CN1947960A - Environment-identification and proceeding work type real-man like robot - Google Patents
Environment-identification and proceeding work type real-man like robot Download PDFInfo
- Publication number
- CN1947960A CN1947960A CN 200610123018 CN200610123018A CN1947960A CN 1947960 A CN1947960 A CN 1947960A CN 200610123018 CN200610123018 CN 200610123018 CN 200610123018 A CN200610123018 A CN 200610123018A CN 1947960 A CN1947960 A CN 1947960A
- Authority
- CN
- China
- Prior art keywords
- anthropomorphic robot
- control
- walking
- man
- robot
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Landscapes
- Manipulator (AREA)
Abstract
A man-shaped robot for recognizing environment and conducting job is composed of a system for coordinating rotary two eyes and controlling movement, a pair of 5-finger manipulators with tactile sense, a two-foot walking mechanism, and a man-machine interface platform based on computer network, relative mathematical models and control algorithm.
Description
Technical field
The present invention relates to a kind of robot, particularly a kind of under heavy, dangerous, abominable or general environment, can environment-identification and replace the people to finish the anthropomorphic robot of various complex jobs.
Background technology
Existing anthropomorphic robot vision mainly comprises two types.The first kind is disclosed by 2005 05 interim " present Research and the development trend of bionical eye " papers of " robot " magazine, based on " Pin-Hole " model, its defective is before each operation, must finish very loaded down with trivial details ccd video camera proofreaies and correct, the motion control of ccd video camera also need be finished extremely complicated coordinate transform between visual space and working space, and also can't realize the coordinated movement of various economic factors and various visual function and every operation based on visual processes of two ccd video cameras effectively.Second class also is disclosed by aforesaid magazine and paper, be to calculate in the three dimensions to be watched attentively the deep or light and range image information such as luminous flux that object enters ccd video camera, thereby observe and be familiar with and watched attentively object, and by calculate the luminous flux energy function of being watched attentively object control ccd video camera in real time tracing and positioning watched attentively object.But, this method very easily is subjected to the various interference effects from environment, and the computational algorithm of luminous flux and energy function is very complicated, be difficult under the dynamic environment in real time and use, and the coordinated movement of various economic factors control problem of two ccd video cameras and various visual function and every operation based on visual processes also can't realize effectively.
In the factory and enterprise of developed countries such as American-European and Japan, industry mechanical arm is widely used for replacing the people to finish all kinds of simple and repetitive works.These industry mechanical arms are to be limited to finish single action in the particular environment basically.For some under dangerous, abominable, heavy or general environment, the complex job that needs dexterous staff and wrist thereof and arm just can finish, for example grab, hold, pinch, press from both sides, push away, draw, insert, by, operation such as cut, cut, strike, beat, tear, twist, lead, drag, grind, cut, dig, frustrate and just seem powerless.
And existing the five fingers shape apery manipulator nearly all is to refer to the multi-joint hand from mechanism form more, the most generally pointing number is 3～5, the joint number of each finger also mostly is 3 cradle heads, the free degree is 3～5, has following problem: 1) the apery manipulator is to adopt based on the kind of drive of tendon and the drive principle of imitating staff mostly.Although the tendon transmission has many advantages, in the tendon transmission system,, can cause that drive system lags behind, tendon is unstable or damage because rigidity, mechanical property, quantity and the path design in finger etc. of tendon have bigger influence for the performance of hand; 2) control method of apery manipulator is to be equipped with independently to the finger tip place that the finger of power/torque sensor carries out position and force feedback control or Flexible Control basically.These control methods only are fed back to basic adversary's fingering row control with the contact force/moment at finger tip place, do not have the contact force influence of considering or controlling numerous sense of touch points on finger and the palm surface; 3) research of apery manipulator does not have or seldom relates in one's hands and wrist and arm control method for coordinating, do not have according to doing effectively and reasonably adjusting from the contact force/moment feedback adversary of manipulating object and inertia, viscosity, the rigidity parameter of wrist and arm thereof yet, with adapt to the different manipulating objects and the needs of operating environment (such as, the apery manipulator is grasping and processing flaccid products how, and perhaps the grasping metal tools is finished and installed and maintenance); 4) apery manipulator research does not relate to the job description language relevant with dexterous operation, and both hands and both hands wrist and the both hands arm that draws oneself up as the people more not coordinated control to a pair of apery manipulator and wrist thereof and arm.The research of existing hand and wrist thereof and arm only relates to grasping mostly, opens drawer, by switch, grab egg, pendulum building blocks, play the musical instrument etc., do not relate to as yet and higher level write, cutting, tear and operation such as drag, excavate, beat, grinding, plane are frustrated.
Because anthropomorphic robot is an a kind of nonlinear instability system, it will bear many restrictions that caused by contact environment in walking.The existing main problem of two sufficient walkings of existing anthropomorphic robot is not consider the reaction force influence of kiss the earth, perhaps only consider the haptic force influence of several sense of touch points on the sole, thereby under the kinetic model condition of unknown on walking ground, be difficult to control walk apery robot stabilizedly.
In sum, existing anthropomorphic robot can not possess people's basic function fully, can not replace the people to finish the operation of various complexity fully under heavy, dangerous, abominable or general environment.
