<aside> ⚠️ This is an Experimental Feature, Please, carefully read the limitation section. You may use it for customer projects only if it works for your particular setup and requirements.
</aside>
https://www.youtube.com/watch?v=exQV2XrGFLA&ab_channel=NativeRobotics
(https://www.youtube.com/watch?v=exQV2XrGFLA&ab_channel=NativeRobotics)
We calculate trajectories by solving optimization equations. We consider all known obstacles, Joints, Accelerations, and Jerk limits altogether. As a result, we have a time-space spline.
The Trajectory
is a spline for a particular Task
.
The Task
has a set of parameters
All the parameters above form Task Signature
On every step (next box), we are calculating the current task signature and look for a stored trajectory with the same signature. If the signature is found - the trajectory will be executed. If not found - a regular trajectory will be executed and the task will be sent for planning.
Planning is going in parallel, so for unknown setup robot will start with regular trajectories. As soon as the optimal trajectory is calculated, it will be stored in the PC.