CN101898594B - Walking method for dynamic biped robot - Google Patents

Walking method for dynamic biped robot Download PDF

Info

Publication number
CN101898594B
CN101898594B CN2010102398439A CN201010239843A CN101898594B CN 101898594 B CN101898594 B CN 101898594B CN 2010102398439 A CN2010102398439 A CN 2010102398439A CN 201010239843 A CN201010239843 A CN 201010239843A CN 101898594 B CN101898594 B CN 101898594B
Authority
CN
China
Prior art keywords
supporting leg
leg
thigh
robot
walking
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.)
Active
Application number
CN2010102398439A
Other languages
Chinese (zh)
Other versions
CN101898594A (en
Inventor
赵明国
张晓悦
董浩
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.)
Tsinghua University
Original Assignee
Tsinghua University
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 Tsinghua University filed Critical Tsinghua University
Priority to CN2010102398439A priority Critical patent/CN101898594B/en
Publication of CN101898594A publication Critical patent/CN101898594A/en
Application granted granted Critical
Publication of CN101898594B publication Critical patent/CN101898594B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Abstract

The invention discloses a walking method for a dynamic biped root and belongs to the technical field of the walking control of robots. The method is characterized by comprising the following steps of: during the walking of the robot, applying a dynamic force to support feet to make the support feet actively finish the extension and contraction actions within a walking period and controlling the extension and contraction positions of the support feet to complement the potential energy of a system. The gait of the walking method is determined by five key frames and described by two angle parameters, wherein the key frames are connected with one another through a continuous smooth curve of a first order derivative. On a prototype, the robot can stably walk at different speeds under the control of the dynamic walking method.

