JP2006043872A - Robot and method for controlling robot - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To avoid excessive load to a muscle of a person by a method for controlling a robot performing a work in cooperation with the person. <P>SOLUTION: This method relates to controlling of the robot by information of a sensor for measuring load to a person by arranging the sensor on the person side. <P>COPYRIGHT: (C)2006,JPO&NCIPI

Description

  The present invention relates to a robot control method, and more particularly, to a robot control method capable of working in cooperation with a human by using information on a human limb body joint angle or the like.

Patent Document 1 discloses a conventional robot control method that can perform work with emphasis on human beings. The prior art will be described with reference to the drawings.
FIG. 5 is a configuration diagram of a robot capable of working in cooperation with a conventional human. In FIG. 5, 1 is a robot arm, 2 is a hand, 3 is an operator, 4 is an object, and 5 is a force sensor. FIG. 6 is a diagram illustrating detection output components of a force sensor used in the control of the object cooperative transport robot. In FIG. 6, a force sensor 5 is provided between the robot arm 1 and the hand 2 at the tip thereof, and the force sensor 5 moves the object 4 from the operator side Q at the object support point P on the robot side of the object 4. The force applied to the robot arm 1 side is detected, and the obtained sensor signal is converted into a horizontal translational force component Fx in the horizontal front-rear direction (PQ direction), a translational force component Fz in the vertical direction (PR direction), and around the point P. Is broken down into rotational force components τ around the horizontal axis. Each joint of the robot arm is provided with a joint actuator for driving the joint and an angle sensor for detecting the joint angle, and the inclination angle θ of the object is detected based on the sensor signal. For the translational force component Fx and the rotational force component τ in the horizontal front-rear direction, the horizontal speed and the rotational speed of the hand at the point P of the object 4 are determined so that the robot moves without resistance in each direction. The Further, the vertical speed at the point P for leveling the object is determined in proportion to the inclination angle θ of the object 4. For that purpose, based on the rotational force component τ around the horizontal axis and the horizontal translational force component Fx in the sensor signal of the force sensor 5, the rotational motion component around the horizontal axis and the translational motion component in the horizontal longitudinal direction are obtained from the robot. The robot arm is driven by setting and outputting a gain so that the resistance force of the hand is reduced. On the other hand, for the hand angle (tilt angle θ of the object 4) detected by the sensor signal of the hand angle sensor, The motion command of the proportional translational motion in the vertical direction is output, and the hand of the robot arm is driven up and down so that the object remains horizontal. Thus, around the point where the robot supports the object (point P in FIG. 6), the object can freely rotate and no torque is applied to the object. The force is almost halved according to the mass distribution of the long object. Furthermore, when a human applies a horizontal force to an object, the robot moves horizontally without resistance. Further, when a human lifts an object and the object is tilted, a vertical upward speed is generated at the robot hand so that the tilt angle becomes zero, so that the object rises while maintaining the level. The same applies to a case where an object held by a human is lowered. FIG. 7 is a block diagram of a control system of a conventional object cooperative transport robot. In FIG. 7, the coordinate conversion unit (a) separates the translational force component Fx in the horizontal front-rear direction, the translational force component Fz in the vertical direction, and the rotational force component τ around the horizontal axis. On the other hand, in the angle sensor of each joint of the robot arm, the inclination angle θ of the object 4 is measured. In the force-motion conversion unit, the translational motion component (speed / acceleration) in the PQ direction at the point P in FIG. 6 is determined based on the translational force component Fx in the horizontal front-rear direction, and the rotational force around the horizontal axis. Based on the component τ, a rotational motion component (angular velocity / angular acceleration) around the point P is determined. These are determined by setting gains that reduce the resistance of the robot in the rotation and translation directions.
As described above, the conventional technology sets the gain based on the force sensor signal such that the rotational force component τ around the horizontal axis and the translational force component Fx in the horizontal front-rear direction reduce the resistance force of the robot. Based on the angle θ of the hand detected by the angle sensor, a motion command for translational motion in the vertical direction of the hand is output so that the object 4 is kept horizontal.

JP 2000-343470 A

However, depending on the posture, a human may start to move (stand up, etc.) from a stretched state. At this time, in spite of the situation in which the muscle output is difficult to exert, the conventional configuration cannot avoid an excessive load even if the human himself performs an unreasonable operation.
Therefore, an object of the present invention is to provide a robot control method that can avoid an excessive load on human muscles and has improved safety.

