CN114137998B - Biped robot balance controller based on quick ankle adjustment - Google Patents
Biped robot balance controller based on quick ankle adjustment Download PDFInfo
- Publication number
- CN114137998B CN114137998B CN202111461183.3A CN202111461183A CN114137998B CN 114137998 B CN114137998 B CN 114137998B CN 202111461183 A CN202111461183 A CN 202111461183A CN 114137998 B CN114137998 B CN 114137998B
- Authority
- CN
- China
- Prior art keywords
- robot
- expected
- zero moment
- moment point
- controller
- 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
Links
- 210000003423 ankle Anatomy 0.000 title claims abstract description 22
- 210000002683 foot Anatomy 0.000 claims abstract description 35
- 239000011159 matrix material Substances 0.000 claims description 6
- 230000001133 acceleration Effects 0.000 claims description 4
- 230000033001 locomotion Effects 0.000 abstract description 5
- 238000010586 diagram Methods 0.000 description 4
- 230000001960 triggered effect Effects 0.000 description 4
- 206010034701 Peroneal nerve palsy Diseases 0.000 description 2
- 238000000034 method Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000001105 regulatory effect Effects 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G05—CONTROLLING; REGULATING
- G05D—SYSTEMS FOR CONTROLLING OR REGULATING NON-ELECTRIC VARIABLES
- G05D1/00—Control of position, course, altitude or attitude of land, water, air or space vehicles, e.g. using automatic pilots
- G05D1/08—Control of attitude, i.e. control of roll, pitch, or yaw
- G05D1/0891—Control of attitude, i.e. control of roll, pitch, or yaw specially adapted for land vehicles
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)
- Manipulator (AREA)
Abstract
The invention provides a balance controller of a biped robot based on rapid ankle adjustment, which comprises a mass center controller, a state discriminator, a zero moment point tracker, a temporary stable state controller and a force/moment tracker, wherein the mass center controller is connected with the state discriminator; the centroid controller calculates the position of an expected zero moment point according to the position and speed feedback of the robot centroid; the state discriminator judges the state of the expected zero moment point: when the expected zero moment point position is in the stable domain, the zero moment point tracker calculates the expected contact force/moment of the feet according to the actual zero moment point position and the expected zero moment point position, and then the expected contact force/moment is tracked by using the force/moment tracker; when the position of the expected zero moment point is in a critical stable region, the temporary stable state controller pulls the expected zero moment point back to the stable region; and triggering the foot-falling point control when the position of the expected zero moment point is in the divergent domain. The invention can make the robot realize rapid balance response and realize stable motion.
Description
Technical Field
The invention belongs to the technical field, and particularly relates to a balance controller of a biped robot based on rapid ankle adjustment.
Background
When the robot performs regular walking, running and other movements, the contact force and the contact moment of the feet and the ground determine the motion state of the robot. Therefore, the balance control of the robot can be realized by the control of the foot contact force and the contact moment, and the balance control can be divided into upper-layer control and lower-layer control; according to the attitude deviation of the robot, the upper layer controls to output expected contact force and contact torque, and the lower layer controls to realize the following of the expected force and the expected torque.
The upper-layer control of the existing method belongs to feedback control of a linear system, the output quantity is a continuous value, the method is effective in dealing with general disturbance, but the regulation capability is limited in dealing with sudden large disturbance such as large external thrust or unknown ground protrusion.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides a biped robot balance controller based on rapid ankle adjustment, so that the robot can realize rapid balance response and realize stable motion.
The present invention achieves the above-described object by the following technical means.