Description

A kind of walking method for dynamic biped robot
Technical field
The present invention is based on a kind of biped power type walking method of passive walking principle, through on the robot supporting leg, adding power, realizes the open loop travel controls of dynamic biped robot.
Background technology
How to realize that quick, stabilized walking is emphasis and difficult point during the biped robot studies.At present, biped robot's traveling method mainly comprises static walking, ZMP walking, and limit cycle walking.Wherein static walking be occur the earliest also be most basic a kind of traveling method; It requires, and the barycenter of robot remains in the polygon that both feet constitute on the ground in the walking process; This method is easy to keep the stable of robot, but has also limited the speed of travel of robot greatly.The zero moment point of ZMP theory demand robot remains in the polygon of both feet formation; This method has reduced artificial constraint than the static state walking to a certain extent;, model machine obtained great success on using; The ASIMO that comprises Honda company, the HRP4 of Japanese AIST research institute, and the Qrio of Sony etc.Yet, be difficult to have again breakthrough at aspects such as natural gait and energy efficiencies according to the robot of ZMP Design Theory.
The limit cycle walking is a kind of new walking theory that occurs in recent years; Its proposition has received the inspiration of human walking; Requiring periodic gait sequence is orbitally stable; Be that gait sequence can form a stable limit cycle in state space, but any instantaneous local stability that do not have in gait cycle.This method is less to the artificial constraint of robot, has utilized the dynamics of robot self under the gravitational field fully, thereby has energy efficiency, speed and antijamming capability that robot ambulation is improved in bigger space.At present; The examples of many successful of employing limit cycle walking principle comprises Spring Flamingo and the dummy model control method thereof of MIT; The Rabbit of France academy of sciences and mixing zero dynamic control method thereof; People's such as Geng RunBot and nervous centralis control method thereof, and the biped robot of CMU and encourage learning method etc. again.These robots have been realized bigger breakthrough at aspects such as the speed of travel, energy efficiency and antijamming capabilities, but gait generation method is comparatively loaded down with trivial details, and some then need use machine learning, to having relatively high expectations of experimental situation.
Passive walking is a kind of prominent example of limit cycle walking, and robot is walked downwards along little slope of inclining, and need not apply any control, and the gravitional force that the slope provides is converted into the required kinetic energy of robot ambulation.The gait that passive walking generates is very natural, and energy efficiency can reach human level, is 1/tens of ZMP walking robot ASIMO approximately.For passive walking is realized on the level land, Cornell university has used the method that increases power at the robot ankle place, lead leg and the ground hind paw that bumps is pedaled ground in per step, and be walking injection energy.Deflt university has then adopted in the way of leading leg with collision on the ground front clamp hip joint, has reached the purpose of mending energy equally.But the energy of above two kinds of methods is mended into all being positioned at collision front and back constantly opportunity; Energy is that instantaneous benefit is gone into; Requirement has high energy density; Therefore to a great extent limit the speed of travel of robot, this energy is mended into method and can be caused bigger disturbance to gait simultaneously, has reduced the stability of walking.
Among the former patent ZL200810116148.6 of this seminar a kind of level land power traveling method has been proposed; It through in walking on one's own initiative elongate support leg and shortening lead leg the potential energy of replenishment system barycenter, realize energy supplement and the balance of loss in the walking.Power biped robot traveling method of the present invention is to be the basis with passive walking principle, initiatively shortens the potential energy that supporting leg comes the replenishment system barycenter again through in a walking is walked, initiatively extending earlier.This method is compared with the way in the patent before; Eliminated equally and led leg and collision on the ground moment is mended the influence that walking stability is caused into energy; Can make robot reach higher walking stability; And it also only needs open loop control, and realization is simple and calculated amount is very little, therefore is applicable to the occasion that real-time is had relatively high expectations.In addition, it is being added on the supporting leg in the control sets, and need not control leading leg, and has reduced the variable of control, and this control method can make system mend into energy be a fixed value, thereby have the advantage of the progressive balance of energy.
Summary of the invention
The objective of the invention is to from passive walking principle, propose a kind of method and realize the power type walking method of biped robot on the level land at adding power on the supporting leg.
Biped robot's model of the present invention is as shown in Figure 1, and wherein 1 is the robot health, and quality is M; 2 is the supporting leg thigh; 3 is the supporting leg shank, and 4 are equivalent supporting leg (top by supporting leg thigh 2 constitutes to supporting leg shank 3 terminal lines), and 5 for leading leg thigh; 6 are the shank of leading leg, and 7 are equivalence lead leg (being made up of to the shank 6 terminal lines of leading leg the top of the thigh 5 of leading leg).Robot has three angle θ, α, β, and wherein θ and α are the critical angle of control robot walking, and θ is equivalent supporting leg 4 and equivalent angle of leading leg between 7; α is the angle between supporting leg thigh 2 and the equivalent supporting leg 4, determines the length of equivalent supporting leg 4, α>0 when the supporting leg knee bends, α when the supporting leg knee stretches=0; β leads leg 5 and lead leg angle between 7 of equivalence, the length that the decision equivalence is led leg, and β when leading leg knee bends>0, β=0 when the swing knee stretches does not need control in the process of walking owing to lead leg, so β keeps a constant.The angle between the convenient definite two equivalent legs of definition critical angle like this and the length of two equivalent legs.One walking of robot is walked by swing process and collision to form, and wherein swing process refers to that robot supporting leg end lands, and is axially preceding swing with the end, and leading leg simultaneously swings to supporting leg the place ahead by the supporting leg rear aloft; Collision refers to lead leg when swing process finishes terminal and the instantaneous time collision takes place on ground, and supporting leg is liftoff simultaneously.Leading leg after the collision converts supporting leg into, and supporting leg converts into leads leg.One walking of robot is walked by beginning after the last step collision, after swing process to collision, finishes.