According to the first aspect of the present invention, there is provided a method for controlling a robot that performs work in cooperation with a human, wherein a sensor for measuring a load on the human is arranged on the human side, and the robot is controlled based on sensor information. It is.
According to a second aspect of the present invention, in the method for controlling a robot according to the first aspect, the sensor is an acceleration sensor incorporated in a wearable unit attached to a human limb joint, and detects a distal acceleration of the joint of the limb. A step of detecting the proximal acceleration, a step of calculating joint angle information of the limb from the acceleration information, and a step of transmitting the limb joint angle information to the robot.
According to a third aspect of the present invention, in the robot control method according to the second aspect, the operation speed of the robot is controlled in accordance with the limb joint angle information.
According to a fourth aspect of the present invention, in the robot control method according to the third aspect, the robot operating speed (v) is expressed by the equation v = (h ₂ −h ₁ ) (v ₂ −v ₁ ) / (h− h ₁ ) + v ₁
It is decided according to. Here, h ₁ is the minimum height, h ₂ is the maximum height, v ₁ is the speed corresponding to h ₁ , v ₂ is the speed corresponding to h ₂ , and h is the height during operation.
According to a fifth aspect of the present invention, in the robot control method according to the third aspect, the operation speed of the robot is stored in a memory corresponding to the limb body joint angle information.
According to a sixth aspect of the present invention, in the robot control method according to the second or third aspect, the limb joint angle information is wirelessly transmitted to the robot.
The present invention according to claim 7 is a robot that performs work in cooperation with a human, and controls the operation speed of the robot based on sensor means arranged on the human side for measuring a human load and information of the sensor means. And a control means.
The present invention according to claim 8 is the robot control method according to any one of claims 2 and 3, wherein when the joint angular velocity of the limb is smaller than a predetermined value, the robot is controlled according to the joint angular velocity of the limb. The motion speed is determined. If the motion speed is large, the motion speed of the robot is controlled according to the joint angle information of the limb.

According to the first aspect of the present invention, since the state of a load on a human can be directly monitored, an object can be gripped and transported without applying an excessive load to the human.
According to the second aspect of the present invention, since the joint angle of the human limb can be directly monitored, the object can be gripped and transported without applying an excessive load to the human muscle.
According to the third aspect of the present invention, assist according to the dynamic characteristics of human muscles is possible, so that an object can be grasped and transported without applying an excessive load to human muscles.
According to the fourth and fifth aspects of the present invention, it is possible to perform an operation most suitable for human muscles, and it is possible to grasp and carry an object without applying an excessive load.
According to the sixth aspect of the present invention, since the cable between the sensor and the robot can be eliminated for the sensor information, the operability can be improved.
According to the seventh aspect of the present invention, it is possible to provide a robot that can perform an operation most suitable for human muscles.
According to the eighth aspect of the present invention, since the robot operates in response to the start and stop of human arbitrary muscle contraction, safety can be improved.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings.

  FIG. 1 is a schematic view of a biological information monitoring robot of the present invention and its outline. The human 119 wears the wearable unit 101 and transmits posture information of his / her limb to the biological information monitoring robot 102 every moment. The configuration of the present invention shown in FIG. 1 shows a case in which one long object 120 is lifted in cooperation with the biological information monitoring robot 102 from a state in which a human 119 is squatting. In this case, when the person 119 is squatting, the quadriceps femoris muscle (frontal muscle of the thigh), which is the knee extensor muscle, is stretched. The wearable unit 101 is not characteristic for the present invention and is equivalent to that used in a known biological information monitoring system. FIG. 2 is a block diagram of the system of the present invention, and the same components as those in FIG. The wearable unit 101 includes a proximal acceleration sensor 104 that detects a posture on the proximal side of the limb, a distal acceleration sensor 106 that detects a posture on the distal side of the limb, and a CPU 113 that performs A / D conversion of each sensor information. The transmitter 114 transmits the output from the CPU 113 to the biological information monitoring robot 102, and the battery 115 supplies power to each component. Similarly to the conventional robot 111, the biological information monitoring robot 102 includes a robot arm 108 that performs cooperative work with the human 119 and a driver 109 that is a control unit for driving the robot arm 108. The driver 109 includes an I / O 110 and performs input / output with respect to an external device. In the case of the present invention, an I / O 110 and a receiver 112 for acquiring sensor information from the wearable unit 101 are provided.

