0
$\begingroup$

It is my understanding that there are 3 fundamental control methods : velocity, position and torque.

In exoskeletons meant to augment an healthy user's strength, a typical control method used is sensitivity amplification control, more specifically torque amplification. I am currently selecting a motor controller and trying to build a control framework for an exoskeleton that would use such a system.

According to the research that I have done, an user applied force would be detected by a sensor and the movement would then be assisted by an actuator according to this relation : (amplification ratio * detected force). In this scenario, is it fair to assume that velocity and position control are not needed?

Improve this question
$\endgroup$
5
  • 1
    $\begingroup$ is it fair to assume that velocity and position control are not needed at all? ... making assumptions can lead to disaster ... without position control, the exoskeleton may try to bend the user body parts into positions that could cause injury $\endgroup$
    – jsotola
    Commented May 29, 2023 at 1:09
  • $\begingroup$ @jsotola But would safety be the only reason for incorporating position control? $\endgroup$
    – Kristof
    Commented May 29, 2023 at 19:12
  • $\begingroup$ The human body? Physical constraints would be used at the joints and since it is only a prototype, I wouldn't have to worry about a suicidal user repeatedly hitting those constraints in order to break them. $\endgroup$
    – Kristof
    Commented May 29, 2023 at 21:06
  • $\begingroup$ @jsotola - if you've got an answer, please post it as an answer below. Comments are not intended for extended discussions, they are for helping to improve questions and answers. Comments are distracting, so we try to keep them to a minimum. They should be considered ephemeral, and any comment which no longer actively helps to improve a question or answer may be deleted at any time to tidy up a post. For extended discussion on a question or answer, please use Robotics Chat $\endgroup$
    – Chuck ♦
    Commented May 30, 2023 at 8:50
  • $\begingroup$ This is a moderately open ended subjective question so I'll do my best to answer it. We prefer practical, answerable questions based on actual problems that you face. Take a look at How to Ask and tour for more information on how stack exchange works, and the Robotics question checklist for details of how to write a good question. $\endgroup$
    – Tully ♦
    Commented May 31, 2023 at 7:31

1 Answer 1

Reset to default
0
$\begingroup$

To start off I think that it's best to clarify that position, velocity, and torque control are not necessarily fundamental types of control, but common types of control mechanisms.

There's already a lot of good resources related to that such as Position Control vs Velocity Control vs Torque Control

A control algorithm's purpose is to bring the state of a system to it's goal state. If you measure the torque and control the torque only, any errors in the system will accumulate and the system's kinematics will diverge to the point of being useless. This if more of an amplifier than a control method. Typically if you're controlling something such as an exoskeleton and you want force/torque controls, they will be operating at the same time as the position controls or as an integrated control algorithm. This will make sure that the state of the controlled system continuously match that of the remote operation control.

In the case of an exoskeleton it would be possible to rely on the fact that the two systems are collocated and simply apply torque amplification of the measured arm torques to the analogous joints. This clearly has some safety concerns raised in the comments. Before embarking on controlling the whole or much of the exoskeleton I'd suggest that you start by testing out different methods on single link joints on a test bench.

Improve this answer
$\endgroup$
2
  • $\begingroup$ Thanks for the answer. You talk about using position control to match the user's intended state with the state of the exoskeleton. How would the intended trajectory be calculated? $\endgroup$
    – Kristof
    Commented May 31, 2023 at 15:33
  • $\begingroup$ The goal of the control system is to converge and track as closely as possible to the intended state. While operational in general the control system can be made to track close enough that it's not necessary to compute a trajectory to get to said intended state, instead it just stays close enough that it's just removing any tracking errors only. $\endgroup$
    – Tully ♦
    Commented May 31, 2023 at 19:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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