CN107765702A - Remotely-piloted vehicle and its control method - Google Patents
Remotely-piloted vehicle and its control method Download PDFInfo
- Publication number
- CN107765702A CN107765702A CN201610680463.6A CN201610680463A CN107765702A CN 107765702 A CN107765702 A CN 107765702A CN 201610680463 A CN201610680463 A CN 201610680463A CN 107765702 A CN107765702 A CN 107765702A
- Authority
- CN
- China
- Prior art keywords
- degree
- freedom
- remotely
- action
- piloted vehicle
- 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
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course or altitude of land, water, air, or space vehicles, e.g. automatic pilot
- G05D1/10—Simultaneous control of position or course in three dimensions
- G05D1/101—Simultaneous control of position or course in three dimensions specially adapted for aircraft
Landscapes
- Engineering & Computer Science (AREA)
- Aviation & Aerospace Engineering (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Automation & Control Theory (AREA)
- Toys (AREA)
Abstract
The embodiments of the invention provide remotely-piloted vehicle and its control method.The remotely-piloted vehicle includes：Signal receiving unit, for receiving control signal, the control signal comprises at least the first row vector, the second row vector and the third line vector；With motion control unit, for the motion with the first row vector of the control signal, the second row vector and the third line vector majorization remotely-piloted vehicle in the first dimension, the second dimension and third dimension respectively；Wherein, the action of the first single-degree-of-freedom, the action of the second single-degree-of-freedom and the action of the 3rd single-degree-of-freedom in the gesture motion for the user that the first row vector, the second row vector and the third line vector of the control signal correspond respectively to remotely-piloted vehicle, it is the linear combination of the action of the first single-degree-of-freedom, the action of the second single-degree-of-freedom and the action of the 3rd single-degree-of-freedom with, the gesture motion.By remotely-piloted vehicle and its control method according to embodiments of the present invention, the motion control of aircraft can be remotely controlled by the gesture motion of user, so as to improve the accuracy of control and Discussing Convenience.
Description
Technical field
The present invention relates to remote control equipment field, can more particularly to pass through the remotely-piloted vehicle and its controlling party of gesture control
Method.
Background technology
In recent years, benefit from technology and enter this and cost and decline, the entry threshold in unmanned plane market increasingly reduces, consumer level without
Man-machine market has been broken out, and civilian unmanned plane enters into average family.
Generally, the not manned aircraft manipulated by remote control equipment or self-contained program's control device can be said to unmanned plane,
The use of this unmanned plane needs a whole set of special purpose device and equipment, and unmanned plane and these equipment form a complete system
System, referred to as UAS.UAS generally includes the outer remote control station of unmanned plane, machine and taken off, retracting device etc..
Unmanned plane includes flight control system, payload and the dress for taking off and reclaiming on body, power set, machine
Put.The body of unmanned plane is roughly the same with manned aircraft, relatively simple for structure, light, widely used nonmetallic materials.
The type of power set is different because of the performance of unmanned plane and purposes, is characterized in that cost is low and requires short life.Fly and control on machine
System processed includes automatic pilot, presetting apparatus, remote control and remote-measuring equipment, television camera, homer, calculating
Machine, automatic takeoff and landing system etc..Unmanned plane can choose to install the said equipment according to different purposes and install other special equipments additional.
The use of unmanned plane was put into effect by the specification of unmanned plane industry laws and regulations according to 2013《It is civilian that nobody drives
Sail aircraft system driver management temporary provisions》, MAV of the weight less than or equal to 7 kilograms, flight range is visual
500 meters of sighting distance inside radius, relative altitude is less than in the range of 120 meters, without certificate administration；The indexs such as weight are higher than above-mentioned standard
Unmanned plane and fly into complicated spatial domain, driver need to include the employer's organization even supervision of civil aviation authority.
For the unmanned plane of most of consumer levels, because user does not have driver's qualification of unmanned plane substantially,
Thus each consumer level unmanned plane is all above-mentioned《Regulation》In fly in required scope.
For the unmanned plane in visual sighting distance, user can enter according to the state of flight for the unmanned plane seen to unmanned plane
Row remote control.In various remote control thereofs, the remote control mode for being referred to as " motion sensing manipulation pattern " has recently been developed, i.e., only needs
Simply adjust mobile phone to act with regard to unmanned plane during flying posture can be manipulated, such as by inclination, rotating mobile etc., change unmanned plane
Flight angle and direction.
Accordingly, there exist the needs for the control mode for improving unmanned plane.(the defects of prior art, is if not energy of the present invention
Enough solve, it is just nonsensical, put forward to have minus effect on the contrary, so advantages of the present invention had better be emphasized, and should not
Excessive the shortcomings that emphasizing prior art)
The content of the invention
It is an object of the invention to for it is above-mentioned in the prior art the defects of and deficiency, there is provided have gesture control manipulation
The novel and improved remotely-piloted vehicle and its control method of mode.
According to an aspect of the present invention, there is provided a kind of remotely-piloted vehicle, including：Signal receiving unit, controlled for receiving
Signal processed, the control signal comprise at least the first row vector, the second row vector and the third line vector；Motion control unit, use
In respectively with remotely-piloted vehicle described in the first row vector, the second row vector and the third line vector majorization of the control signal
Motion in dimension, the second dimension and third dimension；Wherein, the first row vector of the control signal, the second row vector and
The first single-degree-of-freedom that the third line vector is corresponded respectively in the gesture motion of the user of the remotely-piloted vehicle acts, the second list
The free degree acts and the action of the 3rd single-degree-of-freedom, and, the gesture motion is the first single-degree-of-freedom action, second list
The free degree acts and the linear combination of the 3rd single-degree-of-freedom action.
In above-mentioned remotely-piloted vehicle, the dimension includes Spatial Dimension and time dimension.
In above-mentioned remotely-piloted vehicle, the first single-degree-of-freedom action is wrist rotation, and second single-degree-of-freedom is moved
Work is wrist upset, and the 3rd single-degree-of-freedom action is that palm is held.
In above-mentioned remotely-piloted vehicle, each single-degree-of-freedom acts includes motion side for control the remotely-piloted vehicle
To at least one of kinematic parameter with movement velocity.
In above-mentioned remotely-piloted vehicle, first single-degree-of-freedom action, second single-degree-of-freedom action and described the
Three single-degree-of-freedoms act the direction in three dimensions for being respectively used to control the remotely-piloted vehicle in three-dimensional coordinate system
Vector.
In above-mentioned remotely-piloted vehicle, first single-degree-of-freedom is acted for controlling the level of the remotely-piloted vehicle to turn
To angle, second single-degree-of-freedom acts the vertical duction angle for controlling the remotely-piloted vehicle, and the 3rd list is certainly
Flying speed for controlling the remotely-piloted vehicle is acted by degree.
In above-mentioned remotely-piloted vehicle, the control signal is examined by the myoelectricity of the wearable electronic of the user
Survey what is generated with control function by the gesture motion of the user.
In above-mentioned remotely-piloted vehicle, generating the control signal by the gesture motion of the user includes：
Obtain electromyographic signal Z, the signal Z and correspond to the gesture motion of the user；
Below equation is to acquired bioelectrical signals Z processing：
Z '=W ' F (1)
Z ' is the characteristic signal of the bioelectrical signals Z, and W ' is system transfer matrix W pseudo inverse matrix, and F is non-negative control
Signal matrix processed；
Wherein, the system transfer matrix W is obtained by the sparse nonnegative integer factorising algorithm in training process, and
The row vector of the non-negative control signals matrix is synchronous in direct ratio with the action of corresponding single-degree-of-freedom.
In above-mentioned remotely-piloted vehicle, the system transfer matrix W is obtained by the training process comprised the following steps：Make
Object must be trained to carry out including at least one multiple training actions in multiple degrees of freedom action and single-degree-of-freedom action；Will be from every
The electromyographic signal combination that individual training action detects is designated as electromyographic signal matrix Z1；Using sparse nonnegative integer factorising algorithm
The electromyographic signal matrix Z1 is decomposed into non-negative system transfer matrix Wi and sparse non-negative control signals matrix F i, i is represented
Iterations；Non- negative system transfer matrix Wi and sparse non-negative control signals matrix F i are updated by iteration, wherein described
Non- negative control signals matrix F i each row vector represents the single-degree-of-freedom action in one of joint of the arm；With,
Non- negative system transfer matrix Wi after being updated is as system transfer matrix W.
In above-mentioned remotely-piloted vehicle, further comprise：The sparse of non-negative control signals matrix F is controlled based on l1 norms
Property degree, is represented by below equation：
Wherein, F (:, t) be control signal matrix F t column vectors, ' Fro ' is Ni Wusi norms, and m is detection myoelectricity letter
Number the number of channel, T is time span, λ>0 is the regular parameter of the accuracy that equilibrium factor decomposes and F openness degree, on
Mark "+" and "-" represents each free degree positively and negatively.
In above-mentioned remotely-piloted vehicle, the equation (2) is rewritten as：
Wherein e1×2mIt is the row vector that all items are equal to 1, and 01×TEqual to 0.