In the present invention, a human wears a sensor for detecting the posture of the limb on the proximal side and the distal side of the joint of the limb, and transmits the sensor information to the robot (by radio). The robot sets the movement speed of the hand according to the joint angle of the human limb. As shown in FIG. 4, generally, the shorter the muscle is loaded and stretched, the smaller the shortening speed. This muscle dynamic characteristic is known as a muscle Hill type model 103. The upper and lower limits of the speed are not constant values, but are adjusted as appropriate according to the work and application site. Further, the relationship between the angle of the limbs and the motion speed is not limited to the linear relationship as in the graph 103, and may be an inversely proportional or exponential curve. The present invention operates the robot according to the dynamic characteristics of the muscle. That is, the knee acceleration θk (definition of the knee angle θk: the maximum extension time is set to 0 ° and increases with flexion) is monitored by the proximal acceleration sensor 104 and the distal acceleration sensor 106, and the knee is increased with rising. The lifting speed is increased in accordance with the degree of extension (formula (1)).
Zref = Zcur + ((v ₁ −v ₂ ) (θk−θm) / θm + v ₁ ) Δt (1)
Here, Zref is in the vertical direction in the absolute coordinate system, the target position in the next control cycle, Zcur the position of the current control cycle, v ₁ is the minimum velocity, v ₂ is the maximum speed, .theta.m the maximum knee angle, Δt represents a control cycle.
The knee angle measuring device uses an acceleration sensor in the above example, but this has a built-in pendulum and acts as a tilt sensor. Since a pendulum is used, the absolute angle with respect to the vertical direction can be measured. Therefore, as shown in FIG. 4, the thigh inclination angle θ ₁ and the lower leg inclination angle θ ₂ are obtained as sensor outputs, and the knee angle θk is θ ₁ + θ ₂ . As a tilt sensor, a relative angle θk between the proximal limb posture and the distal limb posture may be obtained using a gyro sensor or the like. In that case, the output of the proximal sensor (speed) v _alpha proximal side sensor output (speed) v _beta, when the control period is Delta] t, the relative angle θk becomes (v α -v β) _Δt. However, in order to measure the flexion / extension of the limb joints, each sensor is preferably a posture sensor having two or more axes, but a so-called goniometer may be used. Alternatively, one distance measuring sensor may be mounted in the vicinity of the knee joint, and the distance h from the ground to the knee joint may be measured to correspond to the ascent speed v. (Formula (2))
v = (h ₂ −h ₁ ) (v ₂ −v ₁ ) / (h−h ₁ ) + v ₁ (2)
Here, _{h 2} is the maximum height, _{h 1} is the minimum height. Examples of the distance measuring sensor include a magnetic sensor and an ultrasonic sensor. If each of the sensors 104 and 106 is of a built-in amplifier type, the size and weight can be reduced, and the attachment to the limbs can be improved. Since each of the sensors 104 and 106 needs to detect an operation of about 1 Hz such as rising, it is desirable that the CPU 113 has an A / D resolution of about 10 bits. The sensor information of each sensor 104, 106 is transmitted to the biological information monitoring robot 102 by wire or wirelessly, but the transmission method is preferably about 10 ms in time resolution and about 3 m in transmission distance. When wired, it is superior to wireless in terms of responsiveness and reliability. Further, when configured wirelessly, even if the human 119 and the robot 102 work away from each other, the operability can be improved without being disturbed. In the case of wireless communication, it is desirable that the transmitter 114 be composed of an antenna substrate or a chip antenna in order to reduce the size and weight of the transmission module. The power consumption of each sensor 104, 106, CPU 113, and transmitter 114 can be about 3V, and the current consumption can be about 5mA. The battery 115 uses a lithium ion secondary battery (about 750mAh) for mobile phones. However, a manganese dioxide lithium battery (1400 mAh) for a camera may be used. With such a configuration, it is possible to attach to the limb even if all the components of the wearable unit 101 are integrated, but the sensor unit and other than the sensor unit (telemeter) may be separated. Telemeters usually used for exercise measurements such as athletics have a transmission distance of several hundred meters or more and a weight of about 200 g. Therefore, it is difficult to integrate a sensor with a limb body, Examples of attaching the to the waist and back are known. The wearable unit 101 is attached to the limb in FIG. 3 because it is a supporter type, and the sensor casing is fixed with Velcro (registered trademark). it can. Each case may be fixed to the limb with a belt with Velcro (registered trademark) (registered trademark).
In the above example, the squatting posture is given, but the target motion is not limited to this, and it can be applied to a case where a human and a robot push or pull something at the same time. Also, the target part is not limited to arms, legs, neck, waist, etc. The postures and positions of a plurality of parts may be simultaneously measured and reflected in the operation. That is, according to the present invention, an excessive load on any human muscle can be avoided.
Note that the robot control method described above supports only the angle of the human limbs. For example, even when the human is in a certain posture, the robot operates at a predetermined speed, which is dangerous. It is. Therefore, in order to operate the robot in response to the start and stop of human arbitrary muscle contraction, as shown in FIG. 8, when the angular velocity dθk / dt of the limb is smaller than a predetermined value dθm / dt, The robot is controlled based on the angular velocity, not the limb angle. That is, the rising speed is increased according to the increase in the angular velocity dθk / dt of the limb until the rising speed reaches v1. The method of increasing the rising speed may be linear as shown in FIG. 8, or may be stepped or curved, and is not limited. Thus, since the robot operates in response to the start and stop of voluntary muscle contraction by humans, safety can be improved.