Summary of the invention
The objective of the invention is to the defective at the prior art existence, a kind of described anthropomorphic robot has people's basic function, and it can replace the people to finish the operation of various complexity under heavy, dangerous, abominable or general environment.For example, in machine-building, Chemical Manufacture, the nuclear power maintenance, military engagement, medical operating, rescue and relief work, equipment is installed, scientific investigation, housework, patient's nurse, under the occasions such as Edutainment, described anthropomorphic robot can correctly carry out target identification, tracing and positioning, motion detection, hand eye coordination, obstacle is avoided, walking planning etc. are based on the operation of visual processes, on the ground of the unknown stable and self-discipline the workplace of walking, and cooperate mutually with finishing deftly with its both hands and both hands wrist and both hands arm and grab, hold, pinch, folder, push away, draw, insert, press, cut, cut, strike, beat, tear, paste, lead, drag, mill, cut, dig, the operation of various complexity such as frustrate.
The environment-identification that is used for of the present invention comprises with the anthropomorphic robot that carries out operation:
---rotatable two eye coordinate kinetic control systems, be used to finish based on target identification, tracing and positioning, motion detection, hand eye coordination, the obstacle of visual processes avoid, walking planning operation; Be made of two ccd video cameras, five DC servo motors, a high speed binary channels video tablet and two blocks of PIO plates, two ccd video cameras are equivalent to two eyeballs of people, have upper and lower, left and right motion and spinfunction;
---the tactile the five fingers shape of biobelt apery manipulator, be used to finish grab, hold, pinch, press from both sides, push away, draw, insert, by, cut, cut, strike, beat, tear, paste, lead, drag, grind, cut, dig, frustrate operation; Be distributed in totally 120 * 2=240 of the palm of described apery manipulator and the sense of touch point on five fingers thereof, its palm and five fingers thereof comprise 21 joints altogether, and both hands have 21 * 2=42 the free degree, a pair of arm and a pair of wrist that 7 * 2=14 the free degree arranged;
---two foot walking machine structures are used to control the walking stably on the ground of the unknown of described anthropomorphic robot; Be made of waist and two feet, have 15 joints, waist three degree of freedom, two feet are totally ten two frees degree; The power of two six degree of freedoms and torque sensor are installed on the root of described two feet, are used for testing environment reaction force and moment;
---the man-machine interface platform is support with the computer grid, by calculating relevant Mathematical Modeling and control algolithm, controls described anthropomorphic robot and realizes motor function, perceptional function, function and thinking, operation function, man-machine interaction; For providing interface, simplifies even transparent threedimensional model emulation, job description language, stressed effect analysis, visual motion and operation effectiveness, both hands and double-legged teaching, limbs collision detection, collision-free Trajectory Planning of Welding described anthropomorphic robot in network and cluster and grid.
Anthropomorphic robot height of the present invention equals described two foot walking machine structure height and adds described two eye coordinate kinetic control system height, is 1.5 meters, and the distance between 60 kilograms of body weight, two shoulders is 0.45 meter, and walking speed is greater than 2.0 kilometers/hour.
In the moving control system of rotatable two remainders allocation and transportation of anthropomorphic robot, two eyeball distances are set at 75 millimeters, and the distance between the gyroaxis from the eyeball to the neck is decided to be 85 millimeters.
The appearance and size that a pair of of anthropomorphic robot has the five fingers shape apery manipulator of sense of touch is 450 * 200 millimeters
2, weight is 1.5 kilograms.
Each several part illustrates in greater detail as follows:
About two eye coordinate kinetic control systems
1) with the related neutral net of two eye coordinate motion controls.Mainly flow to according to the nerve pathway and the signal of left and right horizontal semicircular canal in the human brain to left and right sides eyeball, find out horizontal semicircular canal, vestibular nuclear, pontine tegmentum net sample nuclear, cerebral cortex vision neck, corpus geniculatum lateral, look plain the signal circulation relationship of examining, changeing between nerve nucleus, outer straight muscle, the interior straight muscle outward, determined and the related neural network model of two eye coordinate motion controls;
2) with the related Mathematical Modeling of two eye coordinate motion controls.Suppose that the optic nerve cell is simple input signal integration organ, the relation of the excitability input signal of nerve cell, inhibition signal, output signal can be represented by weighted sum.Set up and the related Mathematical Modeling of two eye coordinate motion controls according to the neural network model related like this with two eye coordinate motion controls;
3) with the related physical model of two eye coordinate motion controls.Neural network model and Mathematical Modeling according to two eye coordinate motion control associations, optic nerve path and vision control signal to human two are analyzed, find out two optical axis angles of revolution, the relation between sighting target position and the retina error has been determined two vision control signal flow graph;
4) with the related control method of two eye coordinate motion controls.According to two vision control signal flow graph, determine two eye coordinate kinetic control system block diagrams.Again according to having the fact of incomplete integrator between vestibular nuclear, outer shaft nerve nucleus and the movements binocular and left and right sides vestibular is internuclear, signal conductivity between the outer shaft nerve nucleus of the left and right sides is determined the transfer function of each link in this system block diagram, and then determined the control method of two eye coordinate kinematic systems;
5) it is inner two eye coordinate kinetic control systems and visual processes and control software to be installed on the head of the described anthropomorphic robot that has a pair of the five fingers shape sense of touch manipulator and wrist and an arm, allows it finish various operations based on visual processes such as target identification, tracing and positioning, motion detection, hand eye coordination, obstacle are avoided, walking planning in dynamic environment.