In above-mentioned remotely-piloted vehicle, equation (3) is solved by alternately non-negative least square method, and pass through fixation
A renewal for carrying out iteration in the system transfer matrix W and the signal matrix F another, as shown in below equation：
Wherein described F(k+1)And W(k+1)Solution with closed form.
In above-mentioned remotely-piloted vehicle, the control signal is represented by below equation：
Wherein, F1Correspond to the control signal of the first single-degree-of-freedom action, F2It is single free to correspond to described second
The control signal of degree action, and F2The control signal of the 3rd single-degree-of-freedom action is corresponded to, subscript "+" and "-" represent
Per single-degree-of-freedom positively and negatively；With
The control signal is scaled by scaling correction factor according to below equation：
Wherein, the scaling correction factor τijIt is described substantially free to be determined so that the control signal F is mapped to
The whole dimensional extent of degree action.
According to another aspect of the present invention, there is provided a kind of control method of remotely-piloted vehicle, including：Receive control letter
Number, the control signal comprises at least the first row vector, the second row vector and the third line vector；Respectively with the control signal
First row vector, the second row vector and remotely-piloted vehicle described in the third line vector majorization are in the first dimension, the second dimension and the 3rd
Motion in dimension；Wherein, the first row vector of the control signal, the second row vector and the third line vector correspond respectively to
The action of the first single-degree-of-freedom, the action of the second single-degree-of-freedom and the 3rd in the gesture motion of the user of the remotely-piloted vehicle is single certainly
Acted by degree, and, the gesture motion is the first single-degree-of-freedom action, second single-degree-of-freedom action and the described 3rd
The linear combination of single-degree-of-freedom action.
In the control method of above-mentioned remotely-piloted vehicle, the dimension includes Spatial Dimension and time dimension.
In the control method of above-mentioned remotely-piloted vehicle, the first single-degree-of-freedom action is that wrist rotates, described second
Single-degree-of-freedom action is wrist upset, and the 3rd single-degree-of-freedom action is that palm is held.
In the control method of above-mentioned remotely-piloted vehicle, each single-degree-of-freedom is acted for controlling the remotely-piloted vehicle
At least one of kinematic parameter including the direction of motion and movement velocity.
In the control method of above-mentioned remotely-piloted vehicle, the first single-degree-of-freedom action, second single-degree-of-freedom are moved
Make and the 3rd single-degree-of-freedom acts three dimensions for being respectively used to control the remotely-piloted vehicle in three-dimensional coordinate system
Direction vector on degree.
In the control method of above-mentioned remotely-piloted vehicle, first single-degree-of-freedom is acted for controlling the remote control distributor
Device is horizontally diverted angle, and second single-degree-of-freedom acts the vertical duction angle for controlling the remotely-piloted vehicle, institute
State the 3rd single-degree-of-freedom and act flying speed for controlling the remotely-piloted vehicle.
In the control method of above-mentioned remotely-piloted vehicle, the control signal is set by the wearable electronic of the user
What standby checking with EMG method and control function was generated by the gesture motion of the user.
In the control method of above-mentioned remotely-piloted vehicle, the control signal bag is generated by the gesture motion of the user
Include：
Obtain electromyographic signal Z, the signal Z and correspond to the gesture motion of the user；
Below equation is to acquired bioelectrical signals Z processing：
Z '=W ' F (1)
Z ' is the characteristic signal of the bioelectrical signals Z, and W ' is system transfer matrix W pseudo inverse matrix, and F is non-negative control
Signal matrix processed；
Wherein, the system transfer matrix W is obtained by the sparse nonnegative integer factorising algorithm in training process, and
The row vector of the non-negative control signals matrix is synchronous in direct ratio with the action of corresponding single-degree-of-freedom.
In the control method of above-mentioned remotely-piloted vehicle, the system is obtained by the training process comprised the following steps and turned
Move matrix W：So that training object carries out including at least one multiple training in multiple degrees of freedom action and single-degree-of-freedom action
Action；The electromyographic signal detected from each training action combination is designated as electromyographic signal matrix Z1；Using sparse nonnegative integer
The electromyographic signal matrix Z1 is decomposed into non-negative system transfer matrix Wi and sparse non-negative control signals by factorising algorithm
Matrix F i, i represent iterations；Non- negative system transfer matrix Wi and sparse non-negative control signals matrix are updated by iteration
Fi, wherein each row vector of the non-negative control signal matrix F i represents a list in one of joint of the arm freely
Degree acts；With the non-negative system transfer matrix Wi after being updated is as system transfer matrix W.
In the control method of above-mentioned remotely-piloted vehicle, further comprise：Non- negative control signals are controlled based on l1 norms
The openness degree of matrix F, is represented by below equation：
Meet W, F >=0,
s.t.Wij, Fij≥0 (2)
Wherein, F (:, t) be control signal matrix F t column vectors, ' Fro ' is Ni Wusi norms, and m is detection myoelectricity letter
Number the number of channel, T is time span, λ>0 is the regular parameter of the accuracy that equilibrium factor decomposes and F openness degree, on
Mark "+" and "-" represents each free degree positively and negatively.
In the control method of above-mentioned remotely-piloted vehicle, the equation (2) is rewritten as：
Wherein e1×2mIt is the row vector that all items are equal to 1, and 01×TEqual to 0.
In the control method of above-mentioned remotely-piloted vehicle, equation (3) is solved by alternately non-negative least square method,
And by a renewal come iteration in fixation the system transfer matrix W and the signal matrix F another, such as with lower section
Shown in formula：
Wherein described F(k+1)And W(k+1)Solution with closed form.
In the control method of above-mentioned remotely-piloted vehicle, the control signal is represented by below equation：
Wherein, F1Correspond to the control signal of the first single-degree-of-freedom action, F2It is single free to correspond to described second
The control signal of degree action, and F2The control signal of the 3rd single-degree-of-freedom action is corresponded to, subscript "+" and "-" represent
Per single-degree-of-freedom positively and negatively；With
The control signal is scaled by scaling correction factor according to below equation：
Wherein, the scaling correction factor τijIt is described substantially free to be determined so that the control signal F is mapped to
The whole dimensional extent of degree action.
By the remotely-piloted vehicle and its control method according to the present invention, can be remotely controlled by the gesture motion of user
The motion control of aircraft, so as to improve the Discussing Convenience of manipulation.Also, due to each single free in the gesture motion of user
Degree action corresponds to motion of the remotely-piloted vehicle in a dimension, there is provided the accuracy of control.
Also, in the remotely-piloted vehicle according to the present invention and its control method, wearable electronic can be passed through
Checking with EMG method and control function detect the gesture motion of user, to enable control signal accurately to embody gesture motion
Amplitude and direction, so as to realize the synchronization of remotely-piloted vehicle and direct proportion control.
In addition, in the remotely-piloted vehicle according to the present invention and its control method, by representational most to move
Gesture controls the motion of remotely-piloted vehicle in three dimensions, user can be made intuitively to manipulate remotely-piloted vehicle, so as to carry
Rise the manipulation impression of user.
Brief description of the drawings
Fig. 1 is the schematic diagram of single-degree-of-freedom action and multiple degrees of freedom action；
Fig. 2 is the schematic block diagram that the control of multiple degrees of freedom action is realized with electromyographic signal；
Fig. 3 is to simulate the schematic diagram that electromyography transducer is worn in forearm cross section；
Fig. 4 is the schematic block diagram of remotely-piloted vehicle according to embodiments of the present invention；
Fig. 5 is the schematic diagram of representative gesture according to embodiments of the present invention；
Fig. 6 is the indicative flowchart of the control method of remotely-piloted vehicle according to embodiments of the present invention.
Embodiment
Describe to be used for the open present invention below so that those skilled in the art can realize the present invention.It is excellent in describing below
Embodiment is selected to be only used as illustrating, it may occur to persons skilled in the art that other obvious modifications.Define in the following description
General principle of the invention can apply to other embodiments, deformation program, improvement project, equivalent and do not carry on the back
From the other technologies scheme of the spirit and scope of the present invention.
It will be understood by those skilled in the art that in disclosure of the invention, term " longitudinal direction ", " transverse direction ", " on ",
" under ", "front", "rear", "left", "right", " vertical ", " level ", " top ", " bottom " " interior ", the orientation of the instruction such as " outer " or position close
System is to be based on orientation shown in the drawings or position relationship, and it is for only for ease of the description present invention and simplifies description, without referring to
Show or imply that the device of meaning or element there must be specific orientation, with specific azimuth configuration and operation, therefore above-mentioned art
Language is not considered as limiting the invention.
It is understood that term " one " be interpreted as " at least one " or " one or more ", i.e., in one embodiment,
The quantity of one element can be one, and in a further embodiment, the quantity of the element can be multiple, and term " one " is no
It is understood that as the limitation to quantity.
The term and word used in description below and claim is not limited to literal implication, but only by the present inventor
The present invention can be understood and as one man understand by being used so that.Therefore, to those skilled in the art clearly only for explanation
Purpose rather than provide this hair to limit the purpose of the present invention as defined in appended claims and their equivalent
The following description of bright various embodiments.