  The robot control method of the present invention can be applied to a robot control method that performs work in cooperation with a human.

Biological information monitoring robot and its schematic diagram Block diagram of the present invention The figure which shows the mounting state of this invention The figure which shows the control law of this invention Configuration diagram of conventional technology The figure explaining the detection output component of a prior art Prior art control configuration diagram The figure which shows the control law of this invention

Explanation of symbols

101 Wearable unit (transmitter)
102 Biological information monitoring robot (receiver)
103 Muscle type model 104 Proximal acceleration sensor 105 Wireless unit (transmission side)
106 Distal acceleration sensor 107 Wireless unit (receiving side)
108 Robot arm 109 Driver 110 I / O
111 Conventional Robot 112 Receiver 113 CPU with A / D Conversion Function
114 Transmitter 115 Battery 116 Supporter 117 Magic Tape (Registered Trademark)
118 Power and signal lines to sensor 119 Human 120 Long object

Claims (8)

In the control method of the robot that works in cooperation with humans,
A method for controlling a robot, wherein a sensor for measuring a load on a human is arranged on a human side, and the robot is controlled based on information from the sensor. The sensor is an acceleration sensor incorporated in a wearable unit attached to a human limb body joint,
Detecting the distal acceleration of the joints of the limbs;
Detecting proximal acceleration; and
Calculating the limb joint angle information from the acceleration information;
The robot control method according to claim 1, further comprising a step of transmitting the limb joint angle information to the robot.   The robot control method according to claim 2, wherein an operation speed of the robot is controlled according to the limb joint angle information. The motion speed (v) of the robot is expressed by the equation v = (h ₂ −h ₁ ) (v ₂ −v ₁ ) / (h−h ₁ ) + v ₁
Where h ₁ is the minimum height, h ₂ is the maximum height, v ₁ is a speed corresponding to h _1, v speed ₂ corresponding to h _2, h is to be determined depending on the height of the operation The robot control method according to claim 3, wherein:   4. The robot control method according to claim 3, wherein the operation speed of the robot is stored in a memory corresponding to the limb joint angle information.   4. The robot control method according to claim 2, wherein the limb joint angle information is wirelessly transmitted to the robot.   It is a robot that performs work in cooperation with humans, and is composed of sensor means arranged on the human side for measuring human load and control means for controlling the operation speed of the robot based on information of the sensor means. Characteristic robot.   When the joint angular velocity of the limb is smaller than a predetermined value, the robot operation speed is determined according to the joint angular velocity of the limb, and when it is larger, the robot operation speed is controlled according to the joint angle information of the limb. The robot control method according to claim 2, wherein the robot control method is performed.

Images