The five fingers shape apery manipulator and wrist and the arm that have sense of touch about a pair of
1) structure of the five fingers shape apery manipulator and antennal nerve active mechanism thereof; Mainly based on people's finger and the anatomy of palm and the achievement in research of sense of touch mechanism thereof, determined joint setting, the physical dimension of the five fingers shape apery manipulator and the sense of touch point that is distributed on each finger and the palm disposes and correlation;
2) a pair of has the apery manipulator of 5 fingers and palm; Be provided with and physical dimension according to the joint of determining good the five fingers shape apery manipulator, with the driving mechanism of small-sized DC servomotor as each joint.On each small-sized DC servomotor, install and measure the photoelectric encoder of joint angles, developed the joint control circuit and the servo control algorithm of the five fingers shape apery manipulator;
3) utilize pressure-sensing conductive rubber to make the touch sensor of 5 fingers and palm; Produce the principle of alternating current according to changing behind the pressure-sensing conductive rubber resistance pressurized, with the conductive mechanism of pressure-sensing conductive rubber as each sense of touch point.Design and realized being distributed in the distributed touch sensor on 5 fingers and the palm and detect control circuit;
4) kinematics and the dynamics mathematical model of the five fingers shape apery manipulator; Adopt the propulsion of the five fingers shape apery manipulator and the joint angle track that dynamics is calculated each finger and palm, the check control algolithm.Also adopted inverse kinematics and dynamic analysis in addition each finger and the joint moment of palm and from the reaction force of each contact point etc.;
5) the coordination control algolithm of the five fingers shape apery manipulator; According to coordination control feature and the kinematics and the kinetics equation of the five fingers shape apery manipulator, its Jacobian matrix is carried out LU decomposition and ILX decomposition.Import control point, little field, calculate and derive the joint angle of each joint motions of control and the desired value of angular speed, realized coordination control each finger of the five fingers shape apery manipulator and palm.
About two foot walking machine structures
1) control method is held in both hands in anthropomorphic robot sole counter-force and countertorque Shandong; SERVO CONTROL is held in both hands in the two sufficient walking Shandongs of anthropomorphic robot of (5 on head, 3 in waist, both hands 21 * 2=42, both hands arm and both hands wrist 7 * 2=14,6 * 2=12 of both feet) in order to realize having 76 joints, need on the basis of the forward direction of anthropomorphic robot and inverse kinematics and dynamics mathematical model, design unified robust servo controller on the robot joints space;
2) posture control method of physical restriction such as the counter-force of consideration sole and countertorque; In order to control the contact force of anthropomorphic robot two foots and ground environment, need use the ability of posture control that realizes the anthropomorphic robot health based on the optimum power distribution method of mean square deviation minimum;
3) the real-time center of gravity trace control method of anthropomorphic robot both feet exchange; In order to realize walking stably, the both feet of anthropomorphic robot must periodically land.We adopt discretization inverted swing model to calculate the sole touchdown point of anthropomorphic robot, with realization the zero torque point of any track are followed the tracks of control, in the hope of obtaining the stable walking control of anthropomorphic robot on the kiss the earth of the unknown;
4) inertia that the anthropomorphic robot posture is caused changes insensitive Shandong handful control method; Under the situation that the walking speed owing to anthropomorphic robot accelerates, when forward direction inertia increases it is compensated control;
5) three-dimensional the walking emulation mode and the simulation software in real time of anthropomorphic robot; Come the track of each joint angle of simulation calculation with propulsion and dynamics, body position, the health height is so that the access control algorithm.Adopt inverse kinematics and reverse dynamics simultaneously, the position, speed, the acceleration that calculate each joint angle according to the body position and the posture of anthropomorphic robot are so that analysis of joint moment and from the reaction force on ground;
6) corresponding with said method and the anthropomorphic robot prototype in 76 joints arranged.Wait with the power of DC servo motor, 6DOF and torque sensor, computer system and to realize anthropomorphic robot and control system thereof.
About the man-machine interface platform
1) can control the job description language that anthropomorphic robot moves flexibly, this job description language can realize that anthropomorphic robot body and man-machine interface platform communicate by wireless mode to the control of anthropomorphic robot body by the man-machine interface platform;
2) utilize the concurrency control method of many computers of computer grid technology development, comprise that information sharing, the conflict of many intercomputers avoids, instructs priority to control different threedimensional models in the man-machine interface platform;
3) gravity, frictional force, ground supports power, inertia force etc. are set up Mathematical Modeling to the influence of anthropomorphic robot, and simulate the influence of these environmental activity power to anthropomorphic robot in the man-machine interface platform;
The collision-free Trajectory Planning of Welding of the 4) collision detection algorithm of research anthropomorphic robot health each several part, and realization health each several part;
5) forward of anthropomorphic robot and inverse kinematics model, and in the man-machine interface platform, realize spurring the teaching system that robot, pin are finished exercises and operation with mouse, realize recording and playback of action and operation.
Description of drawings
Fig. 1-the 1st, anthropomorphic robot schematic diagram of the present invention;
Fig. 1-2 is the man-machine interface platform schematic diagram of anthropomorphic robot of the present invention;
Fig. 2 is the input and output vision control signal schematic diagram of two of anthropomorphic robots of the present invention;
Fig. 3 is humanoid robot two eye coordinate kinetic control system block diagrams of the present invention;
Fig. 4 is the five fingers shape apery manipulator design figure of the present invention;
Fig. 5 is that the haptic signal of the five fingers shape apery manipulator of the present invention detects schematic diagram;
Fig. 6 is the configuration structure schematic diagram of the five fingers shape apery manipulator sense of touch point of the present invention;
Fig. 7 is anthropomorphic robot of the present invention (part) installation diagram;
Fig. 8 is anthropomorphic robot of the present invention (part) side view;
Fig. 9 (a) is anthropomorphic robot two sufficient walking test interfaces of the present invention;
Fig. 9 (b) is an anthropomorphic robot body joints test interface of the present invention;
Fig. 9 (c) is an anthropomorphic robot wrist shoulder joint test interface of the present invention.
