JPS62187671A - Walking stability controlling system of multi-articulated walking robot - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Application of the Invention] The present invention provides a dynamic (
In other words, it relates to a control method for stably performing independent walking (in which the center of gravity is outside the soles of the feet).

[Background of the invention]

In order for a robot to walk independently, all necessary functional devices must be installed in the robot body, but in the case of a multi-joint walking robot that has these devices installed in the head, the center of gravity is located at the top, which causes problems. It is stable. Furthermore, in order to make an articulated walking robot walk at high speed, a dynamic walking method in which the center of gravity moves outward from the sole surface of the foot is adopted, so there is a high possibility that the robot will become unstable due to the effects of external and internal disturbances. For example, in multi-joint walking robots that are equipped with all the necessary functional devices, hydraulic drive mechanisms are often used for the joints because large torques are applied to the joints, but in the case of hydraulic drive mechanisms, the joint movements change depending on the oil temperature. occurs, and this acts as a civil war. Also, the floor or ground on which you walk (simply referred to as the ground)
The fact that the ground is uneven and shocks act as disturbances.

The journal of the Robotics Society of Japan, Vol. 1, No. 3 (Oct.
There is ``Analysis and control of bipedal locomotion with constraints'' by Masami Ito and Tatsuo Narukiyo in PP, pp. 31-35. This paper states that in dynamic stability control of bipedal walking robots, it is important to control the period during which support legs are replaced. No consideration has been given to solutions for instability caused by internal disturbances due to changes in dynamic characteristics resulting from changes in

In addition, as a literature on the control of bipedal walking robots with soles, Tsutomu Mita, ``On the control and realization of high-speed bipedal walking robots'' in Robot Engineering (September 1982), PP, 104-114. This paper discusses the control of the sole of the foot, which discusses the control to make the sole of the foot in the air parallel to the ground and the posture control of the supporting leg, which is similar to the above-mentioned literature. No active measures to deal with instability factors are mentioned.

[Object of the Invention] The object of the present invention is to reduce the effects of external disturbances and internal disturbances by using signals from ground information sensors installed on the soles of the feet in dynamic walking control, which is essential for increasing the walking speed of multi-joint walking robots. It is an object of the present invention to provide a control device that performs stable control by eliminating the problem and realizes a method that enables a multi-joint walking Rohit to walk independently.

[Outline of speech]

The present invention provides a walking control method for a multi-joint walking robot that realizes walking by sequentially driving each joint in response to a sequential joint angle command signal based on a preset walking pattern. When the sole contact information sensor detects that the foot has touched the ground at an angle to the left or right,
means for locking the joint angle in the pitch direction based on the preset walking pattern; and means for controlling the joint angle in the roll direction during the locking period so as to correct the horizontally inclined ground contact of the sole of the foot. It is characterized by:

[Embodiments of the invention]

An example in which the present invention is applied to a bipedal walking mouth-getting device will be shown below.

FIG. 2 shows a front view of the biped walking robot device of this embodiment. The biped walking mouth bit device has a total of 12 degrees of freedom for each leg. In other words, in a position corresponding to the human waist,
A waist pitch axis 1 and a waist roll axis 2, a knee pitch axis 3 at a position corresponding to the knee, an ankle pitch axis 4 and an ankle roll axis 5 at a position corresponding to the ankle, and a large axis for changing direction. It has a direction changing yaw axis 6 in its thigh. Each joint shaft has a hydraulic swing actuator 10 for driving the joint. Furthermore, a potentiometer 12 for joint angle detection is attached to the shaft of the hydraulic swing actuator lO. A servo valve 13 is attached to control the operation of the hydraulic swing actuator IO.

In addition, this robot has soles on both legs, and these soles attempt to compensate for instability during dynamic walking to a certain extent. It also plays an important role in enabling work to be completed and stopped. As shown in FIG. 3, a total of eight ground contact information sensors 11, four on each foot, are installed on the sole surface of the foot.

The hydraulic power source is composed of a hydraulic pump motor 9, a hydraulic tank 9', a relief valve 9, etc., and supplies the pressure and oil flow rate necessary to drive the bipedal walking robot device. The flow rate of oil to the hydraulic swing actuator 10 of each joint is adjusted by moving the servo valve 13. Thereby, the drive of the hydraulic swing actuator IO is controlled.

A control unit 7 is installed above the hydraulic power source with a heat insulating material 8 interposed therebetween to prevent it from being affected by the heat generated by the hydraulic pump motor 9. The control unit 7 includes a microcomputer that processes information that determines the walking motion of one lobe, its peripheral equipment, and an amplifier that drives and controls the servo valve 13 that controls the flow rate of oil using an electric current.

The hydraulic power sources 9, 9', 9'' and the servo valves 13 are connected by flexible piping hoses.The control unit 7, the ground information sensor 11, the servo valves 13, and the angle detection potentiometers 12 are connected to each other by flexible piping hoses. They are connected by a signal cable.