The energy conversion principle of biped robot's traveling method of the present invention is as shown in Figure 2, and for the effect of outstanding supporting leg, only at two moments of the whole story in step two legs that draws, middle constantly all having omitted led leg among the figure.Robot is started to walk from A constantly, and this moment equivalent supporting leg and equivalence led leg isometric, and it will be soon liftoff to lead leg, liftoff after, supporting leg freely swings to moment B.Robot in a step swing process through stretching supporting leg knee joint (from moment B to moment C) earlier, bent support leg knee joint (from moment C to moment D) again then, thus reach elongation and shorten the length of equivalent supporting leg 4, come the potential energy of replenishment system.Remain unchanged behind the supporting leg knee joint bending, continue swing forward, two legs bump with a fixing angle and ground, accomplish the walking of one-period.The process that supporting leg stretches is to mend the process into energy to system, and system capacity increases E 1In order to guarantee that two legs are isometric when colliding; And supporting leg can not ad infinitum extend, so in the walking process in a step, supporting leg is also wanted crooked Hui Yuanchang before bumping with ground; The process that supporting leg is crooked is the process that system capacity reduces, and system capacity reduces E 2In order to make robot gross energy increase, i.e. E in the walking process in a step 1And E 2Poor E cGreater than zero, the position that the supporting leg of must strictness controlling well stretches and shortens, promptly guarantee position that supporting leg stretches than the position of shortening more near vertical position, the gross energy of system in swing process increased.In the method that this energy is mended; Only considered that the power that adds on the supporting leg mends the effect into energy to system; Because we have only considered mass concentration at hip, have ignored the model of shank quality, lead leg to the not influence of energy of system; So in walking, keep crooked constant, and hypothesis is led leg, and can to put with the supporting leg angle be the lock hip state of a fixed value.
Biped robot of the present invention has the characteristic of the progressive balance of energy, and the performance of this specific character is following.Walk in the process in a walking, the increase of system capacity is only stretched with bent position relevant, and irrelevant with the initial speed of supporting leg with supporting leg.When the robot supporting leg stretch with bent position fixing after; It mends the energy of going in each step of walking be exactly a fixing constant; When the energy of mending the energy of going into and the collision damage of leading leg is the same; The walking of robot just gets into limit cycle, and the starting velocity after the collision is the speed at fixed point place.If the initial speed of robot is greater than the speed at fixed point place; The energy of collision damage is proportional to rate of onset owing to lead leg; So the energy of collision damage greater than mend into energy, thereby next step starting velocity is reduced, up to equating with fixed point place speed; Otherwise, if the initial speed of robot less than the speed at fixed point place, the energy of collision damage less than mend into energy, thereby next step starting velocity is increased, equate up to speed with the fixed point place.
Biped robot's traveling method of the present invention is characterised in that, contains following steps successively:
Step (1) is constructed a biped robot, and shown in Fig. 3 (b), its step is following:
Step (1.1); Set up being connected of trunk 1 and first thigh 4 and second thigh 5: this trunk 1 is captiveed joint respectively with the first hip joint motor 2 of the coaxial placement in the left and right sides and the body of the second hip joint motor 3; And the rotary output axis of the said first hip joint motor 2 is connected with said first thigh 4; The rotary output axis of the said second hip joint motor 3 is connected with said second thigh 5
Step (1.2); Set up said first thigh 4 and first shank 7; Second thigh 5 is connected with second shank 9: the end of this said first thigh 4 is connected with the body-fixed of the first knee joint motor 6, and the rotary output axis of this first knee joint motor 6 is connected with first shank 7, and the end of this said second thigh 5 is connected with the body-fixed of the second knee joint motor 8; The rotary output axis of this second knee joint motor 8 is connected with second shank 9
Step (1.3) all adopts servomotor at four motors described in step (1.1), the step (1.2), uses S respectively Hip1, S Hip2Represent the anglec of rotation of the said first hip joint motor 2 and the second hip joint motor 3, use S respectively Knee1, S Knee2The anglec of rotation of representing the said first knee joint motor 6 and the second knee joint motor 8, and with said four motors of PC control, wherein: said S Hip1Be the angle of said first thigh 4 with trunk 1 vertical direction, S when the end of this first thigh 4 is positioned at trunk 1 the place ahead Hip1>0, and S when being positioned at trunk 1 rear Hip1<0, said S Hip2Be the angle of said second thigh 5 with trunk 1 vertical direction, S when the end of this second thigh 5 is positioned at trunk 1 the place ahead Hip2>0, and S when being positioned at trunk 1 rear Hip2<0; Said S Knee1Be the angle between said first shank 7 and first thigh 4, when this first shank 7 with respect to said first thigh 4 S during to rear curved Knee1>0, S when both are parallel Knee1=0, said S Knee2Be the angle between said second shank 9 and second thigh 5, when this second shank 9 with respect to said second thigh 5 S during to rear curved Knee2>0, S when both are parallel Knee2=0,
