Hey everyone!
I’d like to update you on everything we at PickNik have been improving with MoveIt open source and the MoveIt Studio Developer Platform lately (TLDR - a lot!).
MoveIt
Here’s some of the latest features available in MoveIt open source for ROS 2 Iron Irwini:
-
Parallel Planning: Speed up and improve the quality of your paths through parallel use of multiple motion planning algorithms.
-
Python Bindings: A new Python wrapper for MoveIt 2 is now available to make scripting and using Python libraries with MoveIt easier and faster.
-
STOMP Planner: MoveIt’s default motion planner, RRTConnect, quickly finds collision-free paths but can lack control over solution quality. This new STOMP planner offers a cost optimization approach, providing more flexibility and control in planning.
-
IK Cost Function API: Cost functions can now be specified to evaluate the fitness of IK solution candidates. Currently, the optimization-based IK solver plugins BioIK and pick_ik support this.
-
PickIK IK Solver: A MoveIt 2-compatible IK solver that improves bio_ik, offering custom cost functions for diverse objectives.
-
IK-based Servoing: MoveIt Servo now supports a plugin architecture for inverse kinematics, allowing users to integrate innovative algorithms for better robot performance, such as pick_ik.
MoveIt Studio
The MoveIt Studio Developer Platform continues to make robot arms smarter, providing an easy to use UI for MoveIt that fully integrates everything teams need to develop advanced robotic arm applications. MoveIt Studio combines Behavior Tree editing, teleoperation, debugging, introspection, and the full power of MoveIt motion planning.
Some of the new features in the latest 2.0 release include:
-
Easy Admittance Controller Tuning: You can now configure and visually edit admittance controller parameters to make force-torque interactions with a manipulator easier and your robots more gentle.
-
Out of the box Pick and Place: Our enhanced ML-powered object picking behaviors include automatic cuboid detection, automatic grasp generation, and dynamic replanning for objects that are adversarially moved.
-
More Visualizations: You can now visualize detected cuboids, generated grasps, and trajectory previews in the user interface to aid in debugging and robot understanding.
-
Teach Pendant-like Endpoint and Joint Jogging: We’ve improved our teleoperation by adding speed control, feedback when encountering singularities, joint limits, or collisions, and support for degrees and radians. We have also added a slider control with a setpoint feature that allows precise target positioning.
-
Improved Behavior Tree Editing: We have made significant improvements to behavior tree editing including the ability to copy/paste Behaviors, undo changes, and leverage subtrees.
-
Faster Simulation Capabilities: We have taken feedback from our customers about GPU issues, and integrated both Gazebo and non-Gazebo simulation options. This makes it easier to install and run MoveIt Studio on laptops and lower powered computers
-
Simplified Installation Process: We now have a less than 5 minute, one-line install process to get an example UR5e running in simulation.
Interested in trying MoveIt Studio? Request a demo today.
Let’s keep pushing the boundaries of robotics together!
Thanks,
Dave Coleman