Biped robot balance controller based on quick ankle regulation includes:
the mass center controller is used for calculating the position of the expected zero moment point according to the position and the speed feedback of the mass center of the robot;
the state discriminator is used for judging the state of the expected zero moment point, triggering the zero moment point tracker when the position of the expected zero moment point is in a stable domain, triggering the temporary stable state controller when the position of the expected zero moment point is in a critical stable domain, and triggering the foot-falling point control when the position of the expected zero moment point is in a divergent domain;
the zero moment point tracker calculates the expected contact force/moment of the feet according to the actual zero moment point position and the expected zero moment point position;
the temporary stable state controller controls the expected zero moment point at the edge of the support domain, so that the foot tilting action is generated through the force/moment tracker, and the state is quickly pulled back to the stable domain;
the control rate of the temporary stability state controller is as follows:
wherein:the position of a zero moment point is expected in the x direction of the humanoid robot,to imitate a humanThe robot plans the zero moment point position in the y direction, geo x For a humanoid robot, the geometrical parameters of the x-direction support domain, geo y For the geometrical parameters of the y-direction support domain of the humanoid robot,is the actual mass center position of the humanoid robot in the x direction,for the actual barycenter position of humanoid robot y direction, x direction and y direction belong to humanoid robot horizontal plane coordinate system, and:
wherein: parameter(s)g is gravitational acceleration, z c Is the height of the centroid of the robot, geo is the geometrical parameter of the supporting domain of the robot, x is the centroid position of the robot,is the robot centroid velocity.
In the above technical solution, the temporary stable state controller considers time optimal control:
u(t)∈U=[-geo,geo]
t 0 =0
wherein: matrix arrayMatrix arraygeo is a geometric parameter of a robot support domain, t is control time, u (t) is a control quantity, and t 0 To control the starting time, t f To control the termination time, x (t) is the state quantity, U is the support domain, x 0 Is in an initial state.
In the above technical solution, the temporary stable state controller introduces a co-modal variableDefining a Hamiltonian:
according to the coordination state condition:
obtaining by solution:
wherein: c. C 1 、c 2 Is a constant.
In the above technical solution, c 1 、c 2 The following conditions are satisfied:
①c 1 > 0 and c 2 >0:u=geo
②c 1 < 0 and c 2 <0:u=-geo
In the above technical solution, the expected zero moment point position is based onAnd calculating to obtain the result, wherein,is the desired position of the ZMP in the mold,it is the planned position of the ZMP,is the position of the actual center of mass,is the position of the planned center of mass,is the actual center of mass velocity and,is the projected centroid velocity, k c And k v Respectively, position term and speed term feedback coefficients.
In the above technical solution, the stable region is: along a polygon formed by the two-foot supporting domain, straight lines which are parallel to all side lines of the supporting domain and have a distance delta are formed towards the inner side, and a region enclosed by the left straight line and the right straight line is a stable domain; the region outside the stable region and inside the two-foot support region is a critical stable region; the area outside the double-foot supporting area is a divergent area; where δ is the stability margin.
In the above technical solution, the force/moment tracker tracks an expected contact force/moment, specifically:
according to the distribution coefficient alpha of the two feet, the expected force and moment of the left foot and the right foot are obtained:
wherein:in order to be the total desired force,in order to be able to sum up the desired torque,in order to expect the force for the right foot,in order to expect the force for the left foot,the moment is expected for the right foot,the desired moment for the left foot;
the admittance controller is designed such that the robot bipedal force/moment tracks the desired contact force/moment.
The beneficial effects of the invention are as follows: the near steady state controller considers time optimal control, introduces a covariate, and the control rate output by the near steady state controller is used for calculating total expected torque, when the expected zero torque point position is in a critical stable domain, the near steady state controller is triggered, the near steady state controller controls the expected zero torque point at the edge of a support domain, so that the foot tilting action is generated through the force/torque tracker, and the state is quickly pulled back to the stable domain.
