3
votes
Machine Learning or Deep Learning libraries used in Robotics with FPGAs?
I cannot comment on 'most common', but I can definitely share several tools and research efforts towards using FPGA for deep-learning. See my survey paper on FPGA-based accelerators for CNN which ...
- 131
2
votes
Is there a more accurate feedback loop with less overshoot or disturbances than PID?
You might want to try a Non linear controller if you want disturbance rejection. But I will suggest you to fine tune your PID first.
Now, back to non linear controls, you need to develop a very good ...
- 163
2
votes
Where is the preffered position to put a simple control loop in a ROS2 Node (C++) for obstacle avoidance
I wouldn't overcomplicate it with what you're trying to do. Its extremely lightweight :-)
this isn't ideal especially if the code took a while to run it could cause sensor messages to be missed
You'...
- 6,264
2
votes
Accepted
Where is the preffered position to put a simple control loop in a ROS2 Node (C++) for obstacle avoidance
In general computations specific to the node are implemented as member functions. If it gets way to big you should think about using different threads / callback groups or split your tasks up into ...
- 116
2
votes
High derivative gain in PD control
Chuck's answer is spot on. Anyway, if you want to derive the reason mathematically, you can start off from the most common form of a PD controller where we employ a setpoint-weighting for the ...
- 4,005
1
vote
High derivative gain in PD control
Real signals have noise. Because noise happens on a per-sample basis, you wind up with a derivative that is constantly fluctuating. A derivative gain acts on this fluctuation and feeds it to the motor,...
- 15.9k
Only top scored, non community-wiki answers of a minimum length are eligible