Step (1.4), the signal input end of each motor links to each other with the control signal output ends of a upper computer respectively described in this step (1.1), step (1.2), the step (1.3);
Step (2) is set a gait cycle T in said upper computer, said gait cycle T refers to collide the time of being experienced to leading leg from a zero hour in step; Wherein, the zero hour, t=0 was meant the liftoff moment of leading leg, and collision constantly t=T is meant and leads leg and ground bumps; Promptly a gait cycle finishes, the moment that next gait cycle begins, at this in a flash; Supporting leg before becomes leads leg, and leading leg before becomes supporting leg.
Step (3) in said upper computer, in a said gait cycle, is provided with five key frames, and two crucial joint angles θ and α are as shown in Figure 4:
First key frame (figure A), when being positioned at t=0, the initial attitude in one step of decision robot, wherein, θ=-θ 0, θ 0Be a non-negative constant, the angle during expression t=0 between said two equivalent legs, it has determined the size of stride; α=α 0, α 0Be a non-negative constant, expression t=0 is the angle of bend of said supporting leg thigh 2 with respect to equivalent supporting leg 4.
Second key frame (figure B) is positioned at t=T 1The time, wherein: α=α 0, identical with α in first key frame, represent that the length of equivalent supporting leg 4 remains unchanged between first key frame and second key frame.What this key frame was represented is the starting point that supporting leg stretches.
The 3rd key frame (figure C) is positioned at t=T 2The time, wherein: α=0, expression supporting leg knee joint stretches, and the hip barycenter rises to vertex.From three key frames of second key frame to the is to mend the process into energy to system, and since this moment the knee joint of leading leg also be in case of bending, so can avoid leading leg in the process of walking shank and ground bump.
The 4th key frame (figure D) is positioned at t=T 3The time, wherein: α=α 0, expression supporting leg knee joint bending returns previous status, is processes that supporting leg shortens from four key frames of the 3rd key frame to the, and this is that two legs are isometric in order to guarantee to collide, and becomes at next step supporting leg and to keep case of bending when leading leg.
The 5th key frame (figure E), when being positioned at t=T, decision collision robot pose constantly, wherein: θ=θ 0, the expression collision is the angle between the two equivalent legs constantly; α=α 0, from the 4th key frame to the five key frames, supporting leg remains unchanged.After bumping with ground, supporting leg becomes leads leg, and leading leg becomes supporting leg.
Step (4.1), said upper computer are controlled said robot ambulation successively according to the following steps, set the cycle T that every row makes a move, three material time T 1, T 2, T 3, be t computing time, and t is since 0, and said upper computer is calculated as follows every θ at a distance from the t time, the value of α
(1)
&alpha; = &alpha; 0 0 &le; t mod T < T 1 &alpha; 0 2 cos &pi; ( t - T 1 ) T 2 - T 1 + &alpha; 0 2 T 1 &le; t mod T < T 2 &alpha; 0 2 cos &pi; ( t - T 3 ) T 2 - T 3 + &alpha; 0 2 T 2 &le; t mod T < T 3 &alpha; 0 T 3 &le; t mod T < T
(2)
Can obtain θ=f by following formula θ(t), α=f α(t) two curves, they are the curve of hip angle and supporting leg knee joint bending angle respectively, and are as shown in Figure 5, variable θ, α is continuous about the first derivative of t.
Step (4.2); In above-mentioned upper computer, be calculated as follows when step number is n (n is t and the T integer part of division result mutually), at above-mentioned walking parameter θ; The α and the kneed angle of bend β that leads leg (because supporting leg just remains unchanged becoming the patella of leading leg, i.e. β=α 0) value under, the anglec of rotation of said each motor of biped robot.When n is odd number, S Hip1, S Knee1Be respectively said supporting leg hip joint and kneed angle, S Hip2, S Knee2Be respectively said hip joint and the kneed angle of leading leg, when n is even number, S Hip2, S Knee2Be respectively said supporting leg hip joint and kneed angle, S Hip1, S Knee1Be respectively said hip joint and the kneed angle of leading leg, that is to say when n is become this odd number and added 1 formed even number by odd number, as the supporting leg of first thigh with walk to finish the back as leading leg of second thigh in a walking and exchange.
S hip 1 = - 1 2 &theta; + &alpha; S hip 2 = 1 2 &theta; + &beta; S knee 1 = 2 &alpha; S knee 2 = 2 &beta; , n = 1,3,5,7,9 . . .
(3)
S hip 1 = 1 2 &theta; + &beta; S hip 2 = - 1 2 &theta; + &alpha; S knee 1 = 2 &beta; S knee 2 = 2 &alpha; , n = 2 , 4 , 6 , 8,10 . . .
(4)
The parameter that gait method of designing of the present invention relates to mainly is time variable T, T 1, T 2, T 3With θ, two angles of α, in operating process, need regulate parameter accordingly according to the actual result of robot.
Power biped robot traveling method of the present invention is to be the basis with passive walking principle, initiatively shortens the potential energy that supporting leg comes the replenishment system barycenter again through in a walking is walked, initiatively extending earlier.This method has been eliminated and has been led leg and collision on the ground moment is mended the influence that walking stability is caused into energy; Can make robot reach higher walking stability; And it only needs open loop control, and realization is simple and calculated amount is very little, therefore is applicable to the occasion that real-time is had relatively high expectations.In addition, it is being added on the supporting leg in the control sets, and need not control leading leg, and has reduced the variable of control, and this control method can make system mend into energy be a fixed value, thereby have the advantage of the progressive balance of energy.
Description of drawings
Fig. 1 is robot model's a scheme drawing.
Fig. 2 has illustrated among the figure for the energy conversion principle figure of walking that process is walked in a walking and supporting leg is mended into energy and pendulum impact expended energy principle.
