Great job! Your video really showed the benefits of the FSR’s! And I like the ATOM tower on the top
I’m thinking of placing the FSR’s on a tube like this but I want to make a incision on one side where I can slice in the end of the tibia. Another cool way should be adding a additional joint like Matt did on his hexapod v4. This way it is possible to get the cool effect where he lift the front legs with his hands in the video.
Don’t worry about the forward kinematics. I’ve started with that a long time ago so I’ve got them on my pc. I’ll send you a copy when I’m home 8)
I’ve build a test setup with a BB2 and SSC so I can easy reach them without having the need of opening up one of my bots! I’ve started with some changes from floating point to fixed point calculations like Acidtech mentioned. I’ll think that I’ve got my first results by the end of the week.
It may be that you just need to install a half sphere rubber bumper like I did on this kit.
That should translate the angle pressure to the FSR. In my kit the rubber end cap slides on to completely cover the FSR and rubber bumper. It is adjusted to where it’s just not touching the rubber bumper. It has a great dynamic response, the full amount that the sensor provides. But it’s not easy to fit this to the end of the Phoenix tibia. But I think it can be done. Another thing about the sensors I used is they are adhesive backed, but the surface it is applied to must be very flat.
I’ll see if I can make something better than this current test setup. I’ve also been thinking of making a tars like you mentioned Xan, similar to Matt’s V4. Actually, I’ve an idea but i think I’ll use that for my next hexapod project…
If you have ever tried to walk on uneven terrain with your eyes closed youl’ll know thats really hard. So only having a touch sensor under the foot is somewhat the same, it’s hard to anticipate and calculate when the foot are going to touch the floor. Maybe having a little IR short range proximity sensor on each leg to measure the distance to the floor/terrain in addition to the FSR? Or maybe a vision system? I’ve no idea how to implement that though…
Anyway, I’m looking forward to work with the terrain walking part 8) . A little slow walking but I think it will work. We just have to find a good working sensor mount solution.
Lots of insects are able to walk on uneven terrain with little to no vision capability. I don’t see the need for additional sensors. Just need to make the best of the FSR, or even just a simple switch on the end of the tibia.
If you are using the ssc-32 as the servo interface, have you tried querying the servo position after giving it a “stop” command to stop a timed movement? If a position value is returned, then you have your servo position. If that doesn’t work to obtain the servo position, then the pot in the servo can be easily tapped and an analog servo position value obtained there. You may want to also look at a simple switch setup for the foot sensor if an actual analog pressure value is not needed.
watched your latest video you made with the “do do do” music in the background… anyway…
whats the deal with the balance modes? is there a gyro in it or something?
1 ) For simplicity I think that the 6 step tripod gate appears to be the least amount of effort to code? please correct me on this if I am incorrect.
As for your experimentation is there anything you would do now to dumb up your code? “not a insult by any means” but any way you can think of to simplify your executions of commands so it can be executed faster?
at the end of your sweet little you tube video youtube.com/watch?v=72MY5GuHSpA I really think that the individual control of the leg with the Futaba remote is very good! I read earlier in your post that your using an 8 channel transmitter, very cool with those extra 4 switches to do something change modes whatever… anyway my idea is this, looks like the front left leg was the one that was able to be controlled by the remote, you ever think of adding some option to select which leg is the one able to be controlled by the transmitter? perhaps when you select a leg to be controlled it simply taps the floor so you know that’s the active one, and then when you press a “switch/do some action” the other leg next to it taps the floor so you can control that one also? <— This would be very cool!
on a bit of an interesting topic, I was wondering which gate method you have found to be physically the fastest gait method getting mr or ms spidy from one end of the floor to the other? “taking a guess with the 6 step tripod gate?” ehh ehh do i win do i win?
my method of control will be the futaba 4 channel transmitter, I have an 8 channel receiver and if its tempting enough I might have to pick up an 8 channel one because well you will most likely make me with all the cool features you will have to include and all that jazz…
The tripod gait might look easier. Actually the code are almost the same for all type of gaits (take a look at Xan’s code). It’s more a matter of different sequence/timing.
Xan are working with a more effective code using fixed point instead of float variables.
