CN109514521A - The servo operation and its method of manpower collaboration Dextrous Hand based on multi-information fusion - Google Patents

The servo operation and its method of manpower collaboration Dextrous Hand based on multi-information fusion Download PDF

Info

Publication number
CN109514521A
CN109514521A CN201811549841.2A CN201811549841A CN109514521A CN 109514521 A CN109514521 A CN 109514521A CN 201811549841 A CN201811549841 A CN 201811549841A CN 109514521 A CN109514521 A CN 109514521A
Authority
CN
China
Prior art keywords
manpower
dextrous hand
finger
gesture
joint
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.)
Granted
Application number
CN201811549841.2A
Other languages
Chinese (zh)
Other versions
CN109514521B (en
Inventor
黄英
张阳阳
朱文瑾
刘平
刘彩霞
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hefei University of Technology
Original Assignee
Hefei University of Technology
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Hefei University of Technology filed Critical Hefei University of Technology
Priority to CN201811549841.2A priority Critical patent/CN109514521B/en
Publication of CN109514521A publication Critical patent/CN109514521A/en
Application granted granted Critical
Publication of CN109514521B publication Critical patent/CN109514521B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • BPERFORMING OPERATIONS; TRANSPORTING
    • B25HAND TOOLS; PORTABLE POWER-DRIVEN TOOLS; MANIPULATORS
    • B25JMANIPULATORS; CHAMBERS PROVIDED WITH MANIPULATION DEVICES
    • B25J3/00Manipulators of master-slave type, i.e. both controlling unit and controlled unit perform corresponding spatial movements
    • B25J3/04Manipulators of master-slave type, i.e. both controlling unit and controlled unit perform corresponding spatial movements involving servo mechanisms

Abstract

The servo operation and its method for the manpower collaboration Dextrous Hand based on multi-information fusion that the invention discloses a kind of, system include: manpower flexible extensible sensor, Dextrous Hand flexible extensible sensor, flexible force-touch sensor, wireless transport module, Dextrous Hand control module and computer module;Computer module includes: gesture-capture module, gesture mapping block and seized condition identification module.Flexible extensible sensor arrangement by being worn on the flexible gloves joint of manpower by the present invention in the joint of robot delicate and being arranged in, and force-touch sensor is arranged at Dextrous Hand finger tip, it is able to achieve the accurate collaboration of manpower and Dextrous Hand operating gesture and perceives Dextrous Hand to the seized condition of target object, backer's hand control Dextrous Hand carries out stablizing crawl target object.