Fig. 3 is the scheme drawing of robot mechanism figure and four crucial motor corners, and wherein Fig. 3 (a) illustrates the robot lateral plan, and the robot front elevation (b) is shown.
Fig. 4 is a gait key frame scheme drawing, and A is first key frame, and B is second key frame, and C is the 3rd key frame, and D is the 4th key frame, and E is the 5th key frame.
Fig. 5 is the change curve of 2 angled key, wherein realizes illustrating the track of θ, is shown in dotted line the track of α.
Fig. 6 is power traveling method realization flow figure.
The specific embodiment
Fig. 3 is the scheme drawing of robot construction used herein and four joint angles; The realization of biped robot's traveling method according to the invention needs five key frames; As shown in Figure 4, in these five key frames, first key frame has determined the attitude of a step initial time; The 5th key frame has determined step collision attitude constantly, has also determined the size of stride simultaneously.Second key frame to the, three key frames are processes of supporting leg elongation; Determined benefit to go into what of energy; The 3rd key frame to the four key frames are processes that supporting leg shortens; To be that two legs are isometric in order guaranteeing to collide, and to make supporting leg become to keep the state that shortens when leading leg that this process is the process of degradation of energy at next step; In order to guarantee that the gross energy of system increases before collision; The energy that the energy of going into must shorten loss greater than supporting leg is mended in supporting leg elongation, and this just needs the relative position of strict control second, third and the 4th these three key frames, as long as the position of elongation more just can guarantee that near vertical position the gross energy of system increases with respect to the position of shortening.
The above parameter T, T 1, T 2, T 3And θ 0, α 0By following mode value.
Rule of thumb provide one group of initial value, confirm T earlier 1, T 2, T 3These three values constantly, this be for the position that guarantees the supporting leg elongation than the position of shortening more near vertical position, for ease, generally all get T 2Key frame is vertical position, i.e. T 2=T/2, T 1And T 3As long as satisfy T 2-T 1<T 3-T 2That's all, then actual robot is operated, keeping T, T 1, T 2, T 3Under the unmodified situation, begin during from t=0 to fall down forward after the walking, α is described like said robot 0Excessive, it is excessive to mend the energy of going into, and makes α 0Subtract 1 °, so repeat, till this robot can be walked; Begin during from t=0 to fall down backward after the walking like said robot, α is described 0Too small, it is too small to mend the energy of going into, and makes α 0Add 1 °, so repeat, till this robot can be walked.
If adjustment α 0Robot ambulation can not be made, then other parameter need be adjusted, if robot falls forward, also can be through reducing T 1(supporting leg is stretched further from vertical position) or reduce T3 (make position that supporting leg shortens more near vertical position) reduce system mend into energy; If robot falls backward, also can be through increasing T 1Perhaps T 3Come the increase system mend into energy.At last can also be through selecting θ again 0And T, come adjustment data according to above step again, make robot can on the level land, realize stable different strides, the walking of different cycles.
In walking, angled key θ shown in Figure 1 and the value of α are calculated according to above-described key frame, and its path of motion is the junction curve between the key frame.Continuous for the cireular frequency that guarantees each joint of robot, use smooth curve to connect between the key frame, promptly the first derivative of curve is continuous.Fig. 5 has provided a kind of use straight line and trigonometric function curve ways of connecting, but scope of the present invention is not limited to this connection mode.
The iterative process that is embodied as following two steps of biped robot's traveling method of the present invention: (1) PC control walking stage; (2) the artificial parameter regulation stage.In the stage (1), at first set one group of θ by manual work 0=50 °, α 0=25 °, T=0.4s, T 1=0.1s, T 2=0.2s, T 3=0.36s; After confirming these parameter values, just use smooth curve to connect key frame as the computing machine of upper computer, calculate each θ and α value constantly; Be converted into four motor angle values of robot afterwards again through (3) and (4) formula; And send to each motor on the robot, make its path of motion rotation, to realize the walking action by appointment.When n is odd number, the S among Fig. 3 Hip1, S Knee1Be supporting leg joint angle, S Hip2, S Knee2Be the joint angle of leading leg, S when n is even number Hip2, S Knee2Be supporting leg joint angle, S Hip1, S Knee1Be the joint angle of leading leg.Treating that robot places begins after the action it to be put in ground running and to observe the walking effect in the air, can not walk if robot is fallen down, and then changes the stage (2) over to.Under the parameter of above-mentioned setting, robot can fall when on the level land, walking backward.
Stage (2) is the whole stage of people's wage adjustment, in this stage, according to the walking effect of robot in the stage (1) hand adjustment is carried out in the parameter setting in the upper computer by the people, and its control method is following: at first that adjustment is two critical angle, wherein α 0Determined the size that walking mends energy,, can make α if fall down forward behind the robot ambulation in the stage (1) 0 Subtract 1 °; If robot is fallen down backward, can make α 0Add 1 °.If robot can be walked, explain that the walking condition satisfies, but fine tuning α 0Optimize collision constantly.If adjustment α 0Robot ambulation can not be made, then other parameter need be adjusted, if robot falls forward, also can be through reducing T 1Perhaps reduce T 3Reduce system mend into energy; If robot falls backward, also can be through increasing T 1Perhaps T 3The increase system mend into energy.At last can also be through selecting θ again 0And T, come adjustment data according to above step again, make can be on the level land stable walking of robot.Under the parameter of in the stage (1), setting, robot begins and can fall forward, through reducing α 0, work as α 0When being reduced to 16 °, robot has been realized the stabilized walking on the level land.