Among the figure: 1 liang of eye coordinate kinetic control system, 2-1 has the five fingers shape apery machinery left hand of sense of touch, 2-3 left finesse and arm, 2-2 has the five fingers shape apery machinery right hand of sense of touch, 2-4 right finesse and arm, 3-1 two foot walking machine structures (left leg), 3-2 two foot walking machine structures (right leg), 4 servomotor control device, 5 transmission/reception antennas; The 4-1 forefinger, 4-2 middle finger, the 4-3 third finger, 4-4 pinkie, 4-5 thumb, 4-6 palm, front and back cradle head 4-7, left-right rotation joint 4-8.
The specific embodiment
Shown in Fig. 1-1 and Fig. 1-2, anthropomorphic robot of the present invention.
Comprised: 1 liang of eye coordinate kinetic control system, 2-1 has the five fingers shape apery machinery left hand of sense of touch, 2-2 has the five fingers shape apery machinery right hand of sense of touch, 2-3 left finesse and arm, 2-4 right finesse and arm, 3-1 two foot walking machine structures (left leg), 3-2 two foot walking machine structures (right leg), 4 small-sized servomotor control device, 5 transmission/reception antennas, 6 man-machine interface platforms.
Wherein, the basic principle of two eye coordinate kinetic control systems is as follows:
Can know that from human binocular vision nerve pathway the exciting signal that comes from horizontal semicircular canal (HSC) is that the first kind neuron (Type I) with vestibular nuclear (VN) is connected.The part of nerve that comes from the nasal side nethike embrane of eyeball and central socket is neural by converging mutually with a part of nerve of the eyeball central socket of opposite side after the optic chiasma simultaneously, again via the second class neuron (Type II) of the VN of pontine tegmentum net sample nuclear (the NRTP) → opposite side of looking rope nuclear (NOT) → opposite side of opposite side and be connected with it.And the nerve that ear side net mould comes, is connected with the second class neuron of the VN of homonymy by outside bridge nuclear (DLPN) through cerebral cortex vision neighbour (VC) from the outside lacquer shape body (LGN) of homonymy.In addition, beginning excitability pathway till the eyestrings from vestibular nuclear is first kind neuron from VN, arrive the second class neuron pathway and moving eye movement nuclear (the OMN) → interior straight muscle (MR) of the VN of opposite side, and by the outer straight muscle (LR) of outer commentaries on classics nerve nucleus (the AN) → opposite side of opposite side, the AN → inboard of opposite side is from bundle MLF → OMN → formations such as MR.The inhibition pathway is the first kind neuron pathway by the VN of first kind neuron → opposite side of VN, the pathway of the AN of first kind neuron pathway → homonymy of VN, and the pathway of first kind neuron → opposite side OMN of VN constitutes.Therefore, arrive the nerve pathway of left and right sides eyeball and the analysis that signal flows to according to above-mentioned left and right horizontal semicircular canal, can find out horizontal semicircular canal, vestibular nuclear, pontine tegmentum net sample nuclear, cerebral cortex vision neck, corpus geniculatum lateral, look plain nuclear, change nerve nucleus outward, the signal circulation relationship between the outer straight muscle, interior straight muscle, and then draw neural network model with two eye coordinate motion associations.Can set up the Mathematical Modeling of two eye coordinates motion thus.
Define two input and output vision control signals, as shown in Figure 2, the head speed of gyration ω that left and right sides semicircular canal detects
lAnd ω
r, two optical axis angle of revolution E
lAnd E
r, sighting target position
lAnd
r, the sighting target net mould error that left and right sides eyeball is caught is divided R
lAnd R
r, sighting target is looked parallel lines etc.In addition, suppose that nerve cell is simple input signal integration organ, the relation of the excitability input signal of nerve cell, inhibition signal, output signal can be represented by weighted sum.According to above definition and supposition two optic nerve path and vision control signal are analyzed, can be found out two optical axis angles of revolution, the relation between sighting target position and the retina error, and draw two vision control signal flow graph.By the block diagram that can determine two eye coordinate kinetic control systems than signal flow diagram.Internuclear according to having the fact of incomplete integrator and left and right sides vestibular between vestibular nuclear, outer shaft nerve nucleus and the movements binocular again, signal conductivity between the outer shaft nerve nucleus of the left and right sides is determined the transfer function of each link in this system block diagram, and then can obtain the control method that systems are moved in two remainder allocation and transportation, as shown in Figure 3.