Although for example the ordinal number of " first ", " second " etc. will be used to describe various assemblies, not limit those herein
Component.The term is only used for distinguishing a component and another component.For example, first assembly can be referred to as the second component, and together
Sample, the second component can also be referred to as first assembly, without departing from the teaching of inventive concept.Term as used herein " and/
Or " include any of one or more projects listed associated and all combinations.
The term being used herein is only used for describing the purpose of various embodiments and is not intended to limit.As used herein,
Singulative is intended to also include plural form, makes an exception unless the context clearly dictates.Will further be understood that term " comprising " and/or
" having " specifies depositing for described feature, number, step, operation, component, element or its combination when using in this specification
, and it is not excluded for the presence or additional of one or more of the other feature, number, step, operation, component, element or its group.
The term being used herein including technology and scientific terminology has the art being generally understood that with those skilled in the art
Language identical implication, so long as not being defined differently than the term.It should be understood that the term tool limited in usually used dictionary
There is the implication consistent with the implication of term of the prior art.
The present invention is further detailed explanation with reference to the accompanying drawings and detailed description：
By bioelectrical signals, for example, it is a kind of emerging operation that electromyographic signal, EEG signals etc., which carry out operation to equipment,
Mode.It is collection electromyographic signal to the mode of operation of equipment by electromyographic signal now, then to myoelectricity by taking electromyographic signal as an example
Signal is handled, and finally exports operation signal.Operation object is received after operation signal according to corresponding to being carried out operation signal
Operation.
Myoelectricity control in, in muscle space the action of the more than one free degree (DOF) single DOF can be broken down into and moved
Make the linear combination of (also known as basic DOF actions), for example, it is a single DOF action that palm, which is held, be defined as DOF1, and hand
Wrist rotation is defined as DOF2.And further, DOF direction is represented with subscript "+" "-".Specifically, DOF1+ represents hand
The action of opening is slapped, and DOF1- represents the action of palm closure, DOF2+ represents to overturn the action of wrist counterclockwise, and DOF2-
Represent the action of upset wrist clockwise.Fig. 1 is the schematic diagram of single-degree-of-freedom action and multiple degrees of freedom action.As shown in figure 1, its
Three objects are shown, two actions that dark color circle represents are single DOF actions, and the action that light color circle represents is two single DOF
The combination of action, that is to say, that acted, it is necessary to activate two DOF that dark circle represents simultaneously to realize that light color circle represents dynamic
Make.
Fig. 2 is the schematic block diagram that the control of multiple degrees of freedom action is realized with bioelectrical signals.Neuro-physiology is ground
Study carefully and show, at spinal nerve aspect (Spinal level), nerve signal can control multiple muscle by way of linear combination
The Union Movement of group.Therefore, if stimulating nerve in a manner of linear combination simultaneously, muscle is moved in a manner of linear combination
Make.As shown in Fig. 2 for directly controlling surface electromyography (sEMG) signal Z (t) of muscle group to be different from controlling the basic free degree
Signal F (t).Specifically, the implication of each variable in Fig. 2 is as follows：
F(t)：Control signal, F (t)=[f1(t),…,fi(t),…,fN(t)]T, each fi(t) one is controlled substantially
The activation of free degree action, so, use can be obtained by controlling the linear combination of control signal that the basic free degree acts
In the control signal of control multiple degrees of freedom action；
S：Muscle Harmonious Matrix, each free degree are converted into the activity of one group of muscle by S, and S simulates the mechanism of spinal nerve；
D(t)：D (t)=[d1(t),…,di(t),…,dM(t)]T, di(t) driven (most to go downwards to the nerve of each muscle
Whole kinesitherapy nerve code)；
Y(t)：Y (t)=[y1(t),…,yi(t),…,yM(t)]T, yi(t) it is the active signal (iEMG) of muscle；
G(t)：Tissue equivalent's filter array；
Z(t)：Z (t)=[z1(t),…zi(t),…,zL(t)] T, zi(t) it is the sEMG signals of a passage.
By establishing the model of muscle Harmonious Matrix and according to the approximation under certain condition, it can be deduced that following two public affairs
Formula：
WL×NIt is system transfer matrix,For Z (t) root mean square (Root Mean Square：RMS).
Control in the control system of operation object, this is two most basic formula, is being represented based on electromyographic signal respectively
Signal processed to electromyographic signal conversion and electromyographic signal to control signal conversion.
In specific operating process, multi channel sEMG signals are gathered using electromyography transducer, and seek each channel
SEMG signals feature, such as the root mean square of the sEMG signals in formula (2), recycle formula (2) to change sEMG signals
For control signal.
Fig. 3 is to simulate the schematic diagram that electromyography transducer is worn in forearm cross section.As shown in Figure 3, it is assumed that control wrist is turned over
The muscle for turning (DOF1) and wrist rotation (DOF2) is 4, i.e.,：Muscle A to muscle D, then can be right in forearm surface wearing 8
The EMG signal sensor of title is to obtain the sEMG signal Z (t) of 8 channels.Certainly, it will be understood by those skilled in the art that EMG
The number of signal transducer can also be other numbers, such as more than 8 or less than 8, also, EMG signal sensor
Setting is also not limited to the mode shown in Fig. 3.
It was found by the inventors of the present invention that the characteristic of the combination of actions and decomposition in this muscle space is especially suitable for carrying out
Motion control.Specifically, motion in three dimensions can equally be divided into linear group of the motion in each dimension
Close, for example, the direction vector in any one three dimensions is all one-dimensional vector under xyz coordinate systems in the x-direction, edge
The direct combination of the one-dimensional vector in y directions and one-dimensional vector in the z-direction.If it is possible to by muscle space it is more from
In being acted by degree each single-degree-of-freedom action is converted into control signal synchronous and in direct ratioly, then as with it is each but from
The control signal of the linear combination of control signal as corresponding to acting degree, also it is fully able to act phase with the multiple degrees of freedom exactly
It is corresponding.
Therefore, inventors have seen that, in the control process of remotely-piloted vehicle, because remotely-piloted vehicle is in itself
Moved in three dimensions, if by the way that multiple degrees of freedom action is converted into control signal according to completely corresponding relation,
So the control signal will control the motion of remotely-piloted vehicle in a manner of directly reacting multiple degrees of freedom action, so that
The control to multiple degrees of freedom action must be converted into the motion control of remotely-piloted vehicle.
For user, gesture motion is often made with one kind and easily controllable action, also, according to
Theory in muscle space, the gesture motion as multiple degrees of freedom action can also equally be decomposed into multiple single-degree-of-freedom actions
Linear combination, wherein the action of each single-degree-of-freedom also corresponds to some specific gesture motion, than wrist as mentioned above
Upset and wrist rotation.So, if user controls the motion of remotely-piloted vehicle based on gesture motion, it is clear that ratio uses other
Mode, such as the control of mobile phone are more convenient and efficient and more directly perceived for a user.
Accordingly, it is desired to provide can be the line of multiple single-degree-of-freedom actions based on multiple degrees of freedom action in above-mentioned muscle space
Property combination principle, the remotely-piloted vehicle of the motion of remotely-piloted vehicle in three dimensions and its control are controlled by gesture motion
Method processed.
One side according to embodiments of the present invention, there is provided a kind of remotely-piloted vehicle, including：Signal receiving unit, it is used for
Control signal is received, the control signal comprises at least the first row vector, the second row vector and the third line vector；Motion control list
Member, for being existed respectively with the first row vector of the control signal, the second row vector and the third line vector majorization remotely-piloted vehicle
Motion in first dimension, the second dimension and third dimension；Wherein, the first row vector of the control signal, the second row vector and
The first single-degree-of-freedom that the third line vector is corresponded respectively in the gesture motion of the user of remotely-piloted vehicle acts, the second list is free
Degree acts and the action of the 3rd single-degree-of-freedom, and, the gesture motion is the action of the first single-degree-of-freedom, the second single-degree-of-freedom action and the
The linear combination of three single-degree-of-freedoms action.
So, in remotely-piloted vehicle according to embodiments of the present invention, user can be controlled by the gesture motion of oneself
The motion of remotely-piloted vehicle processed, and the fortune of remotely-piloted vehicle in three dimensions can be controlled by a gesture motion
It is dynamic, just look like that user controls the motion of remotely-piloted vehicle such by hand-held remote control aircraft, significantly improve remote control and fly
The handling of row device.
Fig. 4 is the schematic block diagram of remotely-piloted vehicle according to embodiments of the present invention.As shown in figure 4, according to of the invention real
Applying the remotely-piloted vehicle 100 of example includes：Signal receiving unit 101, for receiving control signal, the control signal comprises at least the
One row vector, the second row vector and the third line vector；With motion control unit 102, for respectively with the signal receiving unit
First row vector of 101 control signals received, the second row vector and the third line vector majorization remotely-piloted vehicle are first
Motion in dimension, the second dimension and third dimension；Wherein, the first row vector of the control signal, the second row vector and the 3rd
The first single-degree-of-freedom that row vector is corresponded respectively in the gesture motion of the user of remotely-piloted vehicle is acted, the second single-degree-of-freedom is moved
Make and the 3rd single-degree-of-freedom acts, and, the gesture motion is the action of the first single-degree-of-freedom, the second single-degree-of-freedom action and the 3rd
The linear combination of single-degree-of-freedom action.