Claims (5)

1. a walking method for dynamic biped robot is characterized in that, contains following steps successively:
Step (1) is constructed a biped robot, and its step is following:
Step (1.1), set up being connected of trunk and first thigh and second thigh:
This trunk is captiveed joint respectively with the first hip joint motor of the coaxial placement in the left and right sides and the body of the second hip joint motor; And the rotary output axis of the said first hip joint motor is connected with said first thigh; The rotary output axis of the said second hip joint motor is connected with said second thigh
Step (1.2) is set up said first thigh and first shank, and second thigh is connected with second shank:
The end of this said first thigh is connected with the body-fixed of the first knee joint motor, and the rotary output axis of this first knee joint motor is connected with first shank,
The end of this said second thigh is connected with the body-fixed of the second knee joint motor, and the rotary output axis of this second knee joint motor is connected with second shank,
Step (1.3) all adopts servomotor at four motors described in step (1.1), the step (1.2), uses S respectively Hip1, S Hip2Represent the anglec of rotation of the said first hip joint motor and the second hip joint motor, use S respectively Knee1, S Knee2The anglec of rotation of representing the said first knee joint motor and the second knee joint motor, and with said four motors of PC control, wherein:
Said S Hip1Be the angle of said first thigh and trunk vertical direction, S when the end of this first thigh is positioned at trunk the place ahead Hip1>0, and S when being positioned at the trunk rear Hip1<0,
Said S Hip2Be the angle of said second thigh and trunk vertical direction, S when the end of this second thigh is positioned at trunk the place ahead Hip2>0, and S when being positioned at the trunk rear Hip2<0,
Said S Knee1Be the angle between said first shank and first thigh, S when this first shank is crooked backward with respect to first thigh Knee1>0, S during both same straight lines Knee1=0,
Said S Knee2Be the angle between said second shank and second thigh, S when this second shank is crooked backward with respect to second thigh Knee2>0, S during both same straight lines Knee2=0,
Step (1.4), the signal input end of each motor links to each other with the control signal output ends of a upper computer respectively described in this step (1.1), step (1.2), the step (1.3);
Step (2) is set a gait cycle T in said upper computer, said gait cycle T refers to collide the time of being experienced to leading leg from a zero hour in step; Wherein, t=0 was meant and regarded leading leg of second thigh liftoff moment as the zero hour, and collision t=T constantly is meant and leads leg and ground bumps; Promptly a gait cycle finishes, the moment that next gait cycle begins, at this in a flash; The supporting leg that is regarded as second thigh this moment becomes leads leg, and leading leg before becomes supporting leg, and said walking parameter comprises: θ, α, β; Unit is an angle, wherein:
θ; Be equivalent supporting leg (4) and the equivalence angle between (7) of leading leg; Said equivalent supporting leg (4) is used by the top of supporting leg thigh (2) and is represented to the terminal line of this supporting leg shank (3); Equivalence lead leg (7) use by the top of the thigh of leading leg (5) and represent to the terminal line of this shank of leading leg (6)
α is the angle between said supporting leg thigh (2) and the equivalent supporting leg (4), determines the length of equivalent supporting leg (4),
β is the said thigh of leading leg (5) and the equivalence angle between (7) of leading leg, the lead leg length of (7) of decision equivalence, and in the process of walking, leading leg remains unchanged, so β is a fixed constant,
When said equivalence lead leg (7) be positioned at equivalent supporting leg (4) θ>0 before the time, θ in the time of afterwards<0,
α>0 when the knee joint bending of said supporting leg, α when the supporting leg knee joint stretches=0,
β>0 when said knee joint bending of leading leg, the β when knee joint of leading leg stretches=0;
Step (3) in said upper computer, in a said gait cycle, is provided with five key frames, two crucial joint angles θ and α;
First key frame, when being positioned at t=0, the initial attitude in one step of decision robot, wherein, θ=-θ 0, θ 0Be a non-negative constant, expression is the angle between said two equivalent legs during t=0, has determined the size of stride; α=α 0, α 0Be a non-negative constant, expression t=0 is the angle of bend of said supporting leg thigh with respect to equivalent supporting leg,
Second key frame is positioned at t=T 1The time, wherein: α=α 0, identical with α in first key frame, represent that the length of equivalent supporting leg remains unchanged between first key frame and second key frame, what this key frame was represented is the starting point that supporting leg stretches,
The 3rd key frame is positioned at t=T 2The time, wherein: α=0, expression supporting leg knee joint stretches; The hip barycenter rises to vertex; From three key frames of second key frame to the are processes of supporting leg elongation, are to mend the process into energy to system, and knee joint also is in case of bending because lead leg this moment; Shank and ground bump so can avoid leading leg in the process of walking
