IKINE

[lyhbuaa]

I am using the ikine() function to find the inverse solution of the manipulator, and I found that the qlim that set in the link didn't work，In the results, some of the joint Angle exceeds the limit. I don't know why and how to improve.

THANK YOU!

[yf291115925]

ikcon() % Numerical inverse kinematics with joint limits

[petercorke]

Please send a complete example

[HongyuanChen]

@petercorke maybe I can provide a complete example of panda.

Notice that I have removed the panda hand from panda, so I rewrite the mdl_panda.m with changes to this line (other code leaves the same):
panda_withoutTool = SerialLink([L1 L2 L3 L4 L5 L6 L7], 'name', 'PANDA', 'manufacturer', 'Franka-Emika');.

Then I record the current joint_pose and end-effector pose from franka_ros:

joint_pose:[0.17940711027697512, 0.4618104288114321, -0.07474237419220439, -1.3189208764310127, -0.01820848535001278, 1.8252406880905954, 0.8094893078808263]
cartesian_pose(in column-major format):[0.7681760305042956, -0.6365870159800344, 0.06828292058998993, 0.0, -0.6381994665792949, -0.7698674017593761, 0.002371616784093802, 0.0, 0.05105905420752297, -0.04539994266409048, -0.9976631787279365, 0.0, 0.6951395624527535, 0.081864847706614, 0.4775871098049651, 1.0]

When I use the following command, the forward kinematics returns right:

panda_withoutTool.fkine(joint_pose)
ans=  0.7682   -0.6382    0.0511    0.6951
   -0.6366   -0.7699   -0.0454   0.08186
    0.0683    0.0024   -0.9977    0.4776
         0         0         0         1

However, when I use the inverse kinematic function, the answer turns out to violate the joint limit:

panda_withoutTool.ikine(ans)
ans=0.1268    1.3965   -0.0078    0.3852   -0.0503    1.0569    0.7919

It can be seen that the joint 4 violate the joint limits which is defined in .qlim : [-3.0718,-0.0698]