Motion control method, device and equipment of foot-type robot and storage medium

Abstract

The application provides a motion control method and device for a foot-type robot. The foot-type robot comprises at least two foot ends. The method comprises: receiving a first jumping instruction for the foot-type robot when each foot end of the foot-type robot stands in a single unit area independent of each other; wherein the first jumping instruction is used to instruct the foot-type robot to jump from at least two unit areas currently located to a target unit area; and in response to the jumping instruction, controlling the foot-type robot to jump to the target unit area, so that each foot end of the foot-type robot is gathered in the target unit area, and the distance between any two foot ends of the foot-type robot in the target unit area is smaller than the distance between the corresponding two foot ends before jumping. Through the application, the dynamicity, difficulty and watchability of the foot-type robot motion can be improved, and the diversity of the foot-type robot motion is expanded.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of robot motion control, and in particular to a motion control method, device and equipment of a foot-type robot and a computer readable storage medium. BACKGROUND

[0002] The foot-type robot has been paid more and more attention due to its bionics and flexibility. The gait of the foot-type robot has been realized in the related art is very diverse, including walking, diagonal slow running, diagonal fast running, biped jumping, quadruped jumping, lateral foot walking and sprinting.

[0003] However, for the jumping gait of the foot-type robot, the jumping actions that can be realized by the foot-type robot are mostly continuous gait motions on flat ground or terrain with large foot landing area, and the jumping actions under the constraint of limited foot landing points or small foot landing contact area have not been realized. SUMMARY

[0004] The motion control method, device and computer readable storage medium of the foot-type robot provided in the embodiments of the present application can improve the dynamics, difficulty and ornamental value of the motion of the foot-type robot, and expand the diversity of the motion of the foot-type robot.

[0005] The technical solution of the embodiments of the present application is implemented as follows:

[0006] The motion control method of the foot-type robot provided in the embodiments of the present application comprises the following steps.

[0007] When each foot end of the foot-type robot stands in a unit area independent of each other, a jumping instruction for the foot-type robot is received.

[0008] The jumping instruction is used to instruct the foot-type robot to jump from at least two unit areas currently located to a target unit area.

[0009] In response to the jumping instruction, the foot-type robot is controlled to jump to the target unit area, and the distance between any two foot ends of the foot-type robot in the target unit area is smaller than the distance between the corresponding two foot ends before jumping.

[0010] The motion control device of the foot-type robot provided in the embodiments of the present application comprises the following.

[0011] The receiving module is configured to receive a jumping instruction for the foot-type robot when each foot end of the foot-type robot stands in a unit area independent of each other, wherein the jumping instruction is used to instruct the foot-type robot to jump from at least two unit areas currently located to a target unit area.

[0012] a control module configured to control the hopping robot to jump to the target cell region and, in the target cell region, a distance between any two foot ends of the hopping robot is less than a distance between the corresponding two foot ends before the jump, in response to the jump instruction.

[0013] In the above aspect, the target cell region is located in front of the hopping robot, and the control module is further configured to control the hopping robot to jump forward to the target cell region via a take-off phase, a flight phase, and a landing phase, in response to the jump instruction.

[0014] In the above aspect, the control module is further configured to control the hopping robot to sequentially perform a squatting and power accumulating operation, a foot end ground pushing operation, and a take-off starting operation in the take-off phase.

[0015] In the flight phase, the hopping robot sequentially performs an in-air foot and leg retracting operation and an in-air leg extending operation.

[0016] In the landing phase, the hopping robot sequentially performs a foot end landing operation, a leg bending and buffering operation, and a body balancing operation.

[0017] In the above aspect, the control module is further configured to control the hopping robot to bend the leg to lower a center of mass of the hopping robot from an initial height to a take-off height during the squatting and power accumulating operation.

[0018] In the above aspect, the control module is further configured to control the hopping robot to increase a torque of a leg joint to increase a friction between each foot end and a contact surface during the foot end ground pushing operation.