Description

The servo operation and its method of manpower collaboration Dextrous Hand based on multi-information fusion
Technical field
The present invention relates to human-computer interaction, principal and subordinate's cooperating technical field more particularly to a kind of based on multi-information fusion The servo operation and its method of manpower collaboration Dextrous Hand.
Technical background
The purpose that manpower cooperates with Dextrous Hand servo operation is that manpower is cooperateed with robot delicate, by robot spirit Dab hand reappears the movement of manpower in distal end.General servo operation robot is used in the environment that people is difficult to when participating in the cintest, such as deep The fields such as extra large detection, the outer space.The operation of manpower cooperating robot's Dextrous Hand is the interactive mode of a kind of advanced people and robot, Its basic conception: on the one hand by the multisensor syste around manpower and Dextrous Hand that the motion information of manpower and Dextrous Hand is real When detect, and the gesture collaboration of manpower and Dextrous Hand is realized by dexterous hand controls;On the other hand by being mounted on Dextrous Hand On multisensor syste and environmental interaction, and by interaction mode and information Real-time Feedback to manpower, so that people makes reasonably Adjust decision.
Precisely cooperate with for manpower and Dextrous Hand gesture is that manpower cooperates with critical issue in Dextrous Hand servo operation.Mesh Preceding gesture-capture equipment is view-based access control model sensor and data glove.The information collection precision of view-based access control model sensor has Limit, usually can not accurately capture the gesture of manpower.The non-flexible sensor of complicated multiclass multi-quantity is to the wearable of data glove Property and measurement gesture information accuracy bring challenge.Gesture-capture and gesture based on myoelectricity, the contour coupling information of brain electricity There are certain deficiencies in terms of information decoupling for collaboration, temporarily can not achieve accurately measuring for complicated gesture information, can not be accurate And comprehensively the gesture of complicated manpower and Dextrous Hand is captured and mapped.In addition, how to make Dextrous Hand accurately perceive with How the interaction mode of environment improves the dexterity of manpower collaboration Dextrous Hand servo operation, is the key that current system research is asked Topic.
Summary of the invention
It is an object of the present invention to overcome the deficiencies in the prior art, propose a kind of manpower collaboration spirit based on multi-information fusion The servo operation and its method of dab hand, to be able to achieve the accurate collaboration and perception spirit of manpower and Dextrous Hand operating gesture Dab hand carries out stablizing crawl target object to the seized condition of target object so that backer's hand controls Dextrous Hand.
The present invention to achieve the above object of the invention, adopts the following technical scheme that
A kind of the characteristics of servo operation of manpower collaboration Dextrous Hand based on multi-information fusion of the present invention includes: manpower Flexible extensible sensor, Dextrous Hand flexible extensible sensor, flexible force-touch sensor, wireless transport module, Dextrous Hand Control module and computer module;
The computer module includes: gesture-capture module, gesture mapping block and seized condition identification module;
The manpower flexible extensible sensor is arranged in the joint of flexible gloves and is worn on manpower, for acquiring people Manpower joint angles feature when hand operates, and the gesture-capture module is sent to by the wireless transport module;
The gesture-capture module carries out gesture-capture to the manpower joint information, obtains manpower gesture information and transmits To the gesture mapping block;
The joint of robot delicate is arranged in the Dextrous Hand flexible extensible sensor, for acquiring Dextrous Hand behaviour Delicate finger-joint angle character when making, and the gesture-capture module is sent to by the wireless transport module;
The gesture-capture module carries out gesture-capture to the delicate finger-joint information, obtains Dextrous Hand gesture information simultaneously Transmit the gesture mapping block;
The gesture mapping block manpower gesture information and Dextrous Hand gesture information based on the received calculate Dextrous Hand Required expectation driving value, to establish gesture mapping off-line model;
The gesture mapping block obtains the current information of manpower gesture, obtains Dextrous Hand using the gesture mapping model Current expectation driving value and be sent to the Dextrous Hand control module for drive Dextrous Hand grab target object;
On the inside of the flexibility force-touch sensor is arranged at the finger tip of robot delicate, for acquiring Dextrous Hand crawl The pressure information of target object, and the seized condition identification module is passed to by the wireless transport module;
It include: sliding mode, vibrating state and stable state in seized condition identification module setting seized condition;Again According to the received pressure information of institute, offline Classification and Identification model is established to the seized condition using support vector machines;
The seized condition identification module obtains current pressure information using the flexible force-touch sensor, and utilizes institute Seized condition identified off-line module is stated, current seized condition is obtained and is judged, if current seized condition is stable state, Stop the nearly articulations digitorum manus of driving clever hand finger, save and show the present driving value of Dextrous Hand, if current seized condition is sliding State then obtains next expectation driving value of Dextrous Hand using the gesture mapping block and is sent to the Dextrous Hand control Module;If current seized condition is vibrating state, manpower operation and Dextrous Hand operation are re-started, and resurvey and catch It catches.
The characteristics of manpower collaboration Dextrous Hand servo operation of the present invention, lies also in: the flexible force tactile sensing Device is three-decker, is using the melamine sponge after being carbonized as intermediate sensitive layer, in the upper and lower surfaces of the sensitive layer It is respectively arranged with upper electrode layer, lower electrode layer.
The upper electrode layer and the lower electrode layer use copper plate electrode, are respectively adhered on the upper and lower surfaces of sensitive layer.
The characteristics of a kind of manpower based on multi-information fusion of the present invention cooperates with Dextrous Hand servo operation method is by following step It is rapid to carry out:
Joint and the progress for being worn on the flexible gloves of manpower is arranged in manpower flexible extensible sensor by step 1 Manpower operation, recycles CCD camera and resistance instrument to demarcate the joint angles and manpower resistance value of manpower, obtains manpower Mapping coefficient M;
The joint of robot delicate is arranged in Dextrous Hand flexible extensible sensor and carries out Dextrous Hand operation, then The joint angles and Dextrous Hand resistance value of Dextrous Hand are demarcated using CCD camera and resistance instrument, obtain Dextrous Hand mapping Coefficient M ';