FIG. 1 shows an overall diagram of the control system of the biped walking robot device of this embodiment. The joint angle command signal according to the walking/mother turn output from the computer control system 15 is as follows:
Q is amplified by amplifier 16 and transmitted to servo valve 13. The servo valve 13 is connected to the hydraulic power source 9F9Z9"
The flow rate of oil supplied from the hydraulic oscillating actuator 10 is adjusted by moving the spool inside the valve according to the angle command signal from the servo amplifier 16.
drive is controlled. At the same time, the position (angle) feedback control for the joint angle command signal is performed using a signal from the notensiometer 12 that detects the actual joint angle, and walking is performed according to the walking pattern. On the other hand, a signal from the sole contact information sensor 11 causes control operations to be switched as described later.

FIG. 4 shows the configuration of the control device (computer control system 15 and servo amplifier 16 in FIG. 2). The control device has a configuration in which the illustrated boards are arranged on both sides via an internal path 23. The control device is a servo amplifier 16a for the right leg.
, left foot amplifier 16b, microcomputer 1
7. Interface speed 18a, 18b, memory board 19, D/A converter 21% A/D converter 2
Consists of 2.

The microcomputer 17 calculates angle commands for each joint necessary for normal walking using the program and walking pattern built into the memory date 19, and the D/A converter 2.
Output to 1.

Figure 5 shows an outline of the walking cycle turn data.

The walking pattern data represents the motion trajectory that each joint of the bipedal walking robot device should take during walking, and specifically, the walking pattern data represents the trajectory of the joints that correspond to walking as shown in Figure 9. This is joint angle command data extracted discretely from a time series of angle commands.

The microcomputer 17 then extracts joint angle command data to be output to each joint from this walking pattern data, and interpolates between the current joint angle command data and the retrieved data. The angle command resulting from this interpolation is converted and calculated so that it can be output to the D/A converter 21, and output to the D/A converter 21. The D/A converter 21 converts this angle command into a current, and the servo amplifiers 16a and 16b amplify this current and output it to the servo valve.

In addition, the servo amplifier tea is connected via the connector.

A response angle signal from the actual movement of the joint is input from the angle detection potentiometer 12 to 16b, forming a position feedback control circuit. FIG. 6 shows a block diagram of the position feedback control circuit.

70 points of walking control using normal walking and turning as described above
-The chart becomes K as shown in Figure 7.

The drive signal switching circuit 26 shown in FIG. 6 is especially provided in accordance with the present invention, and serves to switch the angle command by turning on and off a drive inhibit signal output from the microcomputer 17, which will be described later. The switching situation of the angle command signal is as shown in FIG. 8, and when the drive inhibit signal is OFF, the angle command output from the drive signal switching circuit 26 is the same as the angle command output from the microcomputer 17. When the drive prohibition signal is turned ON, the angle command output from the drive signal switching circuit 26 is held at the actual angle value at that time, and therefore, regardless of the angle command from the microcomputer 17, the joint is set at the angle at that time. is locked.

FIG. 9 shows the walking state of the bipedal walking robot device of this embodiment, where A and A' are vehicle leg support periods, and B is a support leg switching period during which the leg supporting the center of gravity is replaced. In the present invention, dynamic walking is performed only during this support leg switching period B. Therefore, walking becomes extremely unstable during the period of switching support legs. Joint movement trajectories that enable stable walking are stored in memory in the form of gait patterns, but there is a problem that cannot be solved by simply following these trajectories during the support leg replacement period. Examples include adaptation to uneven terrain, changes in dynamic characteristics due to changes in oil temperature, and impact upon touchdown. In order to deal with these problems, this embodiment uses a control method that randomly adapts joint movement trajectories to the walking situation using signals from the ground contact information sensor of the sole during support leg replacement period B. take. This will be explained below.

The first step is to examine the changes in ground contact information during drifting motion during the support leg switching period.
This is shown in Figures (,) to (e). Walking is performed in the direction of the arrow in the figure, that is, in the direction from (,) to (e) in the figure. In the figure, the black circle indicates the ground information sensor that is in contact with the ground and outputs the ground information ON signal, and the white circle indicates the ground information sensor that is not grounded and outputs the ground information OFF signal.
This shows the ground information sensor that is emitting the signal.

Walking control for the bipedal walking mouth bit device of this embodiment to operate as shown in FIG. 10 during the support leg replacement period is performed as follows, as shown in the flowchart of FIG. 11.