Here, it will be understood by those skilled in the art that illustrate only the control with the remotely-piloted vehicle of the present invention in Fig. 4
Related part in scheme, the universal component without showing those remotely-piloted vehicles.Remote control according to embodiments of the present invention flies
Row device can be currently marketed various technical grade and consumer level unmanned plane, or develop or future will develop
Other remotely-piloted vehicles, the embodiment of the present invention be not intended to this progress any restrictions.
In above-mentioned remotely-piloted vehicle, dimension includes Spatial Dimension and time dimension.Although that is, in above description
In, gesture motion is used to control the motion of remotely-piloted vehicle in three dimensions, but gesture motion is for remotely-piloted vehicle
The control of motion is not limited on Spatial Dimension, can also include time dimension.For example, the fortune except controlling remotely-piloted vehicle
Outside dynamic direction, the parameter relevant with the time of the motion of remotely-piloted vehicle, such as movement velocity, motion can also be controlled to accelerate
Degree etc..
In above-mentioned remotely-piloted vehicle, the action of the first single-degree-of-freedom is wrist rotation, and the action of the second single-degree-of-freedom is wrist
Upset, and the action of the 3rd single-degree-of-freedom is that palm is held.
In above-mentioned remotely-piloted vehicle, each single-degree-of-freedom acts includes the direction of motion for control the remotely-piloted vehicle
With at least one of kinematic parameter of movement velocity.
As described above, it is necessary to consider during by motion of the gesture motion of user to control remotely-piloted vehicle
Whether gesture is easy to user to make in itself, and whether this gesture is suitable to accurately control the motion of remotely-piloted vehicle.Therefore,
In remotely-piloted vehicle according to embodiments of the present invention, typical control gesture-type, and the typical control gesture are defined
Type is suitable to the motion of control remotely-piloted vehicle.It will be understood by those skilled in the art that include wrist upset, wrist rotation and hand
The gesture motion that the palm is held can be made by user with a gesture motion, and every kind of gesture is all single-degree-of-freedom action,
So that the gesture motion of user can be decomposed into the linear combination of these three single-degree-of-freedom gesture motions.
So, the control signal generated by gesture motion includes three row vectors respectively, and each row vector corresponds to
One of single-degree-of-freedom action so that each single-degree-of-freedom action in the gesture motion of user may be used to control distant
Motion of the aircraft in a dimension is controlled, and as described above, the dimension is not limited in Spatial Dimension.
The realization of the effect above on the one hand depend in muscle space multiple degrees of freedom action can be broken down into it is basic from
The linear combination acted by degree, on the other hand, the selection for also relying on gesture enables to gesture motion to be accurately resolved into
The combination of multiple gesture motions for representing single-degree-of-freedom.So fortune by gesture motion → control signal → remotely-piloted vehicle
Move this control chain, it is possible to achieve the Synchronization Control of the motion of remotely-piloted vehicle.In addition, as noted previously, as gesture motion
Movement range can correspond to motion characteristic, equally realize remotely-piloted vehicle motion direct proportion control.
Fig. 5 is the schematic diagram of representative gesture according to embodiments of the present invention.Wherein, Fig. 5 (a) is shown outside wrist
Turn over, Fig. 5 (b) shows wrist varus, and Fig. 5 (c) shows that wrist clockwise rotates, and Fig. 5 (d) shows the wrist inverse time
Pin rotates, and Fig. 5 (e) shows that palm opens, and Fig. 5 (f) shows that palm closes.(part of wrist we figure adjust
Whole (c) and (d) is exchanged, and this is not related for palm for a while, and not so closure opens no suitable form of presentation with clenching fist)
According to the characteristic of gesture motion and the characteristic of the motion of remotely-piloted vehicle in three dimensions, implemented according to the present invention
The remotely-piloted vehicle of example can be using a variety of methods that motion is controlled by gesture motion.
Due in three dimensions, any one motion vector may be expressed as under xyz coordinate systems in the x-direction to
Measure, vectorial and vectorial combination in the z-direction in the y-direction, three single-degree-of-freedoms action in gesture motion can also divide
Not Dui Yingyu these three directions vector.So, user can be come with a gesture motion comprising three single-degree-of-freedom actions
Any vector in three dimensions is represented, so that remotely-piloted vehicle moves towards any point in three dimensions.
For example, the vector on x directions can be corresponded to wrist rotation, corresponded to wrist upset on y directions
Vector, and held with palm to correspond to the vector on z directions.Also, as shown in figure 5, because the action of each single-degree-of-freedom can
To define "+" and "-" direction so that the vector that the single-degree-of-freedom action in "+" direction corresponds in x, y or z positive direction, and
The vector that the single-degree-of-freedom action in "-" direction corresponds in x, y or z negative direction.
So, in remotely-piloted vehicle according to embodiments of the present invention, the action of the first single-degree-of-freedom, the second single-degree-of-freedom are moved
Make and the action of the 3rd single-degree-of-freedom is respectively used to control side of the remotely-piloted vehicle in three dimensions in three-dimensional coordinate system
To vector.
Certainly, it will be understood by those skilled in the art that except be employed as the xyz coordinate systems of orthogonal rectangular coordinate system with
Outside, the coordinate system that can also be applied in other three dimensions, it is only necessary to which each single-degree-of-freedom action corresponds in three dimensions
Direction vector in one of dimension.
Also, the amplitude of the gesture motion of user is used for the length scale for representing vector in a certain direction, according to this
In the remotely-piloted vehicle of inventive embodiments, the corresponding journey of amplitude and the vector magnitude of the gesture motion by pre-setting user
Degree, the direction of remotely-piloted vehicle can be controlled with the specific movement range of the gesture motion according to user.For example, when user wishes
Hope and remotely-piloted vehicle is controlled in the motion on only possible big direction with small movement range, can be each single to set
Free degree action corresponds to larger size vectorial in this direction more by a small margin.Certainly, this corresponding relation can not be
Identical is acted for each single-degree-of-freedom, for example, user can set corresponding to more by a small margin for the first single-degree-of-freedom action
The larger size of direction vector in the first dimension, and set what the second single-degree-of-freedom acted to correspond to more by a small margin the
The less size of direction vector on two-dimensionses.So, user's can by gesture motion come control remotely-piloted vehicle to
The motion of any direction, and user can adjust the fortune between gesture motion and the direction of motion according to the motor habit of oneself
Dynamic relation, so as to obtain optimal Consumer's Experience.(although amplitude may be different with actual use scene, relatively more directly perceived, because
It is visible for amplitude, and dynamics must be controlled by user oneself, the speed of the gesture motion in paragraphs below is exactly power in fact
The concept of degree)
Also, except the amplitude of three single-degree-of-freedoms action of the gesture motion of user controls in three dimensions respectively
Outside the direction of motion, it can also be come with the speed (" dynamics " that is referred to as gesture motion) of the gesture motion of user further
Control the movement velocity of remotely-piloted vehicle.Certainly, it will be understood by those skilled in the art that the inspection of the speed of the gesture motion of user
Survey can be accomplished in several ways, except obtaining the width of electromyographic signal caused by the gesture motion of user by checking with EMG method
, can also the mode such as by acceleration transducer outside degree.Also, the speed of the gesture motion of user can be used for table
Show length scale vectorial in a certain direction, can be directly perceived because the dynamics of gesture motion is directly directly proportional to muscle strengh
Ground is embodied in the power of electromyographic signal, therefore in some actual use scenes advantageously.
In addition, in remotely-piloted vehicle according to embodiments of the present invention, can also using it is other by gesture motion come
The method for controlling motion.For example, the action that user is rotated by wrist is horizontally diverted angle control remotely-piloted vehicle, i.e.
Control steering angle of the remotely-piloted vehicle in the x/y plane of three-dimensional system of coordinate on some reference direction.This reference direction can
To be default constant bearing, for example, due south due north etc. or in initial control remotely-piloted vehicle direction.In addition,
User controls the vertical duction angle of remotely-piloted vehicle by wrist upset, and, control remotely-piloted vehicle is in three-dimensional system of coordinate
The interior steering angle relative to z-axis.
Equally, the amplitude of the gesture motion of user corresponds respectively to the concrete numerical value of steering angle, also, according to the hand of people
The operating angle of wrist so that the maximum actuation amplitude of wrist corresponds to maximum steering angle, can so improve for turning to
The resolution ratio of angle.But if in order to which the convenient of user considers, it is also possible that less movement range then correspond to it is larger
Steering angle so that user need not very significantly list action be obtained with very significantly steering angle.
So, by the gesture motion of user, motion of the remotely-piloted vehicle towards any direction in space can be controlled.
Also, due to only having used two single-degree-of-freedom actions, the 3rd single-degree-of-freedom action can be used for controlling remotely-piloted vehicle
Flying speed, so as to reduce the load for signal detection relative to scheme above, also correspondingly reduce the whole of scheme
Body cost.