Drawings
FIG. 1 is a block diagram of a balancing controller for a rapid ankle adjustment-based biped robot according to the present invention;
FIG. 2 (a) is a schematic illustration of the present invention at a desired zero moment point at steady state;
FIG. 2 (b) is a schematic diagram of the desired zero moment point of the present invention in a critical steady state;
FIG. 2 (c) is a schematic view of the desired zero moment point of the present invention in a divergent state;
FIG. 3 (a) is a schematic representation of a desired ZMP of the present invention directly below the centroid position;
FIG. 3 (b) is a schematic representation of the expected ZMP within the support zone with less disturbance according to the present invention;
FIG. 3 (c) is a schematic representation of the expected ZMP at the edge of the support zone with a large perturbation according to the present invention;
FIG. 4 is a controlled state diagram of the robot under the control of the temporary steady state controller according to the present invention;
fig. 5 is a schematic diagram of calculation of left and right foot force/moment distribution of the humanoid robot.
Detailed Description
The invention will be further described with reference to the following figures and specific examples, but the scope of the invention is not limited thereto.
As shown in figure 1, the balance controller of the biped robot based on rapid ankle adjustment comprises a mass center controller, a state discriminator, a zero moment point tracker, a temporary stability state controller and a force/moment tracker.
(1) And (3) calculating to obtain the position of an expected Zero Moment Point (ZMP) by the mass center controller according to the position and speed feedback of the mass center of the robot:
wherein the content of the first and second substances,is the desired position of the ZMP in the desired position,it is the planned position of the ZMP,is the position of the actual center of mass,is the position of the planned center of mass,is the actual center of mass velocity and,is the projected centroid velocity, k c And k v Respectively a position term and a speed term feedback coefficient;
(2) The state discriminator judges the state of the expected zero moment point to determine the control mode of the trigger; the types of states are shown in fig. 2 (a), (b), and (c):
the stability domain is defined as: along a polygon formed by the two-foot supporting domain, a straight line which is parallel to each side line of the supporting domain and has a distance delta (stability margin) is made inwards, and finally, a region enclosed by the left straight line and the right straight line is a stable domain; the region outside the stable region and inside the two-foot support region is a critical stable region; the area outside the two-foot support area is a divergent area. As shown in fig. 2 (a), when the desired ZMP position is within the stable region (which is now a stable state), the zero-moment point tracker is triggered and the robot can easily remain stable; as shown in fig. 2 (b), when the desired ZMP position is in the critical stability domain (in this case, the critical stability state), the critical stability state controller is triggered to quickly pull the desired ZMP back into the stability domain; as shown in fig. 2 (c), when the desired ZMP position is in the divergent domain (now divergent state), the foot-drop control should be triggered, otherwise the robot will fall, where the foot-drop control is prior art and the present invention does not take into account.
When the robot is standing stably, the ZMP is expected to be directly below the centroid position, as shown in fig. 3 (a); when the robot is disturbed, the actual center of mass position of the robot deviates from the expected center of mass position, the desired ZMP is subjected to feedback adjustment, if the disturbance is small, the desired ZMP is in the support domain, and the foot of the robot can still well fit the ground, as shown in fig. 3 (b); if the perturbation is large, the ZMP is expected to be at the edge of the support field, and the robot foot will naturally generate a foot tilting motion under the control of the foot force/moment tracker, as shown in fig. 3 (c).
(3) The principle of the zero moment point tracker is as follows: the desired contact force/moment magnitude for the feet is calculated from the actual ZMP position and the desired ZMP position, and then the desired contact force/moment is tracked using a force/moment tracker. Wherein the actual ZMP position is measured and calculated by a six-dimensional force sensor mounted on the ankle of both feet and the desired ZMP position is calculated by a centroid controller.
