MATLAB Answers

# Inverse Kinematic Solver solutions

조회 수: 25(최근 30일)
Kyle Lammers 5 Sep 2019
댓글: Cam Salzberger 13 Nov 2020
Hello, I am using the robotic toolbox inverse kinematic solver as shown below.
My main question is why do I get different results everytime this script is ran? There are small differneces in the outputs but then every few iterations the values seem to almost double.
Thank you for any advice you can offer!
clear
robot = importrobot('appleRobot.urdf');
homeConfig = robot.homeConfiguration;
ik = robotics.InverseKinematics('RigidBodyTree', robot);
ik.SolverParameters.AllowRandomRestarts = true;
ik.SolverParameters.MaxIterations = 1500;
x = 24;
y = 25;
z = 26;
%inches to meters
x = x / 39.37; y = y / 39.37; z = z / 39.37;
pos = [x y z];
pose = eye(4);
pose(1:3, 4) = pos';
weights = [0.25 0.25 0.25 1 1 1];
[configSoln,solnInfo] = ik('end', pose, weights,...
homeConfig);
dis = 0; theta = 0; phi = 0;
dis = configSoln(1).JointPosition;
theta = configSoln(2).JointPosition;
phi = configSoln(3).JointPosition;
[dis theta phi];
%solnInfo

댓글을 달려면 로그인하십시오.

### 채택된 답변

Cam Salzberger 6 Sep 2019
Hello Kyle,
The inverse kinematics solver parameter AllowRandomRestarts is likely the culprit here. If you check the documentation:
AllowRandomRestarts — Indicator if random restarts are allowed. Random restarts are triggered when the algorithm approaches a solution that does not satisfy the constraints. A randomly generated initial guess is used. MaxIteration and MaxTime are still obeyed. By default, random restarts are enabled.
Since there is some randomness in the solver process, the results may differ every time. Especially if there are multiple solutions for the same position.
-Cam
##### 댓글 수: 3표시숨기기 이전 댓글 수: 2
Cam Salzberger 13 Nov 2020
Mosen, if you check out the examples attached to the inverseKinematics documentation, this example shows precisely how to do this.
ik = inverseKinematics('RigidBodyTree', pr2);
ik.SolverParameters.AllowRandomRestart = false;

댓글을 달려면 로그인하십시오.

### Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by