Add support for velocity dependent effort limits in robot actuators.

[itsikelis]

In IsaacLab, the actuator model calculates torque limits based on a velocity based rule:

    def _clip_effort(self, effort: torch.Tensor) -> torch.Tensor:
        # save current joint vel
        self._joint_vel[:] = torch.clip(self._joint_vel, min=-self._vel_at_effort_lim, max=self._vel_at_effort_lim)
        # compute torque limits
        torque_speed_top = self._saturation_effort * (1.0 - self._joint_vel / self.velocity_limit)
        torque_speed_bottom = self._saturation_effort * (-1.0 - self._joint_vel / self.velocity_limit)
        # -- max limit
        max_effort = torch.clip(torque_speed_top, max=self.effort_limit)
        # -- min limit
        min_effort = torch.clip(torque_speed_bottom, min=-self.effort_limit)
        # clip the torques based on the motor limits
        clamped = torch.clip(effort, min=min_effort, max=max_effort)
        return clamped

It would be nice to implement this behaviour in the sim node as well.