Motion Generation for Robots and Machines
The core Ruckig algorithm calculates a trajectory to a target state (with position, velocity, and acceleration) starting from any initial state limited by velocity, acceleration, and jerk constraints. This calculation is done within the real-time control loop of the robot or machine, enabling it to instantaneously react to sensor input and replan its motion accordingly. Moreover, Ruckig allows for obstacle environments or intermediate waypoints to calculate complex trajectories on-the-fly, even surpassing offline approaches for motion planning in terms of trajectory duration. Ruckig is the first trajectory generator for arbitrary target states and supports directional velocity and acceleration limits, minimum trajectory durations, and many more features.
RuckigCommunity Version |
RuckigPro Version |
|
|---|---|---|
| Recommended for | Research Education Prototyping | Commercial products Industrial deployments |
|
Arbitrary initial and target states
Start and target are each given by position, velocity, and acceleration. Non-zero target velocities let a robot move through a state without stopping, e.g. for blending or handovers. |
||
|
Jerk limitation
Limits on jerk in addition to velocity and acceleration produce smooth motions. This reduces vibrations, mechanical wear, and tracking errors, and lets machines run closer to their limits. |
||
|
Synchronization of multiple axes
Works with any number of axes, from a single motor to multi-robot systems. All axes reach the target at the same time, or follow a straight line through phase synchronization. |
||
|
Velocity control interface
Switch any axis from position to velocity control at runtime. Use it to bring the system to a safe, jerk-limited stop, or for jogging and manual guidance. |
||
|
Hard real-time capability
Bounded worst-case calculation times for guaranteed performance in time-critical control loops. |
– | |
|
Increased reliability
Numerically robust and extensively tested for the most demanding production environments. |
– | |
|
Position limitation
Set minimum and maximum positions per axis. The trajectory stays within joint limits or the allowed workspace. |
– | |
|
Intermediate waypoints
Pass through a sequence of waypoints, e.g. a collision-free path from a motion planner, in a single time-optimized trajectory. |
– | |
|
Tracking
Follow a continuously moving target signal with minimal lag while respecting all kinematic limits. Ideal for conveyor tracking, teleoperation, and commands from VLA models. |
– | |
|
Support & integration
Direct contact with the developers, help with integration into your control architecture, and long-term support. |
– | |
|
Codebase
Pro also compiles with older standards, for legacy controllers, embedded toolchains, and microcontrollers. |
≥ C++20 ≥ Python 3.10 | ≥ C++03 ≥ Python 3.8 |
| Software license | Open Source (MIT) | Individual 30-day trial |
| GitHub | Request Evaluation | |
| Documentation | ||
The Community Version is the final state of the original research project. It is no longer actively developed and has known open issues. For commercial products and industrial applications, we strongly recommend Ruckig Pro, which is continuously maintained, tested, and supported.
Ruckig powers pick-and-place, welding, assembly, bin-picking, and jogging tasks with time-optimal trajectories that reduce cycle times.
From cobots to humanoids, Ruckig provides safe and smooth motion execution for VLA (Vision-Language-Action) models and teleoperation systems.
When robots must react to vision input in real-time, Ruckig bridges the gap between perception and actuation, e.g. for camera-based control, conveyor tracking, or collision avoidance.
Ruckig optimizes toolpath execution with time-optimal trajectories while respecting machine constraints for improved surface finish and reduced cycle times.
From ground vehicles to flight control, Ruckig's online trajectory generation enables instantaneous replanning when obstacles appear or targets move.
Even on resource-constrained hardware and microcontrollers, Ruckig controls motor drives with precise velocity, acceleration, and jerk limits.
Ruckig integrates into IEC 62304-compliant software architectures for medical robots, with full lifecycle documentation, traceability, and test coverage.
For laser cutting, welding, and marking, Ruckig generates precise scanning trajectories with control rates exceeding 50kHz for exact energy deposition.
Ruckig already controls container cranes, wired cameras in sports arenas, and instruments for minimally invasive surgeries. What about yours?
Ruckig is developed by Griffig Robotics GmbH, a company founded and led by Lars Berscheid. The algorithm originates from his research at the Karlsruhe Institute of Technology (KIT), where he developed the first method for computing time-optimal, jerk-limited motions to arbitrary target states, published as "Jerk-limited real-time trajectory generation for arbitrary target states" at Robotics: Science and Systems (RSS) 2021. He subsequently continued his research at the University of California, Berkeley.
Since publication, the Ruckig Pro version has been integrated into production systems across multiple industries worldwide. Griffig Robotics provides consulting, integration, and long-term support for professional deployments.
We're happy to answer your questions and discuss how Ruckig can help you.
Send us a short note about your application, and we'll set you up with an evaluation license.
or write us an email at hello@ruckig.com