I’m using a 7 channel Futaba transmitter. And I use only 6 of the channels. I’ve modified the transmitter by connecting 7 of the switches in a resistor ladder connected to a analog channel (ch6). And I’m using a tri-state switch on ch5 (mode selection). This gives me 3 modes and 7 functions on each mode.
A little schematic of how I connected the ladder:
The tapping function sounds great, but I’ve already implemented individual control of each leg By using one mode and single leg movement and using 6 of the function switches for each leg.
The tripod gait are always the fastest. Xan’s latest 4 step tripod is the fastest!
That’s neat, I guess each time you add 150 ohms to the switch the output signal for that PWM channel translates to a different static number you can read off your receiver, pretty slick! too bad they don’t make a 12 position switch !!
Either way I hope to have my project built pretty soon so I can test this stuff,
how is your receiver connected ? just a bunch of output channels dumped into the bot board?
have any good coding references for atom pro code? besides just diving into it head first?
Your code is a headache “not that that’s a bad thing” but its like you use all these advanced functions to make your robot move is all that calculation really neccessary?? i mean cant you dumb up your code a bit seems like your stuff is calculated to the uber-teenth decimal place and all your really positioning is 18 servo’s with a small range of motion, it seems like all this can be done with simple ratio’s …“or some other form”. 'I could be wrong in saying all this… but imagine what could be done with a faster processor!
I’m particularly a KISS person “keep it simple stupid”
Although how you do all those body rotations is a bit cool… who knows my batteries should be here sometime until then i play the waiting game :\
LOL, We just putted them in for fun! Just in case the BAP get bored
No serious, we’re controlling 18 servos which need to work all together. The servos are axial but you want to hex to move in a strait line. IK calculations are needed to do that. I think it can be done with less calculations but I’m sure the hex doesn’t move as smooth at all. Just search on youtube for other hexapods and see the difference…
Xan said it very well, and its mainly his code. And to be honest I don’t think the code are going to be easier to understand in the future either, since Xan are working with the fixed point calculation to make the code faster.
Any of you guys have any reference material for studying this code? or did the Atom Pro software have all that listed with descriptions sorta like how visual basic was setup?
Main
High P1
Pause 200
Low P1
Pause 200
Goto main
END
when I hit write to the atom pro, how long is it suppose to take to program this chip? seems like its taking for flipping ever “not sure what the dealio is”
The best way to learn is by trying. Believe me, I know it’s a complex material so what you need to do is split it in to small pieces. But what is your goal? Do you want to learn IK, or programming, or do you want to add something to the code, or…
I think it’s a good idea to learn how the board works. Just read a button and let it turn off and on a LED. Make it switch between the 3 LED’s. Try to move a servo. Etc.
Start controlling one leg with only 2 servos once you know the BAP basics. Try to move it in a straight line. You’ll need IK to do that and you’ll get the IK basics. The IK sheet in Zenta’s PEP can help you out with the calculations. You will bump in all kinds of issues because Excel is far more advanced then the BAP.
I’m sure it will start to make sense when you make progress. I’ve spend a lot of free time in this code the last year. So it would be crazy to think you’ll figure it all out in a couple hours
I’m a smart cookie, but I think Im probably going to start reading the manual and testing with some simple code and see where it gets me, the end goal is make an RC spider do some neat things 4 channels “might invest in an 8 channel one”
I actually got kind of interested in the cockroach walking need to read up on that, but apparently most all these insects use triangle gait or some form of it, but would like to know why cockroaches run so damn fast would be interesting to try and apply this to this spider. “which I am guessing you have all ready done?”
As for IK I will take your suggestion and try and simply learn 1 leg and do it by hand/ti-89 and see where that get’s me take a look at code and go from there…
Respectfully I would really like to learn it on my own without starting with something really complicated I just have a gut feeling that this code doesn’t have to be complicated or “as complicated” as it appears… at least that’s my goal
So I wrote a test program here with the ATOM-Pro IDE software and it seems like its taking a awful long time to program this chip with such a simple peice of code I wrote… anyone have any suggestions?
I’ve started with some changes from floating point to fixed point calculations like Acidtech mentioned. I’ll think that I’ve got my first results by the end of the week.
Or maybe a vision system? I’ve no idea how to implement that though…
LOL, We just putted them in for fun! Just in case the BAP get bored 
I’ve just passed 500 posts and still I’m “only” a roboteer! 
. I probaly have to wait for post 
