if the inner force control loop is perfect. This condition involving the inner force control loop is verified in Fig. 7, where the force generated by the slave actuator accurately tracks the desired value for the control signal Fs. A similar result can be shown for the master-side inner force control loop. It is interesting to note that the performance of the inner-loop force control is less accurate in the free-motion case than in the contact-motion case. Indeed, in free motion where the piston position significantly varies and so do the volume chambers, the pressure variation increases according to (1). This leads to decreased accuracy of the pressure approximation in (9). Therefore, the performance of the inner-loop force control deteriorates during free motion. On the other hand, in contact motion where the position is nearly constant, the pressure variation is small, the pressure approximation in (9) is highly accurate, and thus good inner-loop force tracking performance can be achieved (see Fig. 7).
To better understand the dynamic behavior of the inner force control loop, a spectra analysis is investigated, as illus- trated in Fig. 8. It can be seen that the transfer function of the inner loops approaches to the unitary function at low frequencies (less than 2 Hz). For faster movements (e.g., at 7 Hz), the response is degraded. In conclusion, the pres- sure/force prediction is sensitive to the movement bandwidth of the master/slave manipulator. In our experimental valida- tion, the arm movements were slow enough to be able to assume that a highly transparent teleoperation system (whose response is shown in Fig. 6) can be obtained through the employed hybrid force control (whose response is shown in Fig. 7).Fig. 9 shows the control voltage of the four ON/OFF valves at the slave side. As it can be seen, in free motion as well as under contact, mode 1 (i.e., all valves are closed) and mode 8 (i.e., all valves exhaust) are most used, which allow to keep the difference of pressure in the chambers (Pp – Pn) constant. These two modes allow for the chattering and energy consumption to be significantly reduced. Similar result can be observed at the master side.
b) Three-channel (3CH) case: Another benefit of the general 4CH architecture of Fig. 4 is that by proper adjustment of the control parameters, it is possible to obtain two classes of 3CH control architectures, which can be transparent under ideal conditions [14]. In this way, there is no need for master/operator or slave/environment interaction force mea- surement. The need for fewer force sensors without degrading transparency makes the 3CH architectures attractive from the implementation point of view [14].
Fig. 10 shows the master and slave positions and force tracking profiles for the 3CH teleoperation system in which
if the inner force control loop is perfect. This condition involving the inner force control loop is verified in Fig. 7, where the force generated by the slave actuator accurately tracks the desired value for the control signal Fs. A similar result can be shown for the master-side inner force control loop. It is interesting to note that the performance of the inner-loop force control is less accurate in the free-motion case than in the contact-motion case. Indeed, in free motion where the piston position significantly varies and so do the volume chambers, the pressure variation increases according to (1). This leads to decreased accuracy of the pressure approximation in (9). Therefore, the performance of the inner-loop force control deteriorates during free motion. On the other hand, in contact motion where the position is nearly constant, the pressure variation is small, the pressure approximation in (9) is highly accurate, and thus good inner-loop force tracking performance can be achieved (see Fig. 7).To better understand the dynamic behavior of the inner force control loop, a spectra analysis is investigated, as illus- trated in Fig. 8. It can be seen that the transfer function of the inner loops approaches to the unitary function at low frequencies (less than 2 Hz). For faster movements (e.g., at 7 Hz), the response is degraded. In conclusion, the pres- sure/force prediction is sensitive to the movement bandwidth of the master/slave manipulator. In our experimental valida- tion, the arm movements were slow enough to be able to assume that a highly transparent teleoperation system (whose response is shown in Fig. 6) can be obtained through the employed hybrid force control (whose response is shown in Fig. 7).Fig. 9 shows the control voltage of the four ON/OFF valves at the slave side. As it can be seen, in free motion as well as under contact, mode 1 (i.e., all valves are closed) and mode 8 (i.e., all valves exhaust) are most used, which allow to keep the difference of pressure in the chambers (Pp – Pn) constant. These two modes allow for the chattering and energy consumption to be significantly reduced. Similar result can be observed at the master side.b) Three-channel (3CH) case: Another benefit of the general 4CH architecture of Fig. 4 is that by proper adjustment of the control parameters, it is possible to obtain two classes of 3CH control architectures, which can be transparent under ideal conditions [14]. In this way, there is no need for master/operator or slave/environment interaction force mea- surement. The need for fewer force sensors without degrading transparency makes the 3CH architectures attractive from the implementation point of view [14].
Fig. 10 shows the master and slave positions and force tracking profiles for the 3CH teleoperation system in which
การแปล กรุณารอสักครู่..
