【正文】 rating conditions to observe the control Fig. 12. Effect of pressure feedback control on boom angle. Fig. 13. Effect of feedforward control on control error of Z. characteristics, and at the same time to determine the optimal control parameters ?such as the control gains shown in Fig. 6.. . Effects of feedforward control In the case of position feedback only, increasing gain K to decrease error DZ causes oscillation due p to the time delay in the system, as shown by AOFFB in Fig. 13. That is, K cannot be increased. Applying the feedforward of the arm lever value described in Section can decrease error without increasing K as shown by AONB in the figure. p . Effects of pensation in attitude Level crowding is apt to bee oscillatory at the raised position or when crowding is almost pleted. This oscillation can be prevented by changing gain K according to the attitude, as has been p discussed in Section . The effect is shown in . This shows the result when the level crowding was done at around 2 m above ground. Compared tothe case without the pensation, denoted by OFF in the figure, the ON case with the pensation provides stable response. Fig. 14. Effect of adaptive gain control on control error of Z. . Effects of control inter213。idual attachments For each attachment of the boom, arm, and bucket, the reference angle was changed 58 stepwise from the initial value, and the responses were measured。 selects control modes andcalculates actuating variables。e gain scheduling according to the attitude In articulated machines like hydraulic shovels,dynamic characteristics are greatly susceptible to the attitude. Therefore, it is difficult to control the machine stably at all attitudes with constant gain. To solve this problem, the adaptive gain scheduling according to the attitude is multiplied in the feedback loop Fig. 6.. As shown in Fig. 7, the adaptive gain (KZ or Ku). is characterized as a function of two variables , and Z. means how the arm is extended, and Z means the height of the bucket. 5. Simulation results The level crowding control was simulated by applying the control algorithm described in Section 4 to the hydraulic shovel model discussed in Section 2. In the simulation, our large SK16 hydraulic shovel was employed.. Fig. 8 shows one of the results. Fiveseconds after the control started, load disturbance Fig. 6. Block diagram of control system (Z). Fig. 7. Gain scheduling according to the attitude. was applied stepwise. Fig. 9 shows the use of feedforward control can reduce control error. 6. Semiautomatic control system Based on the simulation, a semiautomatic control system was manufactured for trial, and applied to the SK16 shovel. Performance was then ascertained by field tests. This section will discuss the configuration and functions of the control system. . Configuration As illustrated in Fig. 10, the control system consistsof a controller, sensors, man–machine interface,and hydraulic control system. The controller is based on a 16bit microputer which receives angle input signals of the boom, arm, and bucket from the sensor。 and csflow coefficient. . Link relations In the model shown in Fig. 1, the relation between the cylinder length change rate and the attachment rotational angular velocity is given as follows: (1)leg (2). arm (3). bucket 當 時， Fig. 3. Model of hydraulic cylinder and valve. . Torque relations From the link relations of Section , the supply torque t is given as follows, taking cylinder friction I into consideration: Where, Cci is the viscous friction coefficient and Fi is kiic frictional force of a cylinder. . Response characteristics of the spool Spool action has a great effect on control , we are assuming that the spool has the following firstorder lag against the referenceinput. Where, XX is the reference input of spool displacement and T is a time constant. 3. Angle control system As shown in Fig. 4, the angle u is basically controlled to follow the reference angle u by position feedback. In order to obtain more accurate control, nonlinear pensation and state feedback are added to the position feedback. We will discuss details of these algorithms as follows. Fig. 4. Block diagram of control system . Nonlinear pensation In the ordinary automatic control systems, new control devices such as servo valves are used. In our semiautomatic system, in order to realize the coexistence of manual and automatic operations, we must use the main control valves, which are used in manual operation. In