Skip to content

Analytical IK¤

Closed-form inverse kinematics for known industrial robots. Runs entirely in-process -no Docker, no PyBullet, no ROS- and is the fastest IK option in compas_fab.

When to use¤

  • Prototyping reachability and inverse kinematics on a supported robot.
  • IK from inside Rhino, Grasshopper or any other Python host.
  • Generating IK seeds for a downstream planner.
  • You don't need collision checking. (If you do, see Analytical IK + PyBullet.)

Trade-offs¤

What you get What you give up
Closed-form IK in microseconds Only the supported analytical robot families
Zero setup beyond pip install compas_fab No collision checking
Works from any Python host (CAD, headless, CI) No motion planning, IK / FK only

Setup¤

Nothing to install beyond compas_fab itself. Analytical solvers are part of the package.

Supported robots¤

  • Universal Robots: UR3, UR3e, UR5, UR5e, UR10, UR10e, UR16e
  • Staubli: TX2-60L

  • Abb

    IRB 4600 40/255

See compas_fab.backends for the full list of solver classes.

First example¤

Forward kinematics on a UR5:

from compas_robots import Configuration

from compas_fab.backends import AnalyticalKinematicsPlanner
from compas_fab.backends import UR5Kinematics
from compas_fab.robots import RobotCellLibrary
from compas_fab.robots import TargetMode

# Not loading the robot's geometry because AnalyticalKinematicsPlanner does not use it for collision checking
robot_cell, robot_cell_state = RobotCellLibrary.ur5(load_geometry=False)

# The kinematics_solver must match the robot's kinematics
planner = AnalyticalKinematicsPlanner(UR5Kinematics())

# This is a simple robot cell with only the robot
planner.set_robot_cell(robot_cell)

# Configuration for FK calculation
configuration = Configuration.from_revolute_values([0.0, 4.8, 1.5, 1.1, 1.9, 3.1])
robot_cell_state.robot_configuration = configuration

# AnalyticalKinematicsPlanner.forward_kinematics(), do not support `planning_group` parameter, it must be left as None.
# The `link` option is also not supported and cannot be used.
frame_WCF = planner.forward_kinematics(robot_cell_state, TargetMode.ROBOT)

print("Robot flange frame of the default planning group in the world coordinate system:")
print(frame_WCF)

More examples¤

For collision-aware IK and Cartesian planning, see Analytical IK + PyBullet.

API reference¤