The 4th key frame is positioned at t=T 3The time, wherein: α=α 0, expression supporting leg knee joint bending returns previous status, is processes that supporting leg shortens from four key frames of the 3rd key frame to the, and this is that two legs are isometric in order to guarantee to collide, and becomes at next step supporting leg and to keep case of bending when leading leg,
The 5th key frame, when being positioned at t=T, decision collision robot pose constantly, wherein: θ=θ 0, the expression collision is the angle between the two equivalent legs constantly; α=α 0, from the 4th key frame to the five key frames, supporting leg remains unchanged, and after bumping with ground, supporting leg becomes leads leg, and leading leg becomes supporting leg,
Said parameter T, T 1, T 2, T 3And θ 0, α 0Press following mode value:
The given T of elder generation, and get T 2=T/2 confirms T then 1, T 3, for the position that guarantees supporting leg elongation than the position of shortening more near vertical position, T 1And T 3Must satisfy T 2-T 1<T 3-T 2, θ 0, α 0At satisfied 30 °<θ 0<65 °, 10<α 0Given at random under<50 the situation;
Step (4), said upper computer are controlled said robot ambulation successively according to the following steps:
Step (4.1) is set the cycle T that every row makes a move, three material time T 1, T 2, T 3, be t computing time, and t is since 0, and said upper computer is calculated as follows every θ at a distance from the t time, the value of α
(1)
Figure FSB00000720256300042
(2)
Can obtain θ=f by following formula θ(t), α=f α(t) two curves are respectively the curves of hip angle and supporting leg knee joint bending angle, variable θ, and α is continuous about the first derivative of t,
Step (4.2); In above-mentioned upper computer, be calculated as follows when step number is n, n is t and the T integer part of division result mutually; At above-mentioned walking parameter θ; Under the value of the α and the kneed angle of bend β that leads leg, because supporting leg just remains unchanged becoming the patella of leading leg, so β=α 0, the anglec of rotation of said each motor of biped robot,
(3)
Figure FSB00000720256300044
(4)
When n is odd number, S Hip1, S Knee1Be respectively said supporting leg hip joint and kneed angle, S Hip2, S Knee2Be respectively said hip joint and the kneed angle of leading leg, when n is even number, S Hip2, S Knee2Be respectively said supporting leg hip joint and kneed angle, S Hip1, S Knee1Be respectively said hip joint and the kneed angle of leading leg, that is to say when n is become this odd number and added 1 formed even number by odd number, as the supporting leg of first thigh with walk to finish the back as leading leg of second thigh in a walking and exchange.
2. a kind of walking method for dynamic biped robot according to claim 1 is characterized in that, if operating personal is found following situation, handles respectively:
Said robot begins during from t=0 to fall down forward after the walking, and α is described 0Excessive, it is excessive to mend the energy of going into, and makes α 0Subtract 1 °, so repeat, till this robot can be walked,
Said robot begins during from t=0 to fall down backward after the walking, and α is described 0Too small, it is too small to mend the energy of going into, and makes α 0Add 1 °, so repeat, till this robot can be walked.
3. a kind of walking method for dynamic biped robot according to claim 1 is characterized in that, in any case if operating personal is found adjustment α 0, all can not make robot ambulation stable, then, handle respectively by following situation:
Said robot begins during from t=0 to fall down forward after the walking, explain that the energy that benefit goes into is excessive, through reducing T 1, so that supporting leg when stretching further from vertical position, perhaps reduce T 3, the position when supporting leg is shortened is more near vertical position, with reduce system mend into energy, so repeat, till this robot can be walked,
Said robot begins during from t=0 to fall down backward after the walking, explain that to mend the energy of going into too small, then through increase T 1, more near vertical position, perhaps increase T when supporting leg is stretched 3, the position when supporting leg is shortened is further from vertical position, with reduce system mend into energy, so repeat, till this robot can be walked.
4. a kind of walking method for dynamic biped robot according to claim 1 is characterized in that, at t=T, if operating personal is found following situation, handles respectively:
Said robot forward lean explains that collision too early, makes α constantly 0Subtract 0.2 °, so repeat, when collision the machine person vertically till,
The hypsokinesis of said robot health explains that collision is constantly late excessively, makes α 0Add 0.2 °, so repeat, when collision the machine person vertically till.
5. a kind of walking method for dynamic biped robot according to claim 1 is characterized in that, if no matter how operating personal is adjusted, all can not make said robot ambulation stable, perhaps can walk, but need to change leg speed, then θ 0Increase or reduce 10 °, perhaps increase T or reduce 0.1s, repeating step (2) and step (3), thus obtain different strides and gait cycle, realize the walking of friction speed.
CN2010102398439A 2010-07-28 2010-07-28 Walking method for dynamic biped robot Active CN101898594B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN2010102398439A CN101898594B (en) 2010-07-28 2010-07-28 Walking method for dynamic biped robot