(4) Controller for temporary stable state
Given a quantity of state ofWhere x (t) is a function of the robot centroid position over time,as a function of robot centroid velocity over time. The clinical steady state controller pulls the desired ZMP back into the steady domain quickly, so time-optimal control is considered:
wherein: matrix arrayMatrix ofParameter(s)g is gravitational acceleration, z c Is the height of the centroid of the robot, geo is the geometrical parameter of the support domain of the robot, t is the control time, u (t) is the control quantity, t 0 To control the starting time, t f For control of the termination time, U is the support field, x 0 Is in an initial state;
according to the maximum value principle, obtaining the control rate of the temporary stability state controller:
according to the coordination state condition:
obtaining by solution:
according to constant c 1 、c 2 The positive or negative of (undetermined from the initial state) is known as λ 2 (t) there are four cases; the corresponding control law can be expressed as:
①c 1 > 0 and c 2 >0:u=geo
②c 1 < 0 and c 2 <0:u=-geo
Wherein, the first and the second end of the pipe are connected with each other,T s is in the case of (3), (4), lambda 2 (t) time to cross the horizontal axis.
According to the four conditions, the control is switched once at most, and the controller is a typical bang-bang controller; under the control of the bang-bang controller, the controlled state of the robot is shown in fig. 4.
In fig. 4:
wherein: r is - 、r + Dividing the state plane into two parts, using R respectively - 、R + Is shown, wherein:
the control quantity u is the position of the zero moment point expected by the robot, and thus the control rate of the temporary stability state controller can be expressed as:
wherein:the ZMP position is desired for the x-direction of the humanoid robot,expected ZMP position, geo, for the y-direction of the humanoid robot x For a humanoid robot, the geometrical parameters of the x-direction support domain, geo y For the geometrical parameters of the y-direction support domain of the humanoid robot,is the actual mass center position of the humanoid robot in the x direction,the position of the actual mass center of the humanoid robot in the y direction is shown; wherein the x direction and the y direction belong to a horizontal plane coordinate system of the humanoid robot.
(5) Force/moment tracker
As shown in FIG. 5, the actual ZMP position is (x) z ,y z ) The position of the left ankle is (x) L ,y L ) The right ankle position is (x) R ,y R ) From the positions of these three points, two straight lines L are calculated 1 、L 2 Wherein L is 1 Two points of the left ankle and the right ankle, L 2 Past the actual ZMP and perpendicular to L 1 Is merged with (x) 0 ,y 0 )。
L 1 The equation of (a) is:
y=ax+b
L 2 the equation of (a) is:
according to a straight line L 1 、L 2 Equation (c), calculating (x) 0 ,y 0 ):
Calculating the two-pin distribution coefficient alpha:
Wherein M is the total mass of the robot, a z Acceleration of the robot centroid along the vertical direction;
according to the distribution coefficients of the two feet, the expected force and moment of the left foot and the right foot are obtained:
in order for the robot's bipedal forces and moments to follow the desired forces and moments, the admittance controller is designed, denoted as:
wherein, the delta P is the ankle position regulating quantity,adjustment of ankle posture, F rel For the actual force on the foot, τ rel Is the actual moment, k, experienced by the foot p,f 、k d,f 、k p,τ 、k d,τ Are feedback coefficient matrices.
The present invention is not limited to the above-described embodiments, and any obvious improvements, substitutions or modifications can be made by those skilled in the art without departing from the spirit of the present invention.
