acrobatic capability and mods tend to upset it a little.

I used the accelerometer to figure out which way he had fallen over so he could get back up, but at the moment his head is too heavy for getting off his face.

I'd be greatly interested to know any thoughts or goals you have for autonomy.

[bauermech]

My experience is that the Robonova design is very finely balanced

This is true (w/o any structural mods). The RN-1 does just fine on flat, hard surfaces w/o the need for gyros.

I would at least like to use one port pair as a stability gyro on the RN, that is if one gyro would even be noticed. Would it? It might come in handy when demonstrating my RN on a carpet.

Why anybody would need four gyros is beyond me. I use only one - for forward/backward (push/pull) compensation. A second one could be installed for right and left sway as well. You can assign as many servos as you'd like to each gyro... set their individual direction and response too.

The problem with carpet... Any deviation from the foot maintaining an evenly distributed surface tension, the edges of the foot will sink into the nap. This makes it easier for him fall or get snagged on the fibers. A gyro really won't help TOO much in preventing this.

Does the tilt sensor tell you whether the RN is on its back or front?

Yes

[treborz]

bauermech, you mentioned that you can assign as many servos as you'd like to each gyro... set their individual direction and response too. That is exciting! Does that mean that one gyro could control servos that handle both forward/backward and left/right sway? Also, could these servos be operating at the same time?

Thanks for the tilt sensor info, I just found, in another sample program, the Basic code for handling the face up or face down selection.

[limor]

to get answers for the "which" gyros/accellerometers question, please check out the RoboNova wiki

[bauermech]

Does that mean that one gyro could control servos that handle both forward/backward and left/right sway?

You'd need two gyros... one for forward/back, and the other for right/left - this is all dependant on the orientation of the gyro installed on your RN-1 (the direction the gyro is facing determines the robots reaction in that direction). You CAN however turn on a certain set of servos to compensate for a push/pull while standing fully erect, and an entirely different set of servos to compensate for the push/pull while he is in a squatting position (for example, you wouldn't want the knees to attempt to compensate while squatting, only while standing, but by leaving the hip servos responsive to the gyro will still aid in keeping him from being knocked over). Setting the direction in code will make the servo either follow with, or repel against the force being applied.

Also, could these servos be operating at the same time?

With two gyros (forward/back and right/left) yes, you can have both running at the same time... and have some of the same servos affected by both.

Thanks for the tilt sensor info, I just found, in another sample program, the Basic code for handling the face up or face down selection.

No problem. It's been a while since i visited this forum, but I know there's some code examples in the Robonova WIKI for handling gyros (about 2/3 of the way down)... as well as some other helpful stuff.

[bauermech]

to get answers for the "which" gyros/accellerometers question, please check out the RoboNova wiki

Ha! Limor, looks like we're on the same page

[plingboot]

just my 2penneth, but i'd pitch up for the RN tilt sensor - as this does a nice job when he falls (you can set it up to know whether he's on his back or front). I'd also go for the max sonar sensor (hitec chest kit) with which you can do some fun stuff - see dirtyrobotos great thread on this.

As for gyros - i've used them on both my 'bots and they do offer counterbalance actions or some impressive dance moves as he shakes every screw loose when you use too much gain - but you need to be aware that in some of your motions you'll have to switch them off and back on otherwise they'll actively try hinder the motion.

as for autonomy - i wouldn't get too excited. knowing he's fallen and reacting to sonar input is as close to autonomy as it gets - unless you are an absolute genius with robobasic AND have all day to fiddle around with it.

I'd also find pennies of for the rn bluetooth kit - the IR remote can be pretty frustrating IMHO

There's ship loads of fun to be had twonking around with what can be done, but i wouldn't expect to have him walking around the house opening and closing doors behind him and turning the tv on without asking...

[JavaRN]

as for autonomy - i wouldn't get too excited. knowing he's fallen and reacting to sonar input is as close to autonomy as it gets - unless you are an absolute genius with robobasic AND have all day to fiddle around with it.

Well, here I think we have to clarify what we intend by autonomous. In my opinion autonomous means that he can do assigned tasks without human intervention not that he is completely independent.

In this area there are loads of things to experiment with, it's not just sensors. Put a wireless camera on a pan and tilt and you will have a whole new world of experiments - colour blobs, image recognition, OCR, path finding etc, add a bluetooth module connect it to a computer and then you can do loads of other staff.

In my opinion if you have to do something serious with our RNs then we have to use more intelligent devices - in my case I find that interfacing with my laptop would at least remove the barrier of processing speed and storage space.

Charles

[treborz]

Thanks for all of the replies, I really needed your help to get started. I wish I had known about the WiKi before I bothered everyone with my newbie questions.

BTW, what does WiKi stand for? (another newbie question)

Thanks again!

[tempusmaster]

:) Thanks for all of the replies, I really needed your help to get started. I wish I had known about the WiKi before I bothered everyone with my newbie questions.

BTW, what does WiKi stand for? (another newbie question)

Thanks again!

http://en.wikipedia.org/wiki/Wiki

[davidalecmcinnes]

I was going to start a new thread but the topics kind of exploding in this one.

Well, here I think we have to clarify what we intend by autonomous. In my opinion autonomous means that he can do assigned tasks without human intervention not that he is completely independent.

I have done a little work on autonomy and for me it was more to do with the robot making its own decisions based on its rules and goals rather than completing pre-ordained tasks, in effect completely independent. Robotics is destined to be a big part of our lives in the future and autonomy in that light is one of the keys to realising that.

Autonomy though is unpredictable and difficult due to the real world sensing, modelling and interaction problems and isn’t helped by the astounding number of things we take for granted due to our own human sensors. There are a couple of schools of thought in robotics in relation to autonomy; the behaviour based response and the sense-plan-act (SPA) paradigm.

Robobasic does not allow a realtime response of Robonova to its environment without OS modification or in depth programming. This limits a behaviour based approach which reduces its ability to operate and interact with an uncontrolled and rapidly changing environment in a robust manner.

The sense-plan-act approach would be a more suitable approach for the basic model. A battery of sensors can gather data which can be processed by a number of rules upon which a decision can be made and then a move executed.

Emergent behaviours occur more readily in behaviour based robots (BBR) but the interaction of the sensors and the environment will also lead to some interesting responses.

Simple object avoidance while roaming and recovery from accidents is a basic level of autonomy that should be possible. More rules could then be implemented to govern behaviours such as environmental (time, light, temperature) and personal (boredom, tiredness, hunger) which could also contribute to its behaviour.

As more technology emerges, hierarchical architectures could also be implemented consisting of, for example, a reactive layer, a thinking layer and goal based layer; which is more or less a fusion of the two existing paradigms in robotics (BBR and SPA). Facial recognition, voice recognition and speech could also be implemented with a more powerful core.

With the basic Robonova 1 and the existing technology today, it is possible to achieve a level at which the little chap is independent and has his own quirky character.