Step 2 utilizes the people of j-th of joint on manpower flexible extensible sensor acquisition i-th of finger of manpower Swivel of hand angle characterUtilize j-th of joint on Dextrous Hand flexible extensible sensor acquisition i-th of finger of Dextrous Hand The delicate finger-joint angle character at place
Step 3, the manpower gesture for being obtained j-th of joint on i-th of finger of manpower respectively using formula (1) and formula (2) are believed BreathWith the Dextrous Hand gesture information of j-th of joint on i-th of finger of Dextrous Hand
Step 4 obtains the calibration relationship between manpower gesture information and Dextrous Hand gesture information using formula (3), to build Vertical gesture maps off-line model:
Step 5, the nearly articulations digitorum manus of five fingers for controlling Dextrous Hand respectively using five direct current generators, and by adjusting PWM The angular velocity of satellite motion of the nearly articulations digitorum manus of five fingers of duty ratio T and driving time t the control Dextrous Hand of wave;
The nearly articulations digitorum manus driving value PWM of i-th of finger of Dextrous Hand in n-th of PWM wave duty cycle sequence is obtained using formula (4) The duty ratio T of wavei n:
It obtains being bent in the nearly articulations digitorum manus of i-th of finger of Dextrous Hand of n-th of PWM wave duty cycle sequence using formula (5) maximum Time t consumed in angle pi/2i n:
ti niTi n (5)
In formula (5), μiFor duty ratio Ti nWith time ti nBetween calibration coefficient, n be PWM wave duty cycle sequence number;
Articulations digitorum manus driving PWM wave close in i-th of finger of Dextrous Hand of n-th of PWM wave duty cycle sequence is obtained using formula (6) Duty ratio Ti nThe angular speed w of articulations digitorum manus close with i-th of finger of Dextrous Handi nBetween calibration relationship:
In formula (6),For in the compensation system of the nearly articulations digitorum manus of i-th of finger of Dextrous Hand of n-th of PWM wave duty cycle sequence Number;
Step 6, definition current time are t, and initialize t=1;
Step 7, the manpower joint angle that j-th of joint on current i-th of finger is respectively obtained using formula (7) and formula (8) Spend the variable quantity of featureWith the variation delta Θ of the current information of the manpower gesture of j-th of joint on i-th of fingeri j:
In formula (7) and formula (8),Indicate that the manpower joint angles of j-th of joint on i-th of finger of t moment are special Sign,The manpower joint angles feature for indicating j-th of joint on i-th of finger of t-1 moment, as t=1,For the initial resistivity value of the flexible extensible sensor in i-th of manpower, j-th of finger joint;Indicate t moment The current information of the manpower gesture of j-th of joint on i-th of finger,It indicates on i-th of finger of t-1 moment j-th The current information of the manpower gesture of joint;As t=1,
Step 8, the current expectation driving time Δ t that i-th of finger of Dextrous Hand is obtained using formula (9)i:
Step 9, the current duty cycle that the nearly articulations digitorum manus driving PWM wave of i-th of finger of t moment Dextrous Hand is obtained using formula (10) Ti n(t), so that Dextrous Hand be driven to grab target object:
In formula (10), Ti n(t-1) and Ti n(t-2) i-th of finger of t-1 moment and t-2 moment Dextrous Hand is respectively indicated closely to refer to The duty ratio of joint drive PWM wave, as t=1, Ti n(t-1) and Ti nIt (t-2) is the initial estimate of set duty ratio;For the force-touch sensor arranged on i-th of finger of Dextrous Hand t moment resistance value,It is the i-th of Dextrous Hand The initial resistivity value for the force-touch sensor arranged on a finger;
On the inside of step 10, flexible force-touch sensor are arranged at the finger tip of robot delicate, for acquiring Dextrous Hand Grab the pressure information of target object;
Step 11, setting seized condition include: sliding mode, vibrating state and stable state;Further according to the received pressure of institute Force information establishes offline Classification and Identification model to the seized condition using support vector machines;
Step 12 is obtained current pressure information using the flexible force-touch sensor, and is known using the offline classification Other model obtains current seized condition and is judged, if current seized condition is stable state, stops driving Dextrous Hand hand Refer to nearly articulations digitorum manus, saves and show the current expectation driving time Δ t of i-th of finger of Dextrous HandiWith i-th of t moment Dextrous Hand The current duty cycle T of the nearly articulations digitorum manus driving PWM wave of fingeri n(t) and seized condition;If current seized condition is sliding mode, T+1 is enabled to be assigned to t;And execute step 7;If current seized condition is vibrating state, 6 are thened follow the steps.
Compared with the prior art, the beneficial effects of the present invention are embodied in:
1. the present invention flexible multi-sensor information is merged, by by flexible extensible sensor arrangement in robot It the joint of Dextrous Hand and is arranged in flexible gloves joint and is worn on manpower, and force-touch sensor is arranged in dexterity At finger tip, precise measurement manpower is when controlling Dextrous Hand, the gesture information of manpower and Dextrous Hand, through computer analysis and processing, Obtain the expectation driving value of Dextrous Hand, Dextrous Hand control module passed to by wireless transport module, in real time carry out manpower and The gesture interaction of Dextrous Hand realizes the real-time mapping of manpower and Dextrous Hand gesture;Dextrous Hand is perceived by force-touch sensor Pressure information between target object realizes the judgement to target object seized condition, and realizes Dextrous Hand grasp speed Adaptive adjustment, realize manpower and Dextrous Hand cooperating, and crawl target object that can be stable.
2. measuring Dextrous Hand present invention introduces force-touch sensor by force-touch sensor and grabbing the pressure between object Force information, manpower control Dextrous Hand grab the target object stage, by by force-touch sensor information be sent to computer into Row information processing, analysis obtain Dextrous Hand to the seized condition of target object, realize and stablize crawl to target object.
3. the present invention carries out analytical calculation, adaptive tune according to the pressure information of force-touch sensor, by computer The angular velocity of satellite motion of the nearly articulations digitorum manus of whole Dextrous Hand, ensure that the safety and reliability of crawl target object, realize adaptive The purpose of micro- manipulation.
4. flexible force-touch sensor of the invention, is using melamine sponge as matrix, copper sheet is as electrode.Melamine Amine sponge filled silicon rubber after high temperature cabonization, using copper sheet as upper/lower electrode;It is carbonized in different melamine sponges Under degree and different silicon rubber filling degree, sensor has different ranges and sensitivity, supports under different pressures environment Requirement to transducer range.
Detailed description of the invention
Fig. 1 is flexible extensible sensor plane structure chart in the prior art;
Fig. 2 is the flexible force-touch sensor plane structure chart of the present invention;
Fig. 3 is present system schematic diagram;