Claims (7)
1. Biped robot balance control ware based on quick ankle is adjusted, its characterized in that includes:
the mass center controller is used for calculating the position of the expected zero moment point according to the position and the speed feedback of the mass center of the robot;
the state discriminator judges the state of the expected zero moment point, triggers the zero moment point tracker when the position of the expected zero moment point is in a stable domain, triggers the temporary stable state controller when the position of the expected zero moment point is in a critical stable domain, and triggers the control of the foot falling point when the position of the expected zero moment point is in a divergent domain;
the zero moment point tracker calculates the expected contact force/moment of the double feet according to the position of the actual zero moment point and the position of the expected zero moment point;
the temporary stable state controller controls the expected zero moment point at the edge of the support domain, so that the force/moment tracker generates a foot raising action to quickly pull the state back into the stable domain;
the control rate of the temporary stability state controller is as follows:
wherein:the position of a zero moment point is expected in the x direction of the humanoid robot,planning the position of the zero moment point, geo, for the y-direction of the humanoid robot x For a humanoid robot, the geometrical parameters of the x-direction support domain, geo y For the geometrical parameters of the y-direction support domain of the humanoid robot,is the actual mass center position of the humanoid robot in the x direction,for the actual barycenter position of humanoid robot y direction, x direction and y direction belong to humanoid robot horizontal plane coordinate system, and:
2. The rapid ankle adjustment based biped robot balancing controller of claim 1, wherein the critical stability state controller considers time optimal control:
u(t)∈U=[-geo,geo]
t 0 =0
5. The rapid ankle adjustment based biped robot balance controller of claim 1 wherein the desired zero moment point position is based onAnd calculating to obtain the result, wherein,is the desired position of the ZMP in the desired position,it is the planned position of the ZMP,is the position of the actual center of mass,is the position of the planned center of mass,is the actual center of mass velocity and,is the projected centroid velocity, k c And k v Respectively a position term and a speed term feedback coefficient.
6. The rapid ankle adjustment based biped robot balance controller according to claim 1 wherein the stability domain is: along a polygon formed by the two-foot supporting domain, straight lines which are parallel to all side lines of the supporting domain and have a distance delta are formed towards the inner side, and a region enclosed by the left straight line and the right straight line is a stable domain; the region outside the stable region and inside the two-foot support region is a critical stable region; the area outside the double-foot supporting area is a divergent area; where δ is the stability margin.
7. The rapid ankle adjustment based biped robot balance controller according to claim 1, wherein the force/torque tracker tracks a desired contact force/torque, in particular:
according to the distribution coefficient alpha of the two feet, the expected force and moment of the left foot and the right foot are obtained:
wherein:in order to be the total desired force,in order to obtain the total desired torque,in order to expect the force for the right foot,in order to expect the force for the left foot,the moment is expected for the right foot,the desired moment for the left foot;
the admittance controller is designed such that the robot bipedal force/moment tracks the desired contact force/moment.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111461183.3A CN114137998B (en) | 2021-12-02 | 2021-12-02 | Biped robot balance controller based on quick ankle adjustment |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111461183.3A CN114137998B (en) | 2021-12-02 | 2021-12-02 | Biped robot balance controller based on quick ankle adjustment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN114137998A CN114137998A (en) | 2022-03-04 |
CN114137998B true CN114137998B (en) | 2022-12-06 |
Family
ID=80387221
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111461183.3A Active CN114137998B (en) | 2021-12-02 | 2021-12-02 | Biped robot balance controller based on quick ankle adjustment |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114137998B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN115933723B (en) * | 2023-01-04 | 2023-08-25 | 之江实验室 | Whole-body compliant control method applied to rapid walking of biped robot |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102749919A (en) * | 2012-06-15 | 2012-10-24 | 华中科技大学 | Balance control method of multi-leg robot |
