Hello,
I'd like to understand some basic KUKA concepts in terms of approximation for both LIN-LIN and PTP-PTP cases.
First let's assume we have declared 4 points (XP1, XP2, XP3 and XP4), and our example looks like this:
INI
$APO.CDIS = 20.0
LIN XP1 C_DIS
LIN XP2 C_DIS
LIN XP3 C_DIS
LIN XP4 ; stop here
I have a lot of questions even for this very simple case:
- Does the $ADVANCE value 1, 2 or 3 influence the resulting path?
- If we set different $VEL.CP speeds for each LIN command, the resulting path would remain the same when using distance-based smoothing?
- Does the robot accelerate with $ACC.CP during the blending phase?
- What if the segment speeds are very different, thus the acceleration is not feasible?
- If the Carthesian distance of XP2 and XP3 is less than 10 mm, will P3 be completely skipped and XP4 would be used after XP2 or just the approximation radius gets reduced symmetrically for both the XP1-XP2 and XP2-XP3 segments?
- What if XP3 lies near XP1 while XP2 is "far away" (thus the normal path without smoothing would be a very sharp triangle): as mathematically no parabolic blend is possible without extreme accelerations, would the robot ignore approximation and stop at XP2 or stop beforehand and turn around to XP3 or even report an error already at XP1?
-----------------------------------
My second path of the topic addresses the PTP-PTP smoothing. As far as I understand the C_PTP mode specifies to start approximation whenever the angular distance to the point is less than some percentage of the 100% angle defined in $APO_DIS_PTP[x] for each robot axis (default: 90°).
Let's look at the following case with 3 AXIS points XQ1, XQ2 and XQ3.
INI
$APO.CPTP = 25.0
PTP XQ1 C_PTP
PTP XQ2 C_PTP
PTP XQ3 ; stop here
- With this example of 25% the robot would start to turn towards XQ3 if all axes are less than 90*0.25=22.5° angular distance to XQ2, right?
- Is the trajectory during the PTP approximation near XQ2 also a normal synchronous PTP trajectory or a parabolic curve like in the LIN-LIN case?
- Would the smoothing end at the "middle" of the XQ2-XQ3 trajectory or simply go directly to the XQ3 point?
- Is there also a rule of "maximum half distance can be approximated" for the PTP-PTP segments?
- If so, would this situation be reduced to the "smaller" of the XQ1-XQ2 and XQ2-XQ3 distances in axis space?
Thank you for the clarification...