Fig. 4 is control system figure of the present invention;
Figure label: 1 flexible force-touch sensor;1a sensitive material;1b sensitive material;1c is PU matrix;2 is flexible Force-touch sensor;2a upper electrode layer;2b sensitive layer;2c lower electrode layer.
Specific embodiment
In the present embodiment, a kind of servo operation of the manpower collaboration Dextrous Hand based on multi-information fusion, comprising: manpower Flexible extensible sensor, Dextrous Hand flexible extensible sensor, flexible force-touch sensor, wireless transport module, Dextrous Hand Control module and computer module;
Computer module includes: gesture-capture module, gesture mapping block and seized condition identification module;
Manpower flexible extensible sensor is arranged in the joint of flexible gloves and is worn on manpower, each finger setting two A flexible extensible sensor, as shown in figure 4, for acquiring manpower joint angles feature when manpower operation, and by wireless Transmission module is sent to gesture-capture module;
Gesture-capture module carries out gesture-capture to manpower joint information, obtains manpower gesture information and pass to gesture reflecting Penetrate module;
The joint of robot delicate is arranged in Dextrous Hand flexible extensible sensor, when for acquiring Dextrous Hand operation Delicate finger-joint angle character, and gesture-capture module is sent to by wireless transport module;
Gesture-capture module carries out gesture-capture to delicate finger-joint information, obtains Dextrous Hand gesture information and transmits gesture Mapping block;
Gesture mapping block manpower gesture information and Dextrous Hand gesture information based on the received calculate needed for Dextrous Hand Expectation driving value, thus establish gesture mapping off-line model;
Gesture mapping block obtains the current information of manpower gesture, works as early period using gesture mapping model acquisition Dextrous Hand It hopes driving value and is sent to Dextrous Hand control module for driving Dextrous Hand to grab target object;
On the inside of flexible force-touch sensor is arranged at the finger tip of robot delicate, for acquiring Dextrous Hand crawl target The pressure information of object, and seized condition identification module is passed to by wireless transport module;
It include: sliding mode, vibrating state and stable state in seized condition identification module setting seized condition;Further according to The received pressure information of institute, establishes offline Classification and Identification model to seized condition using support vector machines;
Seized condition identification module obtains current pressure information using flexible force-touch sensor, and using seized condition from Line identification module obtains current seized condition and is judged, if current seized condition is stable state, it is dexterous to stop driving The nearly articulations digitorum manus of hand finger, saves and shows the present driving value of Dextrous Hand, if current seized condition is sliding mode, utilizes hand Gesture mapping block obtains next expectation driving value of Dextrous Hand and is sent to Dextrous Hand control module;If current seized condition is Vibrating state then re-starts manpower operation and Dextrous Hand operation, and resurveys and capture.
In the present embodiment, as shown in Fig. 2, flexible force-touch sensor is three-decker, it is the melamine sea with carbonization The continuous sensitive layer 2b as centre, is respectively arranged with upper electrode layer 2a, lower electrode layer 2c in the upper and lower surfaces of sensitive layer 2b.
Upper electrode layer 2a and lower electrode layer 2c uses copper plate electrode, is respectively adhered on the upper and lower surfaces of sensitive layer.Sensitive layer It obtains: melamine sponge being put into tube furnace, in a nitrogen atmosphere 800 DEG C high temperature pyrolysis 2 hours as follows, obtained The melamine sponge of carbonization, nitrogen flow rate are 2 DEG C/min;Silicon rubber is dissolved in stone brain according to mass volume ratio 1g:10mL In oil, silicone rubber solution is obtained;Finally the melamine sponge of carbonization is impregnated into 10~15min in silicone rubber solution, taken out After drying, that is, obtain sensitive layer;By the loading of different carbonization time and silicon rubber, the range of adjustable sensor and spirit Sensitivity meets the requirement under different application range scene.
The structure type of manpower flexible extensible sensor and Dextrous Hand flexible extensible sensor is referring to patent CN2017112291961, plane structure chart are as shown in Figure 1.
In the present embodiment, it is a kind of based on multi-information fusion manpower collaboration Dextrous Hand servo operation method be as follows It carries out:
Joint and the progress for being worn on the flexible gloves of manpower is arranged in manpower flexible extensible sensor by step 1 Manpower operation, operator dress data glove, and gloves is kept to be attached at manpower, keep stablizing.Recycle CCD camera and electricity Resistance instrument demarcates the joint angles and manpower resistance value of manpower, and CCD camera is directed at the side of finger, and manpower is slowly bent Finger, manpower keep stable as far as possible, obtain the ccd image of uniform location, transfer data to computer, by image procossing, The joint angles of manpower are obtained, the mapping coefficient M between manpower joint angles and sensor electrical information is calculated;
The joint of robot delicate is arranged in Dextrous Hand flexible extensible sensor and carries out Dextrous Hand operation, then The joint angles and Dextrous Hand resistance value of Dextrous Hand are demarcated using CCD camera and resistance instrument, obtain Dextrous Hand mapping Coefficient M ';
Step 2 is closed using the manpower of j-th of joint on manpower flexible extensible sensor acquisition i-th of finger of manpower Save angle characterUtilize the dexterity of j-th of joint on Dextrous Hand flexible extensible sensor acquisition i-th of finger of Dextrous Hand Swivel of hand angle character
Step 3, the manpower gesture for being obtained j-th of joint on i-th of finger of manpower respectively using formula (1) and formula (2) are believed BreathWith the Dextrous Hand gesture information of j-th of joint on i-th of finger of Dextrous Hand
Step 4 obtains the calibration relationship between manpower gesture information and Dextrous Hand gesture information using formula (3), to build Vertical gesture maps off-line model:
Step 5, the nearly articulations digitorum manus of five fingers for controlling Dextrous Hand respectively using five direct current generators, and by adjusting PWM The angular velocity of satellite motion of the nearly articulations digitorum manus of five fingers of duty ratio T and driving time t the control Dextrous Hand of wave;
The nearly articulations digitorum manus driving value PWM of i-th of finger of Dextrous Hand in n-th of PWM wave duty cycle sequence is obtained using formula (4) The duty ratio T of wavei n:
It obtains being bent in the nearly articulations digitorum manus of i-th of finger of Dextrous Hand of n-th of PWM wave duty cycle sequence using formula (5) maximum Time t consumed in angle pi/2i n:
ti niTi n (5)
In formula (5), μiFor duty ratio Ti nWith time ti nBetween calibration coefficient, n be PWM wave duty cycle sequence number;
Articulations digitorum manus driving PWM wave close in i-th of finger of Dextrous Hand of n-th of PWM wave duty cycle sequence is obtained using formula (6) Duty ratio Ti nThe angular speed w of articulations digitorum manus close with i-th of finger of Dextrous Handi nBetween calibration relationship:
In formula (6),For in the compensation system of the nearly articulations digitorum manus of i-th of finger of Dextrous Hand of n-th of PWM wave duty cycle sequence Number;
Step 6, definition current time are t, and initialize t=1;
Step 7, the manpower joint angle that j-th of joint on current i-th of finger is respectively obtained using formula (7) and formula (8) Spend the variable quantity of featureWith the variation delta Θ of the current information of the manpower gesture of j-th of joint on i-th of fingeri j:
In formula (7) and formula (8),Indicate the manpower joint angles feature of j-th of joint on i-th of finger of t moment,The manpower joint angles feature for indicating j-th of joint on i-th of finger of t-1 moment, as t=1, For the initial resistivity value of the flexible extensible sensor in i-th of manpower, j-th of finger joint;Indicate i-th of hand of t moment Refer to the current information of the manpower gesture of upper j-th of joint,Indicate j-th of joint on i-th of finger of t-1 moment Manpower gesture current information;As t=1,
Step 8, the current expectation driving time Δ t that i-th of finger of Dextrous Hand is obtained using formula (9)i:
Step 9, the current duty cycle that the nearly articulations digitorum manus driving PWM wave of i-th of finger of t moment Dextrous Hand is obtained using formula (10) Ti n(t), so that Dextrous Hand be driven to grab target object;The speed of different PWM duty cycle control robot delicate finger-joint movements Degree is different, in the different operational phases, needs different service speeds, controls robot delicate with this and carries out accurate dexterity Operation.
In formula (10), Ti n(t-1) and Ti n(t-2) i-th of finger of t-1 moment and t-2 moment Dextrous Hand is respectively indicated closely to refer to The duty ratio of joint drive PWM wave, as t=1, Ti n(t-1) and Ti nIt (t-2) is the initial estimate of set duty ratio;For the force-touch sensor arranged on i-th of finger of Dextrous Hand t moment resistance value,It is the i-th of Dextrous Hand The initial resistivity value for the force-touch sensor arranged on a finger;Controlling model shown in formula (10) shows its heredity and note The property recalled, the state parameter of the first two time series directly affect future time sequential parameter, improve Dextrous Hand microoperation Dexterity.
When manpower collaboration Dextrous Hand is in the microoperation stage, finger tip linear velocity more preferably embodies micro- behaviour than joint angular speed The degree of work passes through the finger tip linear velocity of following procedure calculating robot's Dextrous Hand.
Movement velocity based on D-H algorithm and robot direct kinematics theoretical calculation finger tip, in D-H coordinate system, coordinate It is that conversion between i and coordinate system k passes through formula (11) and realizes.
It is transformed into shown in the transition matrix such as formula (12) under basis coordinates system, in formula (12), such as c234=cos (θ23+ θ4)。
Shown in the finger tip coordinate such as formula (13) of robot delicate.
Linear velocity of the finger tip under data of short-time series can be approximate with microvariations amount, as shown in formula (14), is slightly variable according to this Change amount tracks the speed movement status of robot Dextrous Hand finger tip, for assisting Dextrous Hand microcontroller.
On the inside of step 10, flexible force-touch sensor are arranged at the finger tip of robot delicate, for acquiring Dextrous Hand Grab the pressure information of target object;
Step 11, setting seized condition include: sliding mode, vibrating state and stable state;Further according to the received pressure of institute Force information establishes offline Classification and Identification model to seized condition using support vector machines;Detailed process is as follows.
Step 1: data normalization is handled;
Step 2: building training sample set and test sample collection;
Step 3: training sample is instructed in the support vector machines that formula (15), formula (16), formula (17), formula (18) indicate Practice, establishes disaggregated model.
Constraint condition are as follows:
In formula (15)-formula (18), ωkIndicate the normal vector of k-th of Optimal Separating Hyperplane, δiIndicate i-th of classification mark of sample Label, k=1,2 ..., K;The wherein bias term of b presentation class hyperplane, c are penalty factor, ξiFor i-th of relaxation factor.
Step 4: the dual form of support vector machines is sought using Lagrange multiplier, support vector machines decision function such as formula (19) shown in:
Step 5: testing using test sample support vector machines, the categorical attribute of sample is obtained, is realized to crawl The exact classification of state.Penalty factor c in multi-class support vector machine is a free parameter, and the selection of parameter is for classifier Performance have a great impact.Using the available parameter of 20 times of cross-validation methods, penalty factor c can grid 0.1, 0.5,1,5,10 } optimized parameter is chosen in.The 60%-90% of total sample is chosen as training sample set, remaining is as test Sample, choose wherein optimal test scale parameter as system parameter.
Step 12 is obtained current pressure information using flexible force-touch sensor, and utilizes offline Classification and Identification model, is obtained To current seized condition and judged, if current seized condition is stable state, stops that clever hand finger is driven closely to refer to pass Section, saves and shows the current expectation driving time Δ t of i-th of finger of Dextrous HandiClosely refer to i-th of finger of t moment Dextrous Hand The current duty cycle T of joint drive PWM wavei n(t) and seized condition;If current seized condition is sliding mode, t+1 assignment is enabled To t;And execute step 7;If current seized condition is vibrating state, 6 are thened follow the steps.
Fig. 3 and Fig. 4, which is respectively shown, realizes systematic schematic diagram and control principle drawing of the invention.This system is a kind of master From the servo operation of formula, system includes man-machine interface, computer and robot delicate.Man-machine interface system can by flexibility The data glove and sensor array composition of stretch sensor composition.The data glove and sensing that flexible extensible sensor is constituted Device array is used to obtain the operating gesture of manpower and Dextrous Hand.Computer system includes that data glove calibration unit and principal and subordinate's hand reflect Penetrate algorithm unit and seized condition recognition unit.Dexterous robot hand system by operation control unit, Dextrous Hand execution unit, can Stretch sensor unit and power tactilely-perceptible unit composition.Man-machine interface system obtains the operating gesture of manpower, passes through computer Network analysis processing, sends dexterous robot hand system to.For dexterous robot hand system for obtaining control information, control is dexterous Hand executes operation task.Power when power tactilely-perceptible unit is used to perceive Dextrous Hand crawl target object feels information, and will test Information passes to computer system, the grasp speed of seized condition and adaptive adjustment Dextrous Hand is analyzed and determined, to realize one The servo operation of manpower collaboration Dextrous Hand of the kind based on multi-information fusion.