That is, in above-mentioned remotely-piloted vehicle, the first single-degree-of-freedom acts is horizontally diverted angle for control remotely-piloted vehicle
Degree, the second single-degree-of-freedom acts the vertical duction angle for controlling remotely-piloted vehicle, and the 3rd single-degree-of-freedom is acted for controlling
The flying speed of remotely-piloted vehicle processed.
It will be understood by those skilled in the art that as long as can control in whole or in part in three single-degree-of-freedom actions is distant
The motion of any direction of aircraft in three dimensions is controlled, the embodiment of the present invention is not intended to specific corresponding pass therebetween
System carries out any restriction.
In above-mentioned remotely-piloted vehicle, control signal is checking with EMG method and the control of the wearable electronic by user
Function is generated by the gesture motion of user.
As noted previously, as the spy that multiple degrees of freedom action is combined by multiple single-degree-of-freedom acting linears in muscle space
Property, as long as ensureing motion of each single-degree-of-freedom action control remotely-piloted vehicle in single dimension, it is possible to comprising multiple
The gesture motion of single-degree-of-freedom action is controlled remotely-piloted vehicle in multiple dimensions including Spatial Dimension and time dimension
Motion, and the design parameter of motion of the remotely-piloted vehicle in each dimension in this multiple dimension, such as direction and speed
Degree can also be controlled by the movement range or responsiveness of gesture motion.So, checking with EMG method and control function, institute are passed through
Corresponding relation between this gesture motion established and the motion of remotely-piloted vehicle is not limited in simple corresponding relation, i.e.,
As existing some gesture control functions, certain gestures correspond to some certain types of motion, and can be achieved on essence
Quasi- control, i.e., synchronization and the control of direct proportion recited above.
, can be with after user makes gesture motion that is, in remotely-piloted vehicle according to embodiments of the present invention
Control signal corresponding with the amplitude and/or speed of the gesture motion is generated by checking with EMG method and control function.Due to the control
Amplitude and/or speed sync and in direct ratio, the remote control distributor that by the control signal is controlled of the signal processed with the gesture motion
The motion of device can also be with the amplitude and/or speed sync of the gesture motion and in direct ratio.In this manner it is achieved that accurate control
The effect of the motion of remotely-piloted vehicle, just look like that user directly holds the remotely-piloted vehicle so as to control remotely-piloted vehicle with hand
Motion is the same.Also, the effect of this control is very intuitively, and to feel unnatural without using family for user.
Below, will be carried out specifically to how to generate above-mentioned control signal according to the gesture motion of the user detected
It is bright.
Here, it will be understood by those skilled in the art that when with multichannel continuous acquisition sEMG signals, Z can be m × n
Matrix, wherein m represent the number of channel, and n represents n time.Correspondingly, F is 2a × n matrixes, and wherein a represents single DOF number
Mesh, as described above, each single DOF can consider positively and negatively, and n equally represents n time.So, system transfer matrix
It is m × 2a matrixes.Hereinafter, for convenience of description, in the case where row matrix column number need not be particularly shown, sEMG is believed
Number matrix, system transfer matrix and control signal matrix are uniformly abbreviated as Z, W and F.
Specifically, above formula (1) and (2) can be rewritten as：
Z '=WF (3)
F=W ' Z ' (4)
Wherein, Z is the sEMG signals obtained by EMG signal sensor, and Z ' is the characteristic signal of sEMG signals, and W is to be
System transfer matrix, W ' is W pseudo inverse matrix, and F is the control signal for the motion of control operation object.
Therefore, in order to obtain the control signal of the synchronization of control operation object and motion in direct ratio, core is
The system transfer matrix W for representing the relation between sEMG signals and control signal is obtained, this will be retouched in detail below
State.
Wherein, Z ' is the feature of the sEMG signals obtained by EMG signal sensor, and F is to be used to control multiple degrees of freedom
The control signal of action, W are system transfer matrixes.
In wearable electronic according to embodiments of the present invention, the root mean square of sEMG signals can be used as EMG
The feature of signal, i.e. Z '=√ Z.
Certainly, it will be understood by those skilled in the art that other features of sEMG signals can also be used, for example, sEMG believes
Number temporal signatures or auto-regressive parameter.Wherein, auto-regressive parameter is the parameter that an autoregression model is modeled to obtain, can
To play whitening action, with improved effect when with sEMG Signal approximation system transfer matrixes.
Before being controlled to the action of the multiple degrees of freedom of operation object, sparse non-negative matrix factorisation can be passed through
(sparse non-negative matrix factorization：SNMF) solves the sEMG that continuous control signal F is provided
The Factorization problem of signal, therefore, sNMF schemes can be referred to as again.
In wearable electronic according to embodiments of the present invention, system is obtained by the training process carried out in advance
Transfer matrix W.Specifically, in the training process, for example, by bio-signals amplifier (EMGUSB2, OT
Bioelettronica, Italy) with 2048Hz sampling rate record sEMG signals.Training object is instructed to carry out a series of
The upset (DOF1) of wrist movement, i.e. wrist or the rotation (DOF2) of wrist, and both combinations.It is randomly chosen these
The order of action.So, the combination of single-degree-of-freedom action and multiple degrees of freedom action provides one group of sEMG signal.The training process
The sEMG signals of middle record are used to train sNMF algorithms, to be use up by the non-negative matrix factorisation to sENG signals
Measure two nonnegative matrixes of approximate sENG signal matrix：System transfer matrix W and control signal matrix F.Updated by iteration
System transfer matrix W and control signal matrix F, wherein so that it is basic that control signal matrix F meets that the row vector of matrix corresponds to
The activation of free degree action, so as to control operation object.By taking the motion of the upset of above-mentioned wrist and rotation as an example, control signal F
=[F1 +；F1 -；F2 +；F2 -；]T, F1 +Represent the motion that wrist is turned up, F1 -Represent the motion of wrist varus, F2 +Represent wrist up time
The motion of pin rotation, and F2 -Represent the motion of wrist rotate counterclockwise.So, obtain representing that sEMG signal matrix Z and control are believed
The system transfer matrix W of transformational relation between number matrix F.
In addition, system transfer matrix can also be obtained by linear recurrence (LR) method.Specifically, will correspond to each
The control signal combination of the free degree is designated as matrix F, and the sEMG signals detected from each free degree combination is designated as into matrix Z, then
Can be by matrix F and Z come estimating system transfer matrix W, as shown in below equation：
W=(FFT+λI)-1FTZ (5)
Wherein I is unit matrix, and λ is regular parameter.It is to obtain that optimal λ, optimal λ can be selected by cross validation
The parameter of minimum average signal-to-noise ratio.
That is, by using sNMF schemes, control signal F is estimated by the sNMF of multichannel sEMG signals, with
Find two the nonnegative matrixes W and F that its product is the good approximation of recorded multichannel sEMG signal matrix.
For the sEMG signals of N- channel T- length, its root-mean-square value is expressed as Z, wherein t row are the sEMG in time t
Signal, and represent with m DOF number.As described above, because each free degree can be further broken into positively and negatively, institute
So that Z can be represented by the product of the non-negative control signals matrix Fs of the non-negative system transfer matrix W and 2m × T of N × 2m, such as with lower section
Described in formula：
ZN×T≈WN×2mF2m×T (6)
In sNMF schemes according to embodiments of the present invention, robustly and simultaneously separation can be identified in a manner of standard is unsupervised
The Factorization of the basic function of generation.By using the program, training with calibration phase, object need not simultaneously follow pre-
The order of definition acts to activate single-degree-of-freedom, and can even activate the action of the more than one free degree simultaneously.In addition, the party
Case can also carry out extraction system transfer matrix with a step.
In sNMF schemes, the solution addition to Factorization constrains and especially needs the solution to maximize output control function
It is openness.Sparsity constraints limit the space of possible NMF solutions.Especially, the solution with basic function corresponds to single-degree-of-freedom,
It is target solution, and most sparse in other infinite solutions.So, have constrained Factorization and do not need default calibration
Stage or the activation of single-degree-of-freedom, and can apply to user execution there is multivariant any task.
Therefore, by generating the sNMF schemes of sparse solution, it may not be necessary to activate the specific collection of generation by single-degree-of-freedom
Calibration data.As noted previously, as in muscle space, multiple degrees of freedom action can be broken down into the line of single-degree-of-freedom action
Property combination.So, the control signal of multiple degrees of freedom action is controlled to be broken down into the control letter of control single-degree-of-freedom action
Number linear combination.So by the way that sparsity constraints are applied into control signal, can be determined using sparsity constraints sparse
Solution.
Openness degree is mathematically generally controlled by l1 norms and l0 norms.For convenience of calculation, it have selected and be based on
The SNMF methods of l1 norms, the object function of the sNMF methods are represented by below equation：
Meet W, F >=0,
s.t.Wij, Fij≥0 (7)
Wherein, F (:, t) be control signal matrix F t column vectors, ' Fro ' is Ni Wusi norms, and λ>0 be balance because
The accuracy and the regular parameter of F openness degree that number decomposes.As described above, optimal λ is selected by cross validation.As above institute
State, subscript "+" and "-" represent each DOF positively and negatively.Above equation can be rewritten as：
Wherein e1×2mIt is the row vector that all items are equal to 1, and 01×TEqual to 0.Alternately non-negative least square can be passed through
(ANLS) method effectively solves equation (6), and by another in a renewal W and F come iteration in fixed W and F
One, as shown in below equation：
Above-mentioned F(k+1)And W(k+1)Solution be classical least square problem, and each solution with closed form.Institute
With, it is possible to understand that sNMF schemes can converge to static point.
By sNMF schemes, institute can be extracted from one step of record of the combination producing of any free degree by user
There is basic function.
The openness degree of control signal matrix is controlled although being described above by taking l1 norms as an example, according to
The remotely-piloted vehicle of the embodiment of the present invention is equally applicable l0 norms to be controlled to the openness degree of control signal matrix
System.Equally, it will be understood by those skilled in the art that wearable electronic according to embodiments of the present invention can also use other
Sparsity constraints control the openness degree of the control signal matrix so that the row vector of control signal matrix is corresponding
In the control signal for controlling the basic free degree, the control signal for controlling multiple degrees of freedom action is decomposed into the basic free degree of control
The linear combination of the control signal of action.
So, obtained representing that the control of control single-degree-of-freedom action is believed from the surface electromyography signal of collection due to directly
Number linear combination control signal matrix, the corresponding physiology of multiple degrees of freedom action of the control mode with performing corresponding task
Muscle activity is associated, so as to intuitively be controlled operation object in a manner of accurately simulating human muscle's activity.
So, wearable electronic according to embodiments of the present invention is by detecting the gesture of user, and will represent multiple degrees of freedom action
User gesture be decomposed into represent single-degree-of-freedom action user gesture linear combination, to be controlled to operation object,
So as to improve the achievable degree of system development.
By sNMF schemes as described above, after system transfer matrix W is obtained from one group of sEMG signal, in order to estimate
The control signal acted on DOF purpose is counted, seeks W pseudo inverse matrix, and is multiplied by the feature of the sEMG signals of new record,
So, the control signal of estimation is：
F (t)=W ' Z ' (t)
It is described by taking the jointly controlling of DOF1, DOF2 and DOF3 these three frees degree as an example, then F (t) is expressed as：
In order to ensure that no component is covered by other components, each component in F is normalized relative to its maximum.Estimate
The control signal of meter is further scaled by scaling correction factor, and this is used to explain signal power (control letter in Factorization processing
Number scope) uncertainty.
Wherein, by F1, F2And F3Respectively as DOF1, DOF2 and DOF3 control signal.Multiplication is determined for each object
Factor τijTo allow the final control signal of training stage to be mapped to the whole dimensional extent of each single-degree-of-freedom action,
Such as the whole joint angles scope of the wrist in Wrist-sport.In this way it can be ensured that the control signal obtained being capable of accurate mould
The muscle of anthropomorphic body controls the process of more DOF actions, intuitively to be controlled operation object.Then, the control letter of acquisition
Number in 6Hz (motion bandwidth) LPF, manipulating objects are then can apply to cause operation object control DOF actions.
In muscle signal domain, it can be acted as the linear combination of basic function.Best basic function is
Corresponding to single DOF, because they are associated with the physiology muscle activity for determining corresponding task.In fact, in muscle signal domain
Activity space can by the active signal corresponding to single DOF any linear combination cover.SNMF schemes can appoint from any
One group of sEMG signal behavior caused by business has a maximum openness Factorization, and this correspond directly to it is associated with single DOF
Basic function.
Except acting the advantages of also setting basic function well from single DOF even if the signal used, sNMF schemes can
Do not have the signal of the constraint from single DOF action generations for Factorization.Therefore, can for more major class signal with nothing
The mode of supervision realizes identical solution, to obtain system transfer matrix W.Therefore, the program can be used for calculating in control
The continuous estimation of the basic function during use of method, as the mode that basic function is adapted to the time.
Therefore, remotely-piloted vehicle according to embodiments of the present invention is distant to realize by choosing representational gesture motion
Control the motion control of aircraft in three dimensions.Wherein, the representative gesture except need be user habitual movement in addition to,
Also need to be in muscle space single-degree-of-freedom action, so as to meet control accuracy and Discussing Convenience both requirement.
Wrist upset, wrist rotation and palm above-mentioned hold these three basic gestures and can meet above-mentioned needs well.Separately
Outside, user easily can make the gesture motion comprising these three basic gestures with an action, so as to control remote control distributor
The motion of device in three dimensions.
Another aspect according to embodiments of the present invention, there is provided a kind of control method of remotely-piloted vehicle, including：Receive control
Signal processed, the control signal comprise at least the first row vector, the second row vector and the third line vector；Believed respectively with the control
Number the first row vector, remotely-piloted vehicle described in the second row vector and the third line vector majorization the first dimension, the second dimension and
Motion in third dimension；Wherein, the first row vector of the control signal, the second row vector and the third line vector are right respectively
The action of the first single-degree-of-freedom, the second single-degree-of-freedom action and the 3rd in the gesture motion of the user of remotely-piloted vehicle described in Ying Yu
Single-degree-of-freedom acts, and, the gesture motion is that first single-degree-of-freedom acts, second single-degree-of-freedom acts and described
The linear combination of 3rd single-degree-of-freedom action.
Fig. 6 is the indicative flowchart of the control method of remotely-piloted vehicle according to embodiments of the present invention.As shown in fig. 6,
The control method of remotely-piloted vehicle according to embodiments of the present invention includes：S1, receives control signal, and the control signal comprises at least
First row vector, the second row vector and the third line vector；S2, respectively with the first row vector of the control signal, the second row vector
With motion of the third line vector majorization remotely-piloted vehicle in the first dimension, the second dimension and third dimension；Wherein, the control is believed
Number the first row vector, the second row vector and the third line vector user that corresponds respectively to the remotely-piloted vehicle gesture motion in
The action of the first single-degree-of-freedom, the action of the second single-degree-of-freedom and the action of the 3rd single-degree-of-freedom, and, the gesture motion be first it is single from
Acted by degree, the second single-degree-of-freedom acts and the linear combination of the 3rd single-degree-of-freedom action.
In the control method of above-mentioned remotely-piloted vehicle, the dimension includes Spatial Dimension and time dimension.
In the control method of above-mentioned remotely-piloted vehicle, first single-degree-of-freedom action is wrist rotation, and second list is certainly
It is wrist upset by degree action, and the action of the 3rd single-degree-of-freedom is that palm is held.
In the control method of above-mentioned remotely-piloted vehicle, each single-degree-of-freedom acts the bag for controlling the remotely-piloted vehicle
Include at least one of kinematic parameter of the direction of motion and movement velocity.
In the control method of above-mentioned remotely-piloted vehicle, first single-degree-of-freedom action, second single-degree-of-freedom action and
3rd single-degree-of-freedom acts the side in three dimensions for being respectively used to control the remotely-piloted vehicle in three-dimensional coordinate system
To vector.
In the control method of above-mentioned remotely-piloted vehicle, first single-degree-of-freedom is acted for controlling the remotely-piloted vehicle
Angle is horizontally diverted, second single-degree-of-freedom acts the vertical duction angle for controlling the remotely-piloted vehicle, and the 3rd list is certainly
Flying speed for controlling the remotely-piloted vehicle is acted by degree.
In the control method of above-mentioned remotely-piloted vehicle, the control signal is the wearable electronic by the user
What checking with EMG method and control function were generated by the gesture motion of the user.
In the control method of above-mentioned remotely-piloted vehicle, generating the control signal by the gesture motion of the user includes：
Electromyographic signal Z is obtained, signal Z corresponds to the gesture motion of the user；
Below equation is to acquired bioelectrical signals Z processing：
Z '=W ' F (1)
Z ' is bioelectrical signals Z characteristic signal, and W ' is system transfer matrix W pseudo inverse matrix, and F is non-negative control
Signal matrix；
Wherein, the system transfer matrix W is obtained by the sparse nonnegative integer factorising algorithm in training process, and should
The row vector of non-negative control signals matrix is synchronous in direct ratio with the action of corresponding single-degree-of-freedom.
In the control method of above-mentioned remotely-piloted vehicle, the system is obtained by the training process comprised the following steps and shifted
Matrix W：So that at least one multiple training that training object include in multiple degrees of freedom action and single-degree-of-freedom action are moved
Make；The electromyographic signal detected from each training action combination is designated as electromyographic signal matrix Z1；Using sparse nonnegative integer because
Electromyographic signal matrix Z1 is decomposed into non-negative system transfer matrix Wi and sparse non-negative control signals matrix by number decomposition algorithm
Fi, i represent iterations；Non- negative system transfer matrix Wi and sparse non-negative control signals matrix F i are updated by iteration,
The wherein non-negative control signals matrix F i each row vector represents the single-degree-of-freedom action in one of joint of the arm；
With the non-negative system transfer matrix Wi after being updated is as system transfer matrix W.
In the control method of above-mentioned remotely-piloted vehicle, further comprise：Non- negative control signals are controlled based on l1 norms
The openness degree of matrix F, is represented by below equation：
Meet W, F >=0,
s.t.Wij, Fij≥0 (2)
Wherein, F (:, t) be control signal matrix F t column vectors, ' Fro ' is Ni Wusi norms, and m is detection myoelectricity letter
Number the number of channel, T is time span, λ>0 is the regular parameter of the accuracy that equilibrium factor decomposes and F openness degree, on
Mark "+" and "-" represents each free degree positively and negatively.
In the control method of above-mentioned remotely-piloted vehicle, party's formula (2) is rewritten as：
Wherein e1×2mIt is the row vector that all items are equal to 1, and 01×TEqual to 0.
In the control method of above-mentioned remotely-piloted vehicle, equation (3) is solved by alternately non-negative least square method,
And by fix a renewal come iteration in the system transfer matrix W and signal matrix F another, such as below equation
It is shown：
The wherein F(k+1)And W(k+1)Solution with closed form.
In the control method of above-mentioned remotely-piloted vehicle, the control signal is represented by below equation：
Wherein, F1Correspond to the control signal of first single-degree-of-freedom action, F2Second single-degree-of-freedom is corresponded to move
The control signal of work, and F2The control signal of the 3rd single-degree-of-freedom action is corresponded to, subscript "+" and "-" expression are each certainly
By spending positively and negatively；With
The control signal is scaled by scaling correction factor according to below equation：
Wherein, scaling correction factor τijIt is determined so that control signal F is mapped to the basic free degree action
Whole dimensional extent.
By remotely-piloted vehicle and its control method according to embodiments of the present invention, can be entered by the gesture motion of user
The motion control of row remotely-piloted vehicle, so as to improve the Discussing Convenience of manipulation.Also, due to each in the gesture motion of user
Single-degree-of-freedom action corresponds to motion of the remotely-piloted vehicle in a dimension, there is provided the accuracy of control.
Also, in remotely-piloted vehicle and its control method according to embodiments of the present invention, wearable electronic can be passed through
The checking with EMG method and control function of equipment detects the gesture motion of user, to enable control signal accurately to embody gesture
The amplitude of action and direction, so as to realize the synchronization of remotely-piloted vehicle and direct proportion control.
In addition, in remotely-piloted vehicle and its control method according to embodiments of the present invention, by most to move generation
The gesture of table controls the motion of remotely-piloted vehicle in three dimensions, user can be made intuitively to manipulate remotely-piloted vehicle,
So as to lift the manipulation of user impression.
The term " unit " herein being referred to can wrap in an embodiment of the present invention according to the context using the term
Include software, hardware or its combination.For example, software can be machine code, firmware, embedded code and application software.Further for example,
Hardware can be circuit, processor, computer, integrated circuit, integrated circuit kernel, Micro Electro Mechanical System (MEMS), passive device
Part or its combination.
It should be understood by those skilled in the art that the embodiments of the invention shown in foregoing description and accompanying drawing are only used as illustrating
And it is not intended to limit the present invention.The purpose of the present invention completely and effectively realizes.The function and structural principle of the present invention exists
Show and illustrate in embodiment, under without departing from the principle, embodiments of the present invention can have any deformation or modification.
Claims (26)
1. a kind of remotely-piloted vehicle, including：
Signal receiving unit, for receiving control signal, the control signal comprise at least the first row vector, the second row vector and
The third line vector；
Motion control unit, for respectively with the first row vector, the second row vector and the third line the vector control of the control signal
Make motion of the remotely-piloted vehicle in the first dimension, the second dimension and third dimension；
Wherein, it is winged to correspond respectively to the remote control for the first row vector of the control signal, the second row vector and the third line vector
The action of the first single-degree-of-freedom, the action of the second single-degree-of-freedom and the action of the 3rd single-degree-of-freedom in the gesture motion of the user of row device,
With
The gesture motion is the first single-degree-of-freedom action, second single-degree-of-freedom action and the 3rd single-degree-of-freedom
The linear combination of action.
2. remotely-piloted vehicle according to claim 1, it is characterised in that
The dimension includes Spatial Dimension and time dimension.
3. remotely-piloted vehicle according to claim 1, it is characterised in that
The first single-degree-of-freedom action is wrist rotation, and the second single-degree-of-freedom action is that wrist is overturn, and the described 3rd
Single-degree-of-freedom action is that palm is held.
4. remotely-piloted vehicle according to claim 1, it is characterised in that
Each single-degree-of-freedom acts includes the direction of motion and movement velocity at least within for control the remotely-piloted vehicle
One of kinematic parameter.
5. remotely-piloted vehicle according to claim 3, it is characterised in that
The first single-degree-of-freedom action, second single-degree-of-freedom action and the 3rd single-degree-of-freedom action are respectively used to control
Make direction vector of the remotely-piloted vehicle in three dimensions in three-dimensional coordinate system.
6. remotely-piloted vehicle according to claim 3, it is characterised in that
First single-degree-of-freedom acts the angle that is horizontally diverted for controlling the remotely-piloted vehicle, second single-degree-of-freedom
The vertical duction angle for controlling the remotely-piloted vehicle is acted, the 3rd single-degree-of-freedom is acted for controlling the remote control
The flying speed of aircraft.
7. remotely-piloted vehicle according to claim 1, it is characterised in that
The control signal is checking with EMG method and control function by the wearable electronic of the user by the user
Gesture motion generation.
8. remotely-piloted vehicle according to claim 7, it is characterised in that the control is generated by the gesture motion of the user
Signal processed includes：
Obtain electromyographic signal Z, the electromyographic signal Z and correspond to the gesture motion of the user；
According to below equation to acquired electromyographic signal Z processing：
Z '=W ' F (1)
Z ' is the characteristic signal of the electromyographic signal Z, and W ' is system transfer matrix W pseudo inverse matrix, and F is non-negative control signals
Matrix；
Wherein, the system transfer matrix W is obtained by the sparse nonnegative integer factorising algorithm in training process, and described
The row vector of non-negative control signals matrix is synchronous in direct ratio with the action of corresponding single-degree-of-freedom.
9. remotely-piloted vehicle according to claim 8, it is characterised in that obtained by the training process comprised the following steps
The system transfer matrix W：
So that training object carries out including at least one multiple training actions in multiple degrees of freedom action and single-degree-of-freedom action；
The electromyographic signal detected from each training action combination is designated as electromyographic signal matrix Z1；
The electromyographic signal matrix Z1 is decomposed into non-negative system transfer matrix Wi using sparse nonnegative integer factorising algorithm
Iterations is represented with sparse non-negative control signals matrix F i, i；
Non- negative system transfer matrix Wi and sparse non-negative control signals matrix F i are updated by iteration, wherein the non-negative control
Signal matrix Fi processed each row vector represents the single-degree-of-freedom action in one of joint of the arm；With
Sparse non-negative system transfer matrix Wi after being updated is as system transfer matrix W.
10. remotely-piloted vehicle according to claim 9, it is characterised in that apply the sparse nonnegative integer Factorization
Algorithm decomposes the electromyographic signal matrix Z1 and further comprised：
The openness degree of non-negative control signals matrix F is controlled based on l1 norms, is represented by below equation：
Wherein, F (:, t) be control signal matrix F t column vectors, ' Fro ' is Ni Wusi norms, and m is detection electromyographic signal
The number of channel, T are time spans, λ>0 be equilibrium factor decompose accuracy and F openness degree regular parameter, subscript "+"
Each free degree is represented with "-" positively and negatively.
11. remotely-piloted vehicle according to claim 10, it is characterised in that the equation (2) is rewritten as：
Wherein e1×2mIt is the row vector that all items are equal to 1, and 01×TEqual to 0.
12. remotely-piloted vehicle according to claim 11, it is characterised in that asked by alternately non-negative least square method
Solve an equation (3), and another come the renewal of iteration by one in fixation the system transfer matrix W and the signal matrix F
One, as shown in below equation：
Wherein described F(k+1)And W(k+1)Solution with closed form.
13. remotely-piloted vehicle according to claim 1, it is characterised in that
The control signal is represented by below equation：
Wherein, F1Correspond to the control signal of the first single-degree-of-freedom action, F2Second single-degree-of-freedom is corresponded to move
The control signal of work, and F2The control signal of the 3rd single-degree-of-freedom action is corresponded to, subscript "+" and "-" represent each
The free degree is positively and negatively；With
The control signal is scaled by scaling correction factor according to below equation：
Wherein, the scaling correction factor τijIt is determined so that the control signal F is mapped to the basic free degree and moved
The whole dimensional extent made.
14. a kind of control method of remotely-piloted vehicle, including：
Control signal is received, the control signal comprises at least the first row vector, the second row vector and the third line vector；
Existed respectively with remotely-piloted vehicle described in the first row vector, the second row vector and the third line vector majorization of the control signal
Motion in first dimension, the second dimension and third dimension；
Wherein, it is winged to correspond respectively to the remote control for the first row vector of the control signal, the second row vector and the third line vector
The action of the first single-degree-of-freedom, the action of the second single-degree-of-freedom and the action of the 3rd single-degree-of-freedom in the gesture motion of the user of row device,
With
The gesture motion is the first single-degree-of-freedom action, second single-degree-of-freedom action and the 3rd single-degree-of-freedom
The linear combination of action.
15. the control method of remotely-piloted vehicle according to claim 14, it is characterised in that
The dimension includes Spatial Dimension and time dimension.
16. the control method of remotely-piloted vehicle according to claim 14, it is characterised in that
The first single-degree-of-freedom action is wrist rotation, and the second single-degree-of-freedom action is that wrist is overturn, and the described 3rd
Single-degree-of-freedom action is that palm is held.
17. the control method of remotely-piloted vehicle according to claim 14, it is characterised in that
Each single-degree-of-freedom acts includes the direction of motion and movement velocity at least within for control the remotely-piloted vehicle
One of kinematic parameter.
18. the control method of remotely-piloted vehicle according to claim 16, it is characterised in that
The first single-degree-of-freedom action, second single-degree-of-freedom action and the 3rd single-degree-of-freedom action are respectively used to control
Make direction vector of the remotely-piloted vehicle in three dimensions in three-dimensional coordinate system.
19. the control method of remotely-piloted vehicle according to claim 6, it is characterised in that
First single-degree-of-freedom acts the angle that is horizontally diverted for controlling the remotely-piloted vehicle, second single-degree-of-freedom
The vertical duction angle for controlling the remotely-piloted vehicle is acted, the 3rd single-degree-of-freedom is acted for controlling the remote control
The flying speed of aircraft.
20. the control method of remotely-piloted vehicle according to claim 14, it is characterised in that
The control signal is checking with EMG method and control function by the wearable electronic of the user by the user
Gesture motion generation.
21. the control method of remotely-piloted vehicle according to claim 20, it is characterised in that moved by the gesture of the user
Making the generation control signal includes：
Obtain electromyographic signal Z, the electromyographic signal Z and correspond to the gesture motion of the user；
According to below equation to acquired electromyographic signal Z processing：
Z '=W ' F (1)
Z ' is the characteristic signal of the electromyographic signal Z, and W ' is system transfer matrix W pseudo inverse matrix, and F is non-negative control signals
Matrix；
Wherein, the system transfer matrix W is obtained by the sparse nonnegative integer factorising algorithm in training process, and described
The row vector of non-negative control signals matrix is synchronous in direct ratio with the action of corresponding single-degree-of-freedom.
22. the control method of remotely-piloted vehicle according to claim 21, it is characterised in that pass through what is comprised the following steps
Training process obtains the system transfer matrix W：
So that training object carries out including at least one multiple training actions in multiple degrees of freedom action and single-degree-of-freedom action；
The electromyographic signal detected from each training action combination is designated as electromyographic signal matrix Z1；
The electromyographic signal matrix Z1 is decomposed into non-negative system transfer matrix Wi using sparse nonnegative integer factorising algorithm
Iterations is represented with sparse non-negative control signals matrix F i, i；
Non- negative system transfer matrix Wi and sparse non-negative control signals matrix F i are updated by iteration, wherein the non-negative control
Signal matrix Fi processed each row vector represents the single-degree-of-freedom action in one of joint of the arm；With
Sparse non-negative system transfer matrix Wi after being updated is as system transfer matrix W.
23. the control method of remotely-piloted vehicle according to claim 22, it is characterised in that application is described sparse non-negative whole
Number factorising algorithm decomposes the electromyographic signal matrix Z1 and further comprised：
The openness degree of non-negative control signals matrix F is controlled based on l1 norms, is represented by below equation：
Wherein, F (:, t) be control signal matrix F t column vectors, ' Fro ' is Ni Wusi norms, and m is detection electromyographic signal
The number of channel, T are time spans, λ>0 be equilibrium factor decompose accuracy and F openness degree regular parameter, subscript "+"
Each free degree is represented with "-" positively and negatively.
24. the control method of remotely-piloted vehicle according to claim 23, it is characterised in that the equation (2) rewrites
For：
Wherein e1×2mIt is the row vector that all items are equal to 1, and 01×TEqual to 0.
25. the control method of remotely-piloted vehicle according to claim 24, it is characterised in that pass through an alternately non-negative most young waiter in a wineshop or an inn
Multiply method to solve equation (3), and by one in fixation the system transfer matrix W and the signal matrix F come repeatedly
The renewal in generation another, as shown in below equation：
Wherein described F(k+1)And W(k+1)Solution with closed form.
26. the control method of remotely-piloted vehicle according to claim 14, it is characterised in that
The control signal is represented by below equation：
Wherein, F1Correspond to the control signal of the first single-degree-of-freedom action, F2Second single-degree-of-freedom is corresponded to move
The control signal of work, and F2The control signal of the 3rd single-degree-of-freedom action is corresponded to, subscript "+" and "-" represent each
The free degree is positively and negatively；With
The control signal is scaled by scaling correction factor according to below equation：
Wherein, the scaling correction factor τijIt is determined so that the control signal F is mapped to the basic free degree and moved
The whole dimensional extent made.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610680463.6A CN107765702A (en) | 2016-08-17 | 2016-08-17 | Remotely-piloted vehicle and its control method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610680463.6A CN107765702A (en) | 2016-08-17 | 2016-08-17 | Remotely-piloted vehicle and its control method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107765702A true CN107765702A (en) | 2018-03-06 |
Family
ID=61260357
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610680463.6A Pending CN107765702A (en) | 2016-08-17 | 2016-08-17 | Remotely-piloted vehicle and its control method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107765702A (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111598451A (en) * | 2020-05-15 | 2020-08-28 | 中国兵器工业计算机应用技术研究所 | Control work efficiency analysis method, device and system based on task execution capacity |
| WO2022141187A1 (en) * | 2020-12-30 | 2022-07-07 | SZ DJI Technology Co., Ltd. | Systems and methods for controlling an unmanned aerial vehicle using a body-attached remote control |
-
2016
- 2016-08-17 CN CN201610680463.6A patent/CN107765702A/en active Pending
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111598451A (en) * | 2020-05-15 | 2020-08-28 | 中国兵器工业计算机应用技术研究所 | Control work efficiency analysis method, device and system based on task execution capacity |
| CN111598451B (en) * | 2020-05-15 | 2021-10-08 | 中国兵器工业计算机应用技术研究所 | Control work efficiency analysis method, device and system based on task execution capacity |
| WO2022141187A1 (en) * | 2020-12-30 | 2022-07-07 | SZ DJI Technology Co., Ltd. | Systems and methods for controlling an unmanned aerial vehicle using a body-attached remote control |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CA2882968C (en) | Facilitating generation of autonomous control information | |
| CN110312471A (en) | The Adaptable System of export control signal from neuron-muscular activity measurement | |
| CN109062398B (en) | Spacecraft rendezvous and docking method based on virtual reality and multi-mode human-computer interface | |
| Väänänen et al. | Gesture driven interaction as a human factor in virtual environments–an approach with neural networks | |
| Smith | Human factors in haptic interfaces | |
| CN111189577A (en) | Sensor calibration and data measurement method, device, equipment and storage medium | |
| Rognon et al. | Soft haptic device to render the sensation of flying like a drone | |
| CN107765702A (en) | Remotely-piloted vehicle and its control method | |
| Sheridan | Human and machine haptics in historical perspective | |
| CN107766773A (en) | Various dimensions control method and control device based on gesture | |
| Müezzinoğlu et al. | Wearable glove based approach for human-UAV interaction | |
| CN111796710B (en) | Method for reproducing image contour features on touch screen | |
| CN113408443B (en) | Gesture posture prediction method and system based on multi-view images | |
| Kato et al. | Classification method of tactile feeling using stacked autoencoder based on haptic primary colors | |
| Abioye et al. | Multimodal human aerobotic interaction | |
| CN107219918A (en) | A kind of design method of data glove interface module | |
| Menezes et al. | Touching is believing-Adding real objects to Virtual Reality | |
| Amir et al. | Sensor based human arm motion and gesture replication system using MATLAB | |
| Di Capua | Inertial motion capture system for biomechanical analysis in pressure suits | |
| CN113199469B (en) | Space arm system, control method for space arm system, and storage medium | |
| Di Capua et al. | Body Pose Measurement System: System Validation and Range of Motion/Kinematic Analysis of Three Pressure Suits | |
| Saggio et al. | Inedited SVM application to automatically tracking and recognizing arm-and-hand visual signals to aircraft | |
| CN107765837A (en) | Wearable electronic | |
| Büttner et al. | Investigating users' natural spatial mapping between drone dimensions and one-hand drone controllers | |
| Heinz et al. | Sebastian Büttner sebastian. buettner@ th-owl. de Institute Industrial IT, Ostwestfalen-Lippe University of Applied Sciences and Arts Lemgo, Germany |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| WD01 | Invention patent application deemed withdrawn after publication | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20180306 |