First, the walking pattern is output at step 1, which is the state shown in (,) in Fig. 10, and then the ground contact information sensor of the heel of the front foot is
Slowly move each joint of the front leg's hip pitch, and the back leg's hip pitch and ankle pitch by the same angle in the direction of the arrow in Figure 12 (,) until N is reached. When one of the ground contact information sensors of the heel of the forefoot is turned ON, the above-mentioned motion prohibition signal is immediately input to all joints in the pitch direction, and all joints in the pitch direction are locked. Next, the ground contact information sensor of the OFF side of the heel of the front foot is also turned ON, that is, the ground contact information sensor of both the heel of the front foot is set to oNHc, as shown in Fig. 12 (b), in the roll direction of the ankles and hips of both feet. Controls the hydraulic swing actuator. When the state shown in FIG. 1O(b) is reached, in which both the ground contact information sensors of the heels of the front feet are turned on, all joints in the roll direction are locked, and the joints in the pitch direction are unlocked.

Next, the center of gravity is moved as shown in FIG. 12(c) from the state shown in FIG. 10(b) to the state shown in FIG. 10(c). For this purpose, control is performed using only pitch-direction joint data of a preset walking pattern. At this time, adjust the ankle pitch joint and hip pinch joint of the hind foot by moving them by the same angle so that the sole of the hind foot does not lift off the ground. The sole of the front foot is in horizontal contact with the ground, and all ground contact information sensors on the heel and toes are ON (Figure 10(c)).
Continue the above movements until you reach the state of

Next, from this state to the state shown in Figure 1O (d) (the heel of the hind foot is raised), control is performed by walking and turning the hind legs. Control the ankle pitch joint of the hind foot so that the toe leaves the ground and licks it. Posture where the movement of the center of gravity is completed (Fig. 10 (,)
When the robot moves to the state of 1), the joints in the roll direction are unlocked and the control method returns to the normal walking/turning method.

Due to the above, it was not possible to deal with the situation by controlling only by walking and e-turns. It is possible to solve the instability of walking due to the influence of uneven ground and changes in dynamic characteristics due to changes in oil temperature, and realize stable control during the support leg switching period.

〔Effect of the invention〕

According to the present invention, by correcting the horizontally tilted contact of the sole of the foot using the signal from the sole contact information sensor, stable dynamic walking is achieved by compensating for the influence of internal disturbances and external disturbances, and the robot It is possible to greatly improve independent walking performance, and because the number of components that make up the control system is small and the number of calculation processing steps is also small, there is no possibility of erroneous processing or malfunction due to control complexity. Few.

[Brief explanation of drawings]

FIG. 1 is an overall view of the control system of a biped robot according to an embodiment of the present invention, FIG. 2 is a front view of the biped robot, and FIG. 3 is a plan view of the sole of the biped mouth pit. Fig. 4 is a schematic diagram of the configuration of the control device of the same embodiment, Fig. 5 is a diagram showing joint angle commands based on walking patterns, Fig. 6 is a block diagram of the joint angle position feedback control circuit, and Fig. 7 is a normal walking/j
Flowchart of walking control based on P-turn, Figure 8 (
a) l (b) + (c) is a diagram showing the angle command from the microcomputer, the angle command and the drive prohibition signal output from the drive signal switching circuit in the above embodiment;
The figure is a diagram showing the state of walking by walking i4 turn, 10th
Figures (,) to (,) are diagrams showing the state of the sole of the foot during the switching period of the supporting leg, Fig. 11 is a flowchart of walking control during the switching period of the supporting leg in the embodiment of the present invention, and Fig. 12 (a)
1(b) and 1(e) are schematic side and front views showing the joint control mode during the support leg switching period. 1... Hip pitch axis, 2... Hip roll axis, 3
...Knee pitch axis, 4...Ankle pitch axis, 5
... Ankle roll axis, 6... Direction change yaw axis, 7
...Control unit, 8.Thermal insulation material, 9..Hydraulic pump motor, 10..Hydraulic swing rotary actuator, 11.
...Ground information sensor, 12...Potentiometer, 13...Servo valve, 1
4...Power supply unit, 15...Computer control system, 16...Serv amplifier, 17...Microcomputer, 18...Interface date, 19...Memory speed, 21...D/A converter , 22... A/D converter, 23... Internal lotus, 24... Amplifier, 25... Limit circuit. 26... Drive code switching circuit, A, A'... Single leg support period, B... Support leg switching period. (L) (b) Direction of movement (c) (d) ( e
) Figure 11 Figure 12 (C) Direction of travel

Claims (1)

[Claims] In the gait control method of a multi-joint walking robot, which realizes walking by sequentially driving each joint in response to a sequential joint angle command signal based on a preset walking pattern, the sole of the robot's foot is in contact with the ground. When the sole contact information sensor detects that the foot has touched the ground tilted in the left-right direction, a means for locking the joint angle in the pitch direction based on the preset walking pattern, and a means for locking the joint angle in the roll direction during the locking period. A walking stability control method for a multi-joint walking robot, characterized by comprising means for controlling the above-described horizontally inclined ground contact of the sole of the foot.