CN108858208A (en) * | 2018-09-05 | 2018-11-23 | 鲁东大学 | A kind of complicated landform anthropomorphic robot adaptive equalization control method, device and system |
CN109605364A (en) * | 2018-10-31 | 2019-04-12 | 北京理工大学 | A kind of anthropomorphic robot falls down detection and stable control method |
CN110244791A (en) * | 2019-07-11 | 2019-09-17 | 北京理工大学 | A kind of biped robot's foot power and moment follow-up control method |
CN111208809A (en) * | 2018-11-21 | 2020-05-29 | 深圳市优必选科技有限公司 | Robot gait planning method and terminal equipment |
CN111252162A (en) * | 2020-02-24 | 2020-06-09 | 北京理工大学 | Foot soft balance control system and method for biped robot |
CN112731953A (en) * | 2020-12-24 | 2021-04-30 | 深圳市优必选科技股份有限公司 | Robot control method, device, computer readable storage medium and robot |
Family Cites Families (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US8942848B2 (en) * | 2011-07-06 | 2015-01-27 | Florida Institute for Human and Machine Cognition | Humanoid robot that can dynamically walk with limited available footholds in the presence of disturbances |
US10800049B2 (en) * | 2018-02-22 | 2020-10-13 | Boston Dynamics, Inc. | Maneuvering mobile robots |
-
2021
- 2021-12-02 CN CN202111461183.3A patent/CN114137998B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN102749919A (en) * | 2012-06-15 | 2012-10-24 | 华中科技大学 | Balance control method of multi-leg robot |
CN108858208A (en) * | 2018-09-05 | 2018-11-23 | 鲁东大学 | A kind of complicated landform anthropomorphic robot adaptive equalization control method, device and system |
CN109605364A (en) * | 2018-10-31 | 2019-04-12 | 北京理工大学 | A kind of anthropomorphic robot falls down detection and stable control method |
CN111208809A (en) * | 2018-11-21 | 2020-05-29 | 深圳市优必选科技有限公司 | Robot gait planning method and terminal equipment |
CN110244791A (en) * | 2019-07-11 | 2019-09-17 | 北京理工大学 | A kind of biped robot's foot power and moment follow-up control method |
CN111252162A (en) * | 2020-02-24 | 2020-06-09 | 北京理工大学 | Foot soft balance control system and method for biped robot |
CN112731953A (en) * | 2020-12-24 | 2021-04-30 | 深圳市优必选科技股份有限公司 | Robot control method, device, computer readable storage medium and robot |
Non-Patent Citations (1)
Title |
---|
动态双足机器人的控制与优化研究进展;田彦涛;《自动化学报》;20160831;第42卷(第8期);第1142-1158页 * |
Also Published As
Publication number | Publication date |
---|---|
CN114137998A (en) | 2022-03-04 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
JP3269852B2 (en) | Posture stabilization control device for legged mobile robot | |
WO2022027914A1 (en) | Gait planning method and device for humanoid robot, and humanoid robot | |
JP2024512235A (en) | Control method and device for legged and wheeled robot, legged and wheeled robot and equipment | |
US4834200A (en) | Method and apparatus for dynamic walking control of robot | |
US7805218B2 (en) | Robot device and control method of robot device | |
CN110815211B (en) | Method for quadruped robot to dynamically cross convex obstacle | |
WO2022199080A1 (en) | Robot stepping control method and apparatus, robot control device and storage medium | |
CN113485398B (en) | Gesture control method for wheeled biped robot | |
CN111290272B (en) | Attitude stationarity adjusting method based on multi-legged robot | |
CN114137998B (en) | Biped robot balance controller based on quick ankle adjustment | |
CN113031639B (en) | Robot step processing method and device, robot control equipment and storage medium | |
CN114248855B (en) | Biped robot space domain gait planning and control method | |
CN110834685B (en) | Method for quadruped robot to dynamically cross concave obstacle | |
CN111290389A (en) | System and method for controlling foot-falling position of biped robot | |
Isvara et al. | Terrain adaptation gait algorithm in a hexapod walking robot | |
JP2006247769A (en) | Legged robot and operation control method thereof | |
CN116449711A (en) | Four-foot robot crawling state planning method and system capable of crossing large obstacle | |
CN112965508B (en) | Multi-information feedback electric parallel wheel-foot robot walking control method and system | |
Saputra et al. | Adaptive motion pattern generation on balancing of humanoid robot movement | |
Ficht et al. | Direct centroidal control for balanced humanoid locomotion | |
CN116237943A (en) | Four-foot robot control method combined with terrain constraint | |
Yan et al. | Research on multi-AUV cooperative obstacle avoidance method during formation trajectory tracking | |
Roditis et al. | Maintaining static stability and continuous motion in rough terrain hexapod locomotion without terrain mapping | |
JP2017043222A (en) | Moving device and control method of moving device | |
CN115933723B (en) | Whole-body compliant control method applied to rapid walking of biped robot |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |