[sancelot]

I have the same kind of problem.

I know which robot configuration (F/N U/D T/B) I have to apply to my FANUC robot targets. I am unable to set it up with the robodk api.

Resulting in bad robot moving.

for reference
https://github.com/RoboDK/RoboDK-API/issues/56

[Albert]

I replied to the thread on GitHub with an example to pick a joint solution based on a desired configuration. I'm also pasting the example here:

Code:

# Retrieve all solutions for a given pose:
all_solutions = robot.SolveIK_All(pose, toolpose, framepose)
joints = []

# Iterate through each solution
for j in all_solutions:
# Retrieve flags as a list for each solution
conf_RLF = robot.JointsConfig(j).list()

# Breakdown of flags:
rear = conf_RLF[0] # 1 if Rear , 0 if Front
lower = conf_RLF[1] # 1 if Lower, 0 if Upper (elbow)
flip = conf_RLF[2] # 1 if Flip , 0 if Non flip (Flip is usually when Joint 5 is negative)

# Look for a solution with Front and Elbow up configuration
#if conf_RLF[0:2] == [0,0]:
if rear == 0 and lower == 0:
print("Solution found!")
joints = j
break

We'll be updating the documentation in the upcoming days to also show this example.

[SamuelEStreamline]

我有办法修改joint target from the API by moving the cartesian coordinates without knowing what the final robot joints will be?

[Albert]

Yes, you can recalculate the target joints by first making the target a Cartesian target, let RoboDK recalculate the joints and then change it to a Joint Target.

This example shows the concept:

Code:

target.setAsCartesianTarget()
target.setPose(new_pose)
target.setParam("Recalculate")
target.setAsJointTarget()
new_joints = target.Joints()

[SamuelEStreamline]

Yes, you can recalculate the target joints by first making the target a Cartesian target, let RoboDK recalculate the joints and then change it to a Joint Target.

This example shows the concept:

Code:

target.setAsCartesianTarget()
target.setPose(new_pose)
target.setParam("Recalculate")
target.setAsJointTarget()
new_joints = target.Joints()

This is wonderful! Thank you so much Albert!