1
$\begingroup$

Can anyone give me a typical frequency that a driverless vehicle could generate steering and motor thrust commands?

I am trying to model a driverless vehicle in MatLab. Right now the vehicle is generating commands at 50Hz and the integrator is solving the dynamics at a 0.01s timestep (100Hz). I just want to know if this is realistic.

I would be even more grateful if you could point to a published source for this number.

$\endgroup$
  • $\begingroup$ This depends a lot on the dynamics of the system (vehicle) itself. Many of the actuators are not capable of responding to high frequency demand inputs. The main reason controllers run higher frequencies is to cut down the delay from sensor to actuator. The plant model should include the actuator dynamics as true as possible, but the integral frequency has almost nothing to do with the controller frequency. Right? $\endgroup$ – Gürkan Çetin Mar 27 '17 at 9:27
  • 1
    $\begingroup$ To get a better answer, you should elaborate on what question you are trying to answer with your simulation. If you don't have a specific set of problems and assumptions in mind, your simulation is guaranteed to give useless results. Unless your simulation is for entertainment purposes. For the current generation of vehicles, command frequency is driven more by sensor latency than the mechanical time constant of the actuators, and multiple command frequencies are used on a sub-system and sub-control-loop specific basis. $\endgroup$ – hauptmech Apr 25 '17 at 4:37
1
$\begingroup$

Tom Sheridan from MIT used to say (I am paraphrasing from 25 years ago so I am sure to have some of this wrong) that robots which operate only in free space should update around 10 Hz; if they interact with environmental impedances then they should update around 100 Hz; if they have to go from free space to hard contact with a stiff environment, then at least 1 kHz. Based on that I would recommend starting at 100 Hz and adjusting from there, depending on the additional dynamics of the system.

$\endgroup$
0
$\begingroup$

MPC lab from Berkeley uses about 50-100 Hz. Although, I think it might be closer to 500Hz-1kHz for the low-level control (motor speeds etc.) implementations (this is close to the rates at which you control quadrotors as well) - I will post links here if I come across any publications that use a faster update rate.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.