Two eye coordinate kinetic control systems as seen from Figure 3.Its specific embodiment is as follows:
1) nerve pathway from the left and right horizontal semicircular canal to left and right sides eyeball and signal are flowed to further analysis of do, find out horizontal semicircular canal, vestibular nuclear, pontine tegmentum net sample nuclear, cerebral cortex vision neck, corpus geniculatum lateral, look plain nuclear, change nerve nucleus outward, the signal circulation relationship between the outer straight muscle, interior straight muscle, and then draw neural network model with two eye coordinate motion associations;
2) by analysis, find out two optical axis angles of revolution, the relation between sighting target position and the retina error to two eye coordinate motion physical models.And, determine the block diagram of two eye coordinate kinetic control systems in conjunction with acquired nerve pathway and signal circulation relationship.Internuclear according to the fact and the left and right sides vestibular that exist incomplete integrator between vestibular nuclear, outer shaft nerve nucleus and the movements binocular again, signal conductivity between the outer shaft nerve nucleus of the left and right sides is determined the transfer function of each link in this control system block diagram, and then obtains the Mathematical Modeling and the control method of two eye coordinate kinetic control systems;
3) use two ccd video cameras, 5 DC servo motors, 1 high speed binary channels image processing board, 1 personal computer to wait the two eye coordinate kinetic control systems that constitute 5DOF.Also be that left and right sides ccd video camera (two eyeballs that are equivalent to anthropomorphic robot) can move upper and lower, left and right, whole system can equally with people's neck rotate simultaneously.And based on following proposal, use said system to test respectively, the Mathematical Modeling that check proposes and the validity of control method and finish every operation: a) watched attentively object and fix based on visual processes, under the situation that neck rotates, check two to being watched attentively the identification of object, follow the trail of and positioning performance; B) neck is fixed, and is watched attentively under the situation of object move left and right, checks two identification, tracking and positioning performances to being watched attentively object; C) when being watched attentively object on two eyeball center lines and near two the time, check the control characteristic of two pupil distance concentrate around one point, as spokes on acis motions; D) when a left side or right eye are input as 0, also be the control characteristic that left eye or right eye are checked the motion of two eye coordinates when closing; E) check two pupil distance concentrate around one point, as spokes on acis athletic meeting disappear the condition that two control characteristic improves or degenerates under what conditions; F) two eye coordinate athletic meeting are out of control under what conditions in check, determine the stable condition of two eye coordinate kinetic control systems; Finish with this system that target identification, tracing and positioning, motion detection, hand eye coordination, obstacle are avoided, walking is planned.
Two eye coordinate kinetic control systems use two ccd video cameras, 5 DC servo motors, 1 high speed binary channels image processing board, 2 blocks of PIO plates, 1 high-performance personal computer to constitute two eye coordinate kinetic control systems of 5DOF.Also be left and right sides ccd video camera (two eyeballs that are equivalent to anthropomorphic robot) can upper and lower, left and right motion and whole device can rotate as people's neck.Two eyeball distances of this device propose and are decided to be 75 millimeters, and the distance between the gyroaxis from the eyeball to the neck is decided to be 85 millimeters, in the hope of consistent with people's eyeball.Read the visual information of ccd video camera by high-performance personal computer, and detect the translational speed of sighting target, send into control computer via the PIO plate again in plane of delineation central visual axis.Detect the DC servo motor angle of revolution with the photoelectric encoder that is installed on each DC servo motor of sighting device.The head speed of gyration can be by drawing behind the photoelectric encoder output signal differential.Import by the reference that the Mathematical Modeling and the control algolithm of proposition can be calculated two eye coordinate kinetic control systems, and control left and right sides ccd video camera upper and lower, left and right motion and whole system and people's the same gyration of neck with the PID controller.Utilize this system can check two eye coordinate kinetic characteristics described in the experimental program, and can finish that sighting target identification, tracing and positioning, motion detection, hand eye coordination, obstacle are avoided, walking is planned.
With the five fingers shape apery manipulator of sense of touch, its specific embodiment is as follows as seen from Figure 4:
1) constructs and the sense of touch function according to people's finger-joint and palm, with the driving mechanism of small-sized DC servomotor as each articulations digitorum manus of the five fingers shape apery manipulator, determine the configuration of sense of touch point on each finger and the palm, and with the conductive mechanism of pressure-sensing conductive rubber as each sense of touch point.Design distributed touch sensor and testing circuit thereof according to each sense of touch measuring control point and conductive mechanism.Following Fig. 5 and shown in Figure 6, its resistance is subjected to and can changes behind the pressure-sensitive conducting rubber pressurized at sense of touch point P place, produces alternating current i, gives AD converter after the process operational amplifier detects corresponding voltage signal Vmn.Send computer to handle by the I/O interface board again.So just can detect, handle and realize the distributed feeling function of the five fingers shape apery manipulator;
2) set up kinematics and the kinetics equation and the Jacobian matrix thereof of the five fingers shape apery manipulator.And Jacobian matrix carried out LU decomposes and ILX decomposes, control so that realize each finger-joint of the five fingers shape apery manipulator and the coordination of palm.Coordination control feature according to the five fingers shape apery manipulator imports control point, little field, and studies its data configuration and expansion Jacobian, calculates and derive the joint angle of each joint motions of control and the desired value of angular speed then;
3), can measure the motion of people's finger and palm and grasping moment etc. in advance as various movement locus generators, and adopt propulsion and dynamics to calculate the track of each joint angle in order to allow the five fingers shape apery manipulator correctly fulfil assignment.Can draw the palm position so also can the access control algorithm with height.Also adopt inverse kinematics and dynamics in addition, the position, speed, the acceleration that calculate each joint angle according to the position and the posture of the five fingers shape apery manipulator are so that analyze each joint moment and from reaction force of each contact point etc.
Constitute 21 basic joints of the five fingers shape apery manipulator with 21 small-sized DC servomotors.With the conductive mechanism of pressure-sensing conductive rubber, and be installed on the reaction force of the surperficial root of the five fingers shape apery manipulator with the detection contact point as touch sensor; Use high-performance personal computer and I/O plate to constitute the control device of the five fingers shape apery manipulator.A computer reads photoelectric encoder and the operational amplifier output of distributed touch sensor, the angle of revolution of detection DC servo motor and the electric current and the voltage change of each distributed sense of touch point that is installed in each articulations digitorum manus place.Send into another Computer Processing again; Reference the input of calculating the five fingers shape apery manipulator control system by the Mathematical Modeling that proposes and control algolithm and the grip of each finger and palm etc. control with the robust control algorithm that the five fingers shape apery manipulator is respectively pointed and the motion of palm; In actual environment, the tactile the five fingers shape of apparatus manipulator finishes and grabs, holds, pinches, presss from both sides, pushes away, draws, inserts, by, the operation of various complexity such as cut, cut, strike, beat, tear, paste, lead, drag, grind, cut, dig, frustrate.
By Fig. 7 and the visible two foot walking machine structures of Fig. 8, its specific embodiment is as follows:
1) the robust SERVO CONTROL of each joint angle when realizing the anthropomorphic robot walking at first needs to set up the same forward direction and an inverse kinematics and a dynamics mathematical model with the anthropomorphic robot in 76 joints with the people.On these model based, design the Shandong handful servo controller that an energy is worked under the inertia fluctuation of anthropomorphic robot.And, obtain robust servo controller unified on the robot joints space by calculating the inertial matrix of anthropomorphic robot;
2) in order to control the contact force of anthropomorphic robot two foots and ground environment, need a kind of reaction force robust controller of design.The mixed-control mode of position-based and power is adopted in the control of each pin of apery device people.And use the ability of posture control that realizes anthropomorphic robot based on the optimum power distribution method of mean square deviation minimum.Under the situation that the walking speed owing to anthropomorphic robot accelerates, when forward direction inertia increases it is compensated control;
3) at single pin or double-legged driving phase, stablize anthropomorphic robot with attitude control.In order to realize walking stably, the both feet of anthropomorphic robot must periodically land.We adopt discretization inverted swing model to calculate the sole touchdown point of anthropomorphic robot, to realize that the zero torque point of any track is followed the tracks of control, also promptly at each sampling period discretization inverted swing model, in the hope of obtaining the stable walking control algolithm of anthropomorphic robot on the kiss the earth of the unknown;
4) walking that comes the emulation anthropomorphic robot with two kinds of methods.First kind is the track that employing propulsion and forward momentum model come each joint angle of simulation calculation, body position, and the health height is so that verify various control algolithms.Second kind is to adopt inverse kinematics and reverse kinetic model, the position, speed, the acceleration that calculate joint angle according to the body position and the posture of anthropomorphic robot, so that analysis of joint moment and from the reaction force on ground, and the walking of measuring the people is as various tracking generators;
5) wait with torque sensor, high-performance computer with the power of 76 DC servo motors, two 6DOFs and realize a same anthropomorphic robot and a control system thereof with 76 joints with the people.And the various algorithms that adopt following experimental program to check to propose above.
For forward direction and inverse kinematics and the kinetic model of setting up anthropomorphic robot, can be used as anthropomorphic robot and have inertia, friction and graviational interaction and do not have the freely falling body manipulator of fixed contact point are set up its kinematics and kinetics equation.On these equations based, can in the joint space of anthropomorphic robot, design the H ∞ robust servo controller that an energy is worked under the inertia fluctuation of anthropomorphic robot according to the mixed sensitivity method for designing with the CAD Autocad.By calculating the inertial matrix of anthropomorphic robot, just can obtain robust servo controller unified on the robot joints space again.
In order to control anthropomorphic robot and environment contact force, the mixed-control mode of position and power can be used to control each pin of anthropomorphic robot.While is according to the parallel equation of motion of the mass cg of anthropomorphic robot and use the ability of posture control that can realize anthropomorphic robot based on the optimum power distribution method of mean square deviation minimum.In order to set up the overall stabilized walking model of anthropomorphic robot, can adopt discretization inverted swing model to calculate the sole touchdown point of anthropomorphic robot, the zero torque point of any track be followed the tracks of the walking control algolithm of controlling and obtaining anthropomorphic robot on the kiss the earth of the unknown with realization.
In addition, the walking that can measure the people in advance is as various tracking generators, and adopt propulsion and forward momentum to calculate track, body position, the health height of joint angle just can the access control algorithm.Also adopt inverse kinematics and reverse dynamics, according to the position that the body position and the posture of anthropomorphic robot are calculated joint angle, speed, acceleration, analysis of joint moment and from the reaction force on ground.This research intends constituting with 12 DC servo motors 12 basic joints of anthropomorphic robot both feet, and the power of two 6DOFs and torque sensor are installed on the heel portion of anthropomorphic robot with the testing environment reaction force.This research also will use 2 high-performance personal computer and PIO plate to constitute the control device of anthropomorphic robot.The height of anthropomorphic robot is decided to be 1.5 meters, and the distance between two shoulders is 0.45 meter, consistent in the hope of with the people.And read the photoelectric encoder that is installed in each joint by high-performance personal computer, detect the angle of revolution of DC servo motor.Send into another computer via the PIO plate again.Import by the reference that the Mathematical Modeling and the control algolithm of proposition can be calculated the anthropomorphic robot control system, and just can control the stable walking of anthropomorphic robot under various contact environments with the control algolithm that proposes.
Man-machine interface as seen from Figure 9, its specific embodiment is as follows:
We will adopt OpenGL 2.0 and VC++6.0 as developing instrument, the man-machine interface platform of exploitation anthropomorphic robot under Windows 2000 Professional, and anthropomorphic robot body and man-machine interface platform communicate by wireless mode.At first, set up gravity, frictional force, ground supports power and inertia force Mathematical Modeling, and convert the expression matrix form that OpenGL can realize to the anthropomorphic robot influence.Pass through C/S model then, man-machine interface platform main program is at server end, work such as be responsible for that force analysis, collision detection, emulation animation are played up, and control instruction is then sent by the emulation client, allow multi-client to send control instruction to emulation server simultaneously, finish parallel control.Then, set up anthropomorphic robot job description language, and use this job description language control robot to make the two hands coordination operation, exercises and operations such as biped walking, the feasibility of testing this job description language.
For the grid computing in the anthropomorphic robot man-machine interface platform, to analyze existing ripe application framework (as MFC), Distributed Calculation instrument (as MPI, ICE) earlier, absorb its architecture advantage, thereby design more healthy and stronger, higher extendible architecture; Utilize existing ripe tool storage room (as Boost, Globus) as much as possible, thereby shorten the construction cycle, improve stability; Implement exploitation then stage by stage, step by step, add new function with the form of module, subsequent development reduces the influence to existed system as far as possible; Development process is followed spin model; Design complete test macro at last.
Claims (4)
1, a kind of anthropomorphic robot that is used for environment-identification and carries out operation is characterized in that comprising:
---rotatable two eye coordinate kinetic control systems are used to finish that target identification, tracing and positioning, motion detection, hand eye coordination, obstacle are avoided, walking planning operation; Be made of two ccd video cameras, five DC servo motors, a high speed binary channels video tablet and two blocks of PIO plates, two ccd video cameras are equivalent to two eyeballs of people, have upper and lower, left and right motion and spinfunction;
---the tactile the five fingers shape of biobelt apery manipulator, be used to finish grab, hold, pinch, press from both sides, push away, draw, insert, by, cut, cut, strike, beat, tear, paste, lead, drag, grind, cut, dig, frustrate operation; The palm and the sense of touch point on five fingers thereof that are distributed in described apery manipulator have 120 * 2=240, its palm and five fingers thereof comprise 21 joints altogether, and both hands have 21 * 2=42 the free degree, a pair of arm and a pair of wrist that 7 * 2=14 the free degree arranged;
---two foot walking machine structures are used to control the walking stably on the ground of the unknown of described anthropomorphic robot; Described two foot walking machine structures are made of waist and two feet, have 15 joints, and waist has 3 frees degree, two feet totally 12 frees degree; The power of two 6DOFs and torque sensor are installed on the root of described two feet, are used for testing environment reaction force and moment;
---the man-machine interface platform is support with the computer grid, by calculating relevant Mathematical Modeling and control algolithm, controls described anthropomorphic robot and realizes motor function, perceptional function, function and thinking, operation function, man-machine interaction; For providing interface, simplifies even transparent threedimensional model emulation, job description language, stressed effect analysis, visual motion and operation effectiveness, both hands and double-legged teaching, limbs collision detection, collision-free Trajectory Planning of Welding described anthropomorphic robot in network and cluster and grid.
2, anthropomorphic robot according to claim 1, it is characterized in that it highly equals described two foot walking machine structure height and adds described two eye coordinate kinetic control system height, be 1.5 meters, the distance between 60 kilograms of body weight, two shoulders is 0.45 meter, and walking speed is greater than 2.0 kilometers/hour.
3, according to claim 1 and 2 described anthropomorphic robots, it is characterized in that in the described rotatable two eye coordinate kinetic control systems that two eyeball distances are set at 75 millimeters, the distance between the gyroaxis from the eyeball to the neck is decided to be 85 millimeters.
4, anthropomorphic robot according to claim 3 is characterized in that the appearance and size that described a pair of has a five fingers shape apery manipulator of sense of touch is 450 * 200 millimeters
2, weight is 1.5 kilograms.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200610123018 CN1947960A (en) | 2006-10-26 | 2006-10-26 | Environment-identification and proceeding work type real-man like robot |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 200610123018 CN1947960A (en) | 2006-10-26 | 2006-10-26 | Environment-identification and proceeding work type real-man like robot |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN1947960A true CN1947960A (en) | 2007-04-18 |
Family
ID=38017635
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200610123018 Pending CN1947960A (en) | 2006-10-26 | 2006-10-26 | Environment-identification and proceeding work type real-man like robot |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN1947960A (en) |
Cited By (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2010070666A1 (en) * | 2008-12-19 | 2010-06-24 | Manush Lochan Madan | Robot building platform |
| CN101976518A (en) * | 2010-10-14 | 2011-02-16 | 吉林大学 | Intelligent walking robot teaching aid |
| CN102896630A (en) * | 2011-07-25 | 2013-01-30 | 索尼公司 | Robot device, method of controlling the same, computer program, and robot system |
| CN103753566A (en) * | 2014-01-24 | 2014-04-30 | 成都万先自动化科技有限责任公司 | Airport cleaning service robot |
| CN104971502A (en) * | 2015-07-22 | 2015-10-14 | 黑龙江大学 | Multi-driving autonomous networked boxing model robot system and control method thereof |
| CN104971501A (en) * | 2015-07-22 | 2015-10-14 | 黑龙江大学 | Multi-driving autonomous boxing model robot system and control method thereof |
| CN104971503A (en) * | 2015-07-22 | 2015-10-14 | 黑龙江大学 | Single-driving autonomous remote control network boxing robot system and control method thereof |
| CN105013186A (en) * | 2015-07-22 | 2015-11-04 | 黑龙江大学 | Multi-drive automatic-remote-control networked boxing model robot system and control method thereof |
| CN105013185A (en) * | 2015-07-22 | 2015-11-04 | 黑龙江大学 | Single-driven automatic networked boxing model robot system and control method thereof |
| CN105031933A (en) * | 2015-07-22 | 2015-11-11 | 黑龙江大学 | Single-driving remote-control boxing model robot system and control method of robot system |
| CN106737578A (en) * | 2016-12-15 | 2017-05-31 | 山东大学 | A kind of quadruped robot |
| CN107030704A (en) * | 2017-06-14 | 2017-08-11 | 郝允志 | Educational robot control design case based on neuroid |
| CN107127760A (en) * | 2017-07-12 | 2017-09-05 | 清华大学 | A kind of track combined anthropomorphic robot of foot |
| CN110434866A (en) * | 2019-08-02 | 2019-11-12 | 吕梁学院 | A kind of intelligent nursing device and its control method |
-
2006
- 2006-10-26 CN CN 200610123018 patent/CN1947960A/en active Pending
Cited By (18)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2010070666A1 (en) * | 2008-12-19 | 2010-06-24 | Manush Lochan Madan | Robot building platform |
| CN101976518A (en) * | 2010-10-14 | 2011-02-16 | 吉林大学 | Intelligent walking robot teaching aid |
| CN101976518B (en) * | 2010-10-14 | 2012-05-16 | 吉林大学 | Intelligent walking robot teaching aid |
| CN102896630A (en) * | 2011-07-25 | 2013-01-30 | 索尼公司 | Robot device, method of controlling the same, computer program, and robot system |
| CN102896630B (en) * | 2011-07-25 | 2016-06-01 | 索尼公司 | Robot device and control method, computer program and robot system |
| CN103753566A (en) * | 2014-01-24 | 2014-04-30 | 成都万先自动化科技有限责任公司 | Airport cleaning service robot |
| CN105013186A (en) * | 2015-07-22 | 2015-11-04 | 黑龙江大学 | Multi-drive automatic-remote-control networked boxing model robot system and control method thereof |
| CN104971503A (en) * | 2015-07-22 | 2015-10-14 | 黑龙江大学 | Single-driving autonomous remote control network boxing robot system and control method thereof |
| CN104971501A (en) * | 2015-07-22 | 2015-10-14 | 黑龙江大学 | Multi-driving autonomous boxing model robot system and control method thereof |
| CN105013185A (en) * | 2015-07-22 | 2015-11-04 | 黑龙江大学 | Single-driven automatic networked boxing model robot system and control method thereof |
| CN105031933A (en) * | 2015-07-22 | 2015-11-11 | 黑龙江大学 | Single-driving remote-control boxing model robot system and control method of robot system |
| CN104971502A (en) * | 2015-07-22 | 2015-10-14 | 黑龙江大学 | Multi-driving autonomous networked boxing model robot system and control method thereof |
| CN106737578A (en) * | 2016-12-15 | 2017-05-31 | 山东大学 | A kind of quadruped robot |
| CN106737578B (en) * | 2016-12-15 | 2018-11-13 | 山东大学 | A kind of quadruped robot |
| CN107030704A (en) * | 2017-06-14 | 2017-08-11 | 郝允志 | Educational robot control design case based on neuroid |
| CN107127760A (en) * | 2017-07-12 | 2017-09-05 | 清华大学 | A kind of track combined anthropomorphic robot of foot |
| CN110434866A (en) * | 2019-08-02 | 2019-11-12 | 吕梁学院 | A kind of intelligent nursing device and its control method |
| CN110434866B (en) * | 2019-08-02 | 2023-06-23 | 吕梁学院 | Intelligent nursing device and control method thereof |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN1947960A (en) | Environment-identification and proceeding work type real-man like robot | |
| US11772266B2 (en) | Systems, devices, articles, and methods for using trained robots | |
| Liu et al. | High-fidelity grasping in virtual reality using a glove-based system | |
| Ott et al. | A humanoid two-arm system for dexterous manipulation | |
| Ayusawa et al. | Motion retargeting for humanoid robots based on simultaneous morphing parameter identification and motion optimization | |
| Mick et al. | Reachy, a 3D-printed human-like robotic arm as a testbed for human-robot control strategies | |
| Bartsch et al. | Development and control of the multi-legged robot mantis | |
| CN102814814A (en) | Kinect-based man-machine interaction method for two-arm robot | |
| CN115469576B (en) | Teleoperation system based on human-mechanical arm heterogeneous motion space hybrid mapping | |
| Vlădăreanu et al. | Complex walking robot kinematics analysis and PLC multi-tasking control | |
| Li et al. | An efficient motion generation method for redundant humanoid robot arm based on the intrinsic principles of human arm motion | |
| Falck et al. | DE VITO: A dual-arm, high degree-of-freedom, lightweight, inexpensive, passive upper-limb exoskeleton for robot teleoperation | |
| Park et al. | A whole-body integrated AVATAR system: Implementation of telepresence with intuitive control and immersive feedback | |
| Ji et al. | Motion trajectory of human arms based on the dual quaternion with motion tracker | |
| Ponce et al. | Modeling and simulation for designing a line walking chameleon-like legged robot | |
| Deng et al. | Human-like posture correction for seven-degree-of-freedom robotic arm | |
| Tanzini et al. | New interaction metaphors to control a hydraulic working machine's arm | |
| Abut et al. | Interface design and performance analysis for a haptic robot | |
| Ellenberg et al. | A common interface for humanoid simulation and hardware | |
| Qin et al. | Design and development of a cable-driven elephant trunk robot with variable cross-sections | |
| CN115781666A (en) | Control method for robot whole body simulation system | |
| Xie et al. | Human-robot coupling dynamic modeling and analysis for upper limb rehabilitation robots | |
| CN202507280U (en) | Humanoid robot | |
| Leng et al. | Parameter design of biped robot motion system based on multi-objective optimization | |
| Bing et al. | System design and experiment of bionics robotic arm with humanoid characteristics |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C12 | Rejection of a patent application after its publication | ||
| RJ01 | Rejection of invention patent application after publication |
Open date: 20070418 |