Replacing revolute joint with a fixed joint prevents control of other joints

Question

Peanut 2022-6-10

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/1738085-replacing-revolute-joint-with-a-fixed-joint-prevents-control-of-other-joints

回答： Peanut 2022-6-29

I'm trying to reduce the 7DoF Kinova Gen 3 to a 6DoF model. To that end, I wanted to fix the HalfArm_Links.

robotMat = loadrobot("kinovaGen3");
halfArmTForm = robotMat.Bodies{1,3}.Joint.JointToParentTransform;
newJoint = rigidBodyJoint('FixedActuator3','fixed');
setFixedTransform(newJoint,halfArmTForm);
replaceJoint(robotMat,"HalfArm2_Link",newJoint);
show(robotMat,randomConfiguration(robotMat))

However, when doing this, and displaying randomConfigurations or jointConfigs as calculated by IK solvers, actuator4 seems to drive what was originally actuator5. Similarly, actuator 5 drives actuator6 etc. So actuator7 actually doesn't drive anything more, and the joint between HalfArm2_Link and ForeArm_Link appears fixed.

This issue seems to disappear when I change the joint type to 'prismatic'.

I'm using R2021a and have spotted this question, but I'm not seeing the same NaN issues.