Claims (4)

1. a kind of servo operation of the manpower collaboration Dextrous Hand based on multi-information fusion, feature includes: manpower flexibility can Stretch sensor, Dextrous Hand flexible extensible sensor, flexible force-touch sensor, wireless transport module, Dextrous Hand control mould Block and computer module;
The computer module includes: gesture-capture module, gesture mapping block and seized condition identification module;
The manpower flexible extensible sensor is arranged in the joint of flexible gloves and is worn on manpower, for acquiring manpower behaviour Manpower joint angles feature when making, and the gesture-capture module is sent to by the wireless transport module;
The gesture-capture module carries out gesture-capture to the manpower joint information, obtains manpower gesture information and passes to institute State gesture mapping block;
The joint of robot delicate is arranged in the Dextrous Hand flexible extensible sensor, when for acquiring Dextrous Hand operation Delicate finger-joint angle character, and the gesture-capture module is sent to by the wireless transport module;
The gesture-capture module carries out gesture-capture to the delicate finger-joint information, obtains Dextrous Hand gesture information and transmits The gesture mapping block;
The gesture mapping block manpower gesture information and Dextrous Hand gesture information based on the received calculate needed for Dextrous Hand Expectation driving value, thus establish gesture mapping off-line model;
The gesture mapping block obtains the current information of manpower gesture, obtains working as Dextrous Hand using the gesture mapping model Preceding expectation driving value is simultaneously sent to the Dextrous Hand control module for driving Dextrous Hand to grab target object;
On the inside of the flexibility force-touch sensor is arranged at the finger tip of robot delicate, for acquiring Dextrous Hand crawl target The pressure information of object, and the seized condition identification module is passed to by the wireless transport module;
It include: sliding mode, vibrating state and stable state in seized condition identification module setting seized condition;Further according to The received pressure information of institute, establishes offline Classification and Identification model to the seized condition using support vector machines;
The seized condition identification module obtains current pressure information using the flexible force-touch sensor, and grabs described in utilization State identified off-line module is taken, current seized condition is obtained and is judged, if current seized condition is stable state, is stopped The nearly articulations digitorum manus of clever hand finger is driven, saves and show the present driving value of Dextrous Hand, if current seized condition is sliding mode, Next expectation driving value of Dextrous Hand then is obtained using the gesture mapping block and is sent to the Dextrous Hand control module; If current seized condition is vibrating state, manpower operation and Dextrous Hand operation are re-started, and resurvey and capture.
2. manpower according to claim 1 cooperates with Dextrous Hand servo operation, it is characterized in that: the flexible force tactile passes Sensor is three-decker, is using the melamine sponge after being carbonized as intermediate sensitive layer (2b), at the sensitive layer (2b) Upper and lower surfaces be respectively arranged with upper electrode layer (2a), lower electrode layer (2c).
3. manpower according to claim 2 cooperates with Dextrous Hand servo operation, it is characterized in that: the upper electrode layer (2a) Copper plate electrode is used with the lower electrode layer (2c), is respectively adhered on the upper and lower surfaces of sensitive layer.
4. a kind of manpower based on multi-information fusion cooperates with Dextrous Hand servo operation method, it is characterized in that carrying out as follows:
Manpower flexible extensible sensor is arranged in and is worn on the joints of the flexible gloves of manpower and goes forward side by side pedestrian's hand by step 1 Operation recycles CCD camera and resistance instrument to demarcate the joint angles and manpower resistance value of manpower, obtains manpower mapping Coefficient M;
The joint of robot delicate is arranged in Dextrous Hand flexible extensible sensor and carries out Dextrous Hand operation, is recycled CCD camera and resistance instrument demarcate the joint angles and Dextrous Hand resistance value of Dextrous Hand, obtain Dextrous Hand mapping coefficient M′;
Step 2 is closed using the manpower of j-th of joint on manpower flexible extensible sensor acquisition i-th of finger of manpower Save angle characterUtilize j-th joint on Dextrous Hand flexible extensible sensor acquisition i-th of finger of Dextrous Hand Delicate finger-joint angle character
Step 3, the manpower gesture information for obtaining j-th of joint on i-th of finger of manpower respectively using formula (1) and formula (2) With the Dextrous Hand gesture information of j-th of joint on i-th of finger of Dextrous Hand
Step 4 obtains the calibration relationship between manpower gesture information and Dextrous Hand gesture information using formula (3), to establish hand Gesture maps off-line model:
Step 5, the nearly articulations digitorum manus of five fingers for controlling Dextrous Hand respectively using five direct current generators, and by adjusting PWM wave The angular velocity of satellite motion of the nearly articulations digitorum manus of five fingers of duty ratio T and driving time t control Dextrous Hand;
It is obtained using formula (4) in the nearly articulations digitorum manus driving value PWM wave of i-th of finger of Dextrous Hand of n-th of PWM wave duty cycle sequence Duty ratio Ti n:
Articulations digitorum manus bending maximum angle close in i-th of finger of Dextrous Hand of n-th of PWM wave duty cycle sequence is obtained using formula (5) Time t consumed in pi/2i n:
ti niTi n (5)
In formula (5), μiFor duty ratio Ti nWith time ti nBetween calibration coefficient, n be PWM wave duty cycle sequence number;
Accounting in the nearly articulations digitorum manus driving PWM wave of i-th of finger of Dextrous Hand of n-th of PWM wave duty cycle sequence is obtained using formula (6) Sky ratio Ti nThe angular speed w of articulations digitorum manus close with i-th of finger of Dextrous Handi nBetween calibration relationship:
In formula (6),For in the penalty coefficient of the nearly articulations digitorum manus of i-th of finger of Dextrous Hand of n-th of PWM wave duty cycle sequence;
Step 6, definition current time are t, and initialize t=1;
Step 7, the manpower joint angles spy that j-th of joint on current i-th of finger is respectively obtained using formula (7) and formula (8) The variable quantity of signWith the variation delta Θ of the current information of the manpower gesture of j-th of joint on i-th of fingeri j:
In formula (7) and formula (8),Indicate the manpower joint angles feature of j-th of joint on i-th of finger of t moment,The manpower joint angles feature for indicating j-th of joint on i-th of finger of t-1 moment, as t=1, For the initial resistivity value of the flexible extensible sensor in i-th of manpower, j-th of finger joint;Indicate i-th of hand of t moment Refer to the current information of the manpower gesture of upper j-th of joint,Indicate j-th of joint on i-th of finger of t-1 moment Manpower gesture current information;As t=1,
Step 8, the current expectation driving time Δ t that i-th of finger of Dextrous Hand is obtained using formula (9)i:
Step 9, the current duty cycle T that the nearly articulations digitorum manus driving PWM wave of i-th of finger of t moment Dextrous Hand is obtained using formula (10)i n (t), so that Dextrous Hand be driven to grab target object:
In formula (10), Ti n(t-1) and Ti n(t-2) t-1 moment and the nearly articulations digitorum manus of i-th of finger of t-2 moment Dextrous Hand are respectively indicated The duty ratio for driving PWM wave, as t=1, Ti n(t-1) and Ti nIt (t-2) is the initial estimate of set duty ratio; For the force-touch sensor arranged on i-th of finger of Dextrous Hand t moment resistance value,For i-th of hand of Dextrous Hand The initial resistivity value for the force-touch sensor arranged on finger;
On the inside of step 10, flexible force-touch sensor are arranged at the finger tip of robot delicate, for acquiring Dextrous Hand crawl The pressure information of target object;
Step 11, setting seized condition include: sliding mode, vibrating state and stable state;Further according to the received pressure letter of institute Breath, establishes offline Classification and Identification model to the seized condition using support vector machines;
Step 12 obtains current pressure information using the flexible force-touch sensor, and utilizes the offline Classification and Identification mould Type obtains current seized condition and is judged, if current seized condition is stable state, stops driving clever hand finger close Articulations digitorum manus saves and shows the current expectation driving time Δ t of i-th of finger of Dextrous HandiWith i-th of finger of t moment Dextrous Hand The current duty cycle T of nearly articulations digitorum manus driving PWM wavei n(t) and seized condition;If current seized condition is sliding mode, t+1 is enabled It is assigned to t;And execute step 7;If current seized condition is vibrating state, 6 are thened follow the steps.
CN201811549841.2A 2018-12-18 2018-12-18 Servo operation system and method for human hand cooperation dexterous hand based on multi-information fusion Active CN109514521B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201811549841.2A CN109514521B (en) 2018-12-18 2018-12-18 Servo operation system and method for human hand cooperation dexterous hand based on multi-information fusion