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN2010102398439A CN101898594B (en) 2010-07-28 2010-07-28 Walking method for dynamic biped robot

Publications (2)

Publication Number Publication Date
CN101898594A CN101898594A (en) 2010-12-01
CN101898594B true CN101898594B (en) 2012-05-16

Family

ID=43224606

Family Applications (1)

Application Number Title Priority Date Filing Date
CN2010102398439A Active CN101898594B (en) 2010-07-28 2010-07-28 Walking method for dynamic biped robot

Country Status (1)

Country Link
CN (1) CN101898594B (en)

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102591344B (en) * 2012-03-05 2013-12-11 中国人民解放军国防科学技术大学 Time and position control method of four-footed bionic robot
ES2759298T3 (en) * 2016-02-18 2020-05-08 Hexowheel Support structure
CN109991989B (en) * 2017-12-29 2021-12-17 深圳市优必选科技有限公司 Dynamic balance method and device of robot in idle state and storage medium
CN112792807A (en) * 2020-12-24 2021-05-14 深圳市优必选科技股份有限公司 Robot control method, device, computer readable storage medium and robot
CN113934208B (en) * 2021-09-14 2023-08-04 中国北方车辆研究所 Roller-skating gait control method of driven-wheel type quadruped robot
WO2023108472A1 (en) * 2021-12-15 2023-06-22 迈宝智能科技(苏州)有限公司 Rigid-flexible mixed exoskeleton motion control method, device and system

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5311109A (en) * 1992-03-31 1994-05-10 Honda Giken Kogyo Kabushiki Kaisha Locomotion control system for legged mobile robot
CN101011985A (en) * 2007-01-26 2007-08-08 清华大学 Decoupling legs mechanism for apery robot
CN101229826A (en) * 2008-02-28 2008-07-30 清华大学 Lower limb mechanism of biped robot
CN101298260A (en) * 2008-06-11 2008-11-05 清华大学 Low-power consumption two-foot walking moving system and walking control method thereof
CN101323325A (en) * 2008-07-04 2008-12-17 清华大学 Power type walking method of dual-foot robot

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5311109A (en) * 1992-03-31 1994-05-10 Honda Giken Kogyo Kabushiki Kaisha Locomotion control system for legged mobile robot
CN101011985A (en) * 2007-01-26 2007-08-08 清华大学 Decoupling legs mechanism for apery robot
CN101229826A (en) * 2008-02-28 2008-07-30 清华大学 Lower limb mechanism of biped robot
CN101298260A (en) * 2008-06-11 2008-11-05 清华大学 Low-power consumption two-foot walking moving system and walking control method thereof
CN101323325A (en) * 2008-07-04 2008-12-17 清华大学 Power type walking method of dual-foot robot

Also Published As

Publication number Publication date
CN101898594A (en) 2010-12-01

Similar Documents

Publication Publication Date Title
CN100569579C (en) A kind of power type walking method for biped robot
CN101898594B (en) Walking method for dynamic biped robot
CN101531006B (en) Power type walking method for biped robot
CN105511465B (en) A kind of gait control method and apparatus of biped robot
US11883959B2 (en) Articulated, closed kinematic chain planar monopod
Pratt et al. Exploiting natural dynamics in the control of a 3d bipedal walking simulation
Wisse et al. Adding an upper body to passive dynamic walking robots by means of a bisecting hip mechanism
Griffin et al. Straight-leg walking through underconstrained whole-body control
Kurazume et al. Straight legged walking of a biped robot
Smit-Anseeuw et al. The energetic benefit of robotic gait selection—a case study on the Robot RAMone
CN108572553A (en) A kind of movement closed loop control method of quadruped robot
CN103407513A (en) Flat ground power type biped robot walking method adopting spring coupling
Weng et al. Natural walking with musculoskeletal models using deep reinforcement learning
Li et al. Trajectory generation of straightened knee walking for humanoid robot iCub
Kimura et al. Towards 3D adaptive dynamic walking of a quadruped robot on irregular terrain by using neural system model
CN102910218B (en) Double-feet passive walking state control method with knee bending behavior
CN112068578A (en) Method and system for adjusting stability of asymmetric rimless model passive rhythm walking
Bazargan-Lari et al. Dynamics and control of locomotion of one leg walking as self-impact double pendulum
CN114200947B (en) Humanoid gait planning method, system and device of biped robot
Griffin et al. Capture point trajectories for reduced knee bend using step time optimization
Zhao et al. Humanoid robot gait generation based on limit cycle stability
Che et al. Kinematics analysis of leg configuration of an ostrich bionic biped robot
Fu et al. A walking control strategy combining global sensory reflex and leg synchronization
Ossadnik et al. ULT-model: Towards a one-legged unified locomotion template model for forward hopping with an upright trunk
Li et al. Simulating human walking on special terrain: up and down slopes

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
C14 Grant of patent or utility model
GR01 Patent grant