[0019] In the above aspect, the control module is further configured to control the hopping robot to perform a take-off action forward and upward at a target pitch angle to increase a height of a knee joint and a height of the center of mass, which is higher than the initial height, during the take-off starting operation.

[0020] In the above aspect, the control module is further configured to control the hopping robot to drive each leg to retract a corresponding foot end upward below a body center during the in-air foot and leg retracting operation.

[0021] In the above aspect, the control module is further configured to control the hopping robot to control a front leg to stretch toward the target cell region to enable each foot end of the hopping robot to land on the target cell region during the in-air leg extending operation.

[0022] In the above aspect, the control module is further configured to control the hopping robot to control each foot end to contact the contact surface in the target cell region simultaneously during the foot end landing operation.

[0023] In the process that the legged robot performs the leg bending buffering, the legged robot performs a leg bending action of bending each leg and increases a torque of a leg joint to reduce a speed of the legged robot.

[0024] In the process that the legged robot performs the body balancing, the legged robot adjusts a body posture to move a center of mass of the legged robot into the target cell region.

[0025] In the above scheme, the legged robot comprises a tail, and the control module is further configured to control the legged robot to continuously perform at least one of the following operations in the target region based on at least one degree of freedom: a head shaking action, a tail swinging action, and the like.

[0026] In the above scheme, the degree of freedom comprises at least one of roll, pitch, and yaw.

[0027] In the above scheme, the target cell region is located at the center of each cell region where the foot end is located, and the control module is further configured to control the legged robot to jump upward to the target cell region through a take-off stage, a flight stage, and a landing stage in response to the jumping instruction.

[0028] In the above scheme, the control module is further configured to control the legged robot to sequentially perform squatting and accumulating power and starting flight in the take-off stage.

[0029] In the flight stage, the legged robot performs an in-flight foot and leg folding operation.

[0030] In the landing stage, the legged robot sequentially performs a foot end landing, a leg bending buffering, and a body balancing operation.

[0031] In the above scheme, the control module is further configured to control the legged robot to bend the legs to lower the center of mass of the legged robot from an initial height to a take-off height in the process that the legged robot performs the squatting and accumulating power.

[0032] In the process that the legged robot performs the starting flight, the legged robot performs a bouncing upward to increase a height of a knee joint and a height of the center of mass, the height of the center of mass being higher than the initial height.

[0033] In the process that the legged robot performs the in-flight foot and leg folding operation, the legged robot controls a torso to be in an upright state, and each leg of the legged robot drives a corresponding foot end to be folded upward below the torso.

[0034] In the scheme, the quadruped robot, the control module is further configured to, in response to the jumping instruction, control the quadruped robot to jump to the target unit area by alternately using two front legs and two back legs when the jumping instruction indicates to perform a bipedal jump.

[0035] In the scheme, the process of the quadruped robot jumping by alternately using two front legs and two back legs includes a power storage phase, a quadruped ground-pushing phase, a front-leg-off-and-back-leg-ground-pushing phase, a flight phase, and a landing phase.

[0036] In the scheme, the unit area and the target unit area are both end face areas of a single pile in a plum-pudding pile, and the unit area and the target unit area have the same shape and size.

[0037] In the scheme, the control module is further configured to predict a jumping trajectory of the quadruped robot to obtain a reference jumping trajectory.

[0038] In the scheme, the control module is further configured to obtain a constraint condition corresponding to the jumping instruction, and the constraint condition is used to constrain the jumping of the quadruped robot.

[0039] In the scheme, the control module is further configured to determine state parameters of at least two key points of the reference jumping trajectory based on the reference jumping trajectory and the constraint condition.

[0040] In the scheme, the control module is further configured to, in response to the jumping instruction, control the quadruped robot to jump to the target unit area based on the state parameters of the key points.

[0041] In the scheme, the control module is further configured to obtain at least two key points in the reference jumping trajectory.

[0042] In the scheme, the key points include at least one of the following:

[0043] a starting point of the reference jumping trajectory, a vertex of the reference jumping trajectory, and an ending point of the reference jumping trajectory.

[0044] The embodiment of the present application provides a quadruped robot, which comprises:

[0045] a robot body having at least two foot ends;

[0046] a controller arranged on the robot body and configured to execute the motion control method of the quadruped robot provided by the embodiment of the present application.

[0047] The embodiment of the present application provides a computer-readable storage medium storing executable instructions for causing a processor to execute the motion control method of the quadruped robot provided by the embodiment of the present application.

[0048] The embodiments of the present application have the following beneficial effects:

[0049] According to the embodiments of the present application, the foot-type robot can be controlled to stand in the independent unit areas respectively from the foot ends, jump to the target unit area, and the foot ends of the foot-type robot are collected in the target unit area. Thus, the jumping action to the target unit area can be realized under the limited condition of the landing target unit area, the dynamicity, difficulty and watchability of the foot-type robot movement are improved, and the diversity of the foot-type robot movement is expanded. BRIEF DESCRIPTION OF DRAWINGS

[0050] Figures 1A-1B is an architecture schematic diagram of a movement control system 100 of a foot-type robot provided by the embodiments of the present application;

[0051] Figure 1C is a structure schematic diagram of a four-foot robot provided by the embodiments of the present application;

[0052] Figure 1D is a structure schematic diagram of a control device provided by the embodiments of the present application;

[0053] Figure 2 is a structure schematic diagram of an electronic device 500 for implementing a movement control method of a foot-type robot provided by the embodiments of the present application;

[0054] Figures 3A-3B is a flow schematic diagram of a movement control method of a foot-type robot provided by the embodiments of the present application;

[0055] Figure 4 is a schematic diagram of a plum-blossom pile type terrain area provided by the embodiments of the present application;

[0056] Figure 5 is a schematic diagram of an initial state of a foot-type robot provided by the embodiments of the present application;

[0057] Figure 6 is a flow schematic diagram of state parameter acquisition provided by the embodiments of the present application;

[0058] Figure 7 is a schematic diagram of a take-off phase of a foot-type robot provided by the embodiments of the present application;

[0059] Figure 8 is a schematic diagram of a take-off phase of a foot-type robot provided by the embodiments of the present application;

[0060] Figure 9 is a schematic diagram of a landing phase of a foot-type robot provided by the embodiments of the present application;

[0061] Figure 10is a schematic diagram of an angle implementation of a legged robot based on degrees of freedom provided by embodiments of the present application;

[0062] Figure 11 is another flowchart of a motion control method of a legged robot provided by embodiments of the present application;

[0063] Figure 12 is another schematic diagram of a take-off phase of a legged robot provided by embodiments of the present application;

[0064] Figure 13 is a schematic diagram of operations performed in a take-off phase of a legged robot provided by embodiments of the present application;

[0065] Figure 14 is another schematic diagram of a landing phase of a legged robot provided by embodiments of the present application;

[0066] Figure 15 is a schematic diagram of a continuous jumping process of a legged robot provided by embodiments of the present application;

[0067] Figure 16 is another schematic diagram of a continuous jumping process of a legged robot provided by embodiments of the present application;

[0068] Figure 17 is a schematic diagram of a planar model of a quadruped robot provided by embodiments of the present application. DETAILED DESCRIPTION

[0069] In order to make the objects, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings, and the described embodiments should not be regarded as limiting the present application, and all other embodiments obtained by those of ordinary skill in the art without making creative efforts fall within the scope of protection of the present application.

[0070] In the following description, "some embodiments" are described, which describe a subset of all possible embodiments, but it can be understood that "some embodiments" can be the same subset or different subsets of all possible embodiments, and can be combined with each other without conflict.

[0071] If similar descriptions of "first / second" appear in the application file, the following description is added, in the following description, the terms "first\second\third" referred to only distinguish similar objects, and do not represent a specific order of the objects, and it can be