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201811549841.2A CN109514521B (en) 2018-12-18 2018-12-18 Servo operation system and method for human hand cooperation dexterous hand based on multi-information fusion

Publications (2)

Publication Number Publication Date
CN109514521A true CN109514521A (en) 2019-03-26
CN109514521B CN109514521B (en) 2020-06-26

Family

ID=65796117

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201811549841.2A Active CN109514521B (en) 2018-12-18 2018-12-18 Servo operation system and method for human hand cooperation dexterous hand based on multi-information fusion

Country Status (1)

Country Link
CN (1) CN109514521B (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109955254A (en) * 2019-04-30 2019-07-02 齐鲁工业大学 The remote operating control method of Mobile Robot Control System and robot end's pose
CN110253634A (en) * 2019-07-08 2019-09-20 杭州高烯科技有限公司 A kind of remote sensing manipulator
CN111185906A (en) * 2020-01-10 2020-05-22 上海大学 Leap Motion-based dexterous hand master-slave control method
CN111376269A (en) * 2020-03-04 2020-07-07 北京海益同展信息科技有限公司 Object grabbing method and device, storage medium and electronic equipment
CN111941423A (en) * 2020-07-24 2020-11-17 武汉万迪智慧科技有限公司 Man-machine interaction mechanical gripper control system and method
CN114046809A (en) * 2021-11-11 2022-02-15 广东省科学院半导体研究所 Optical sensing device, sensing equipment and system

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN202572399U (en) * 2012-05-15 2012-12-05 郑州大学 Synchronous smart mechanical arm of robot
CN105127973A (en) * 2015-09-14 2015-12-09 华中科技大学 Embedded under-actuated prosthetic hand control system based on CyberGlove
WO2015134336A3 (en) * 2014-03-02 2015-12-17 Drexel University Wearable devices, wearable robotic devices, gloves, and systems, methods, and computer program products interacting with the same
CN105818129A (en) * 2016-04-12 2016-08-03 华南理工大学 Humanoid hand control system based on data glove
CN206048251U (en) * 2016-07-04 2017-03-29 浙江理工大学 Gesture identification Apery manipulator system based on Multi-sensor Fusion
CN106695736A (en) * 2016-07-04 2017-05-24 浙江理工大学 Gesture identification human-simulated mechanical arm system based on multi-sensor fusion and synchronizing method
US20170322629A1 (en) * 2016-05-04 2017-11-09 Worcester Polytechnic Institute Haptic glove as a wearable force feedback user interface
CN206733009U (en) * 2017-04-18 2017-12-12 吉林大学 A kind of bionic mechanical hand based on Fusion
CN108972494A (en) * 2018-06-22 2018-12-11 华南理工大学 A kind of Apery manipulator crawl control system and its data processing method

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN202572399U (en) * 2012-05-15 2012-12-05 郑州大学 Synchronous smart mechanical arm of robot
WO2015134336A3 (en) * 2014-03-02 2015-12-17 Drexel University Wearable devices, wearable robotic devices, gloves, and systems, methods, and computer program products interacting with the same
CN105127973A (en) * 2015-09-14 2015-12-09 华中科技大学 Embedded under-actuated prosthetic hand control system based on CyberGlove
CN105818129A (en) * 2016-04-12 2016-08-03 华南理工大学 Humanoid hand control system based on data glove
US20170322629A1 (en) * 2016-05-04 2017-11-09 Worcester Polytechnic Institute Haptic glove as a wearable force feedback user interface
CN206048251U (en) * 2016-07-04 2017-03-29 浙江理工大学 Gesture identification Apery manipulator system based on Multi-sensor Fusion
CN106695736A (en) * 2016-07-04 2017-05-24 浙江理工大学 Gesture identification human-simulated mechanical arm system based on multi-sensor fusion and synchronizing method
CN206733009U (en) * 2017-04-18 2017-12-12 吉林大学 A kind of bionic mechanical hand based on Fusion
CN108972494A (en) * 2018-06-22 2018-12-11 华南理工大学 A kind of Apery manipulator crawl control system and its data processing method

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109955254A (en) * 2019-04-30 2019-07-02 齐鲁工业大学 The remote operating control method of Mobile Robot Control System and robot end's pose
CN110253634A (en) * 2019-07-08 2019-09-20 杭州高烯科技有限公司 A kind of remote sensing manipulator
CN111185906A (en) * 2020-01-10 2020-05-22 上海大学 Leap Motion-based dexterous hand master-slave control method
CN111376269A (en) * 2020-03-04 2020-07-07 北京海益同展信息科技有限公司 Object grabbing method and device, storage medium and electronic equipment
CN111376269B (en) * 2020-03-04 2021-11-09 北京海益同展信息科技有限公司 Object grabbing method and device, storage medium and electronic equipment
CN111941423A (en) * 2020-07-24 2020-11-17 武汉万迪智慧科技有限公司 Man-machine interaction mechanical gripper control system and method
CN114046809A (en) * 2021-11-11 2022-02-15 广东省科学院半导体研究所 Optical sensing device, sensing equipment and system

Also Published As

Publication number Publication date
CN109514521B (en) 2020-06-26

Similar Documents

Publication Publication Date Title
CN109514521A (en) The servo operation and its method of manpower collaboration Dextrous Hand based on multi-information fusion
JP4878842B2 (en) Robot drive method
Glauser et al. Deformation capture via soft and stretchable sensor arrays
CN103192387A (en) Robot and control method thereof
CN1785608A (en) Control platform of multifinger mechanical skillful closed ring real time action
CN104570731A (en) Uncalibrated human-computer interaction control system and method based on Kinect
CN107225576A (en) A kind of control method based on soft finger
WO2019173678A1 (en) Optimal hand pose tracking using a flexible electronics-based sensing glove and machine learning
JP2023542055A (en) Interactive tactile perception method for object instance classification and recognition
Delgado et al. Adaptive tactile control for in-hand manipulation tasks of deformable objects
Sun et al. Machine learning for haptics: Inferring multi-contact stimulation from sparse sensor configuration
Ge et al. A real-time gesture prediction system using neural networks and multimodal fusion based on data glove
CN109521784B (en) Touch sensing type wearable upper limb exoskeleton unmanned aerial vehicle control system and method
Xiong et al. A target grabbing strategy for telerobot based on improved stiffness display device
CN113792792B (en) Robot-based haptic data perception analysis method
CN115070781B (en) Object grabbing method and two-mechanical-arm cooperation system
WO2022166770A1 (en) Bilateral teleoperation system and control method therefor
Zhou et al. Learning to predict friction and classify contact states by tactile sensor
Zhong et al. A flexible wearable e-skin sensing system for robotic teleoperation
Guo et al. Aerial Interaction with Tactile Sensing
Gu et al. Fine manipulative action recognition through sensor fusion
CN111230872A (en) Object delivery intention recognition system and method based on multiple sensors
CN112083832A (en) Transplantable touch screen device based on IMU and infrared induction
Cretu et al. Estimation of deformable object properties from shape and force measurements for virtualized reality applications
CN110434834A (en) A kind of man-machine collaboration mechanical arm

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant