BlackWido

I’m glad you like it! It was a good idea to use the other sticks as well :slight_smile:

This should be doable but it is a lot of work. But it needs a lot of precise controlling to get it to work. We’re already working on real terrain adaption with force sensors which is taking up all my time right now but with this we can walk over obstacles by only moving the left stick forwards! Sounds fun doesn’t it :wink:

But don’t let me hold you back, feel free to give it a try and adapt my code yourself!

Xan

BTW: You can edit a previous post by pressing the edit button on the top of your post :wink:

Hi Joe,

I’m not sure if I understood your suggestion completely. But it seems to be a good idea. This sound similar to the “oneleg” balancing I demonstrated in my balance video (after 1:53), but also combined with walking gait? Like a pose shift as I demonstrated in this video? But if you also want to incorporate terrain adaption we do need sensors on the leg as Xan mentioned.

I believe walking over obstacles above 15 cm is a bit tricky with Phoenix, simply because of the length of the tibia and femur.

Hi zenta;

Yes is somethink like that…

Im allways try to walk in diferents terrains, grass, rocks…
But with phoenix code, Im not having good results because it dont lift the legs too more… I have to work in it…

youtube.com/v/LQJqFnFrPTk

Firsr I think we need an another tripod gait who lifts more the legs and have high body position…
Im trying the powerpod code… with another atom basic tha I have… and in bad terrain go well, so I have to start (and learn… this is the best way to learn something^^) programing new gait to move more like a tripod… but lifting more the legs…

to your phoenix one leg control is impresionant… yes is some like that… but read carefull Im going to explain it best I can:

Yes, sensors are goog idea to automatize terrain adaptation… but it can be controled by you…
If we press R1 and on rigth stick one direction, The deepness that we press that direction determines how high the leg have to be possed… so the directions of stick(8 dir, separe their zones in 6 because they are 6 legs). At the same time we can mome forwand and back… and the legs must be move one by one.
For example 15cm upstair I go forwand and when Im near tho upstairs while Im going forwand, I will press R1 and softly the rigth stick in direction of the leg tha we want move, so that leg go up and pose more high than normal because we press that direction softly. After the other leg, other, other, and other…
With this we can have 3D control but is theory… Is hard to program and hard to control…
and I think same, the best way automatize this, is with sensors like Xans says and is more inteligent…

thaks for thinking on it :wink:

Did someone ever dive into the “step up” mode?

I mean to have the hexapod lift his front legs and then walk forward on the 4 other legs so he can step up a higher ground.

Thanx in advance.

Not that I’m aware of.

Should be something fun to try but I think it whould be a sequence. It should be hard to realize doing the calculations on the fly. Maybe once the terrain adaption is fully up and running. Who knows :wink:

Will i need sensors on the feet for this?
By the way, if i get some microswitches, how would i attach them? to the ssc-32 or the botboard and wich pins?

Terrain adaption is still a work in progress. And yes, you need some really reliable switches/setup for this. I’m currently working on the Hybrid that uses some micro switches. It’s discussed in this threat. The first post contains a link to the switch setup.

I’ve also tried attaching FSR’s to the phoenix legs. But these don’t give me the wanted hysteresis and reliability. I doubt if this will ever work with FSR’s. Here’s the link to the FSR post.

Keep in mind that I’m not saying this will work with terrain adaption but it’s worth a try once terrain adaption is fully working. One step at the time. :wink:

Xan

Well i made a plan for the script and started writing. Ive done about 1/3 of the code.

I will keep you posted!
This is what i have until now.

[code]IF (DualShock(2).bit6 = 0) and LastButton(1).bit6 THEN ;Cross Button test
for Index = 0 to 31
serout SSC_OUT,SSC_BAUTE,"#",DEC Index,“P1500”]
next
serout SSC_OUT,SSC_BAUTE,“T200”,13] ;default is T200

	pause 288

serout ssc_out,SSC_BAUTE,"#",dec RMCOXAPIN,“P1500#”,dec RMTIBIAPIN,“P1500#”,dec LMFEMURPIN,“P1500#”, |
dec LMTIBIAPIN,“P1500#T288”,13]

			;Step up Posture

serout ssc_out,SSC_BAUTE,"#",dec RMCOXAPIN,“P1200#”,dec RMFEMURPIN,“P1600#”,dec RMTIBIAPIN,“P1500#”, |
dec LMCOXAPIN,“P1800#”,dec LMFEMURPIN,“P1400#”,dec LMTIBIAPIN,“P1500#T2304”,13]

		pause 288

serout ssc_out,SSC_BAUTE,"#",dec RFFemurPin,“P900#”,dec RFTibiaPin,“P1200#”,dec RFCoxapin,"P,dec LFFemurPin,“P2000#”, |
dec LFTibiaPin,“P1600#T2304”,13]

		pause 288

serout ssc_out,SSC_BAUTE,"#",dec RRFEMURPIN,“P1700#”,dec RRTIBIAPIN,“P1600#”,dec RMFEMURPIN,“P1800#”, |
dec RMTIBIAPIN,“P1600#”,dec RFCOXAPIN,“P1200#”,dec RFFEMURPIN,“P2100#”,dec LRFEMURPIN,“P1300#”, |
dec LRTIBIAPIN,“P1400#”,dec LMFEMURPIN,“P1200#”,dec LMTIBIAPIN,“P1400#”,dec LFCOXAPIN,“P1800#”, |
dec LFFEMURPIN,“P900#T2304”,13]

			pause 144

serout ssc_out,SSC_BAUTE,"#",dec RRFemurPin,“P1300#”,dec RRCoxapin,“P1300#”, |
dec LRFemurpin,“P1700#”,dec LRCoxaPin,“P1700#”, |
dec LFFemurpin,“P1000#”,dec RFFemurPin,“P2000#”,dec LFTibiapin,“P1400#”,dec RFTibiapin,“P1600#”, |
dec LFTibiapin,“P1400#T2304”,13]

			Pause 288
			
		serout ssc_out,SSC_BAUTE,"#",dec RRCoxaPin,"P1300#",dec LMCoxaPin,"P1700#",dec LRCoxaPin,"P1700#",dec RMCoxapin,"P1300#", |
			[/code]

Hi Janssen,

I’ve just been looking over your Black Wido project again for the umteenth time, and I’m not clear on the TLDivFactor. The tutorial says “number of movements on the ground (red arrows)”, and indeed for example 1, there are 4 arrows = TLDivFactor. Example 2 has TLDivFactor = 8, and I again see four red arrows.

Example 1 shows steps 2 through 6 along a base line, example 2 shows steps 3 through 11 on a base line. OK, maybe not all the little circles and arrows (and a paragraph on the back of each) are shown in the drawing? That would work.

I see ripple and tripod gates, but what I’m wanting for my 'quad is a creep/crawl gait. I think it would be done this way?

IF (GaitType = 8) THEN ;Creep Gait 4 steps
	LRGaitLegNr = 3  
	RFGaitLegNr = 1
	RRGaitLegNr = 2
	LFGaitLegNr = 4
	  
	NrLiftedPos = 1
	HalfLiftHeigth = FALSE	
	TLDivFactor = 2	  
	StepsInGait = 4
    NomGaitSpeed = 150
  ENDIF


IF (GaitType = 9) THEN ;Creep Gait 8 steps
	LRGaitLegNr = 5  
	RFGaitLegNr = 1
	RRGaitLegNr = 3
	LFGaitLegNr = 7
	  
	NrLiftedPos = 1
	HalfLiftHeigth = FALSE	
	TLDivFactor = 6
	StepsInGait = 8	    
    NomGaitSpeed = 100
  ENDIF

Do these look at all correct? I hope to be trying them soon!

EDIT: Jonny, seems my TLD’s (and I don’t know what else) were wrong!

I’d also like to try (from Peter’s web site):

RR-LR-LF-RF
RR-LR-RF-LF
RR-LF-LR-RF
RR-RF-LR-LF

The last one seems the most reasonable!

Nice gait engine!

Have you considered adding a “COG Shift” parameter? I’m thinking it could signify when to shift COG to left or right to help with the balance of quadruped robots. the shift would be cyclic, and thus have a left and right maxim step.

Or maybe some clues from KÃ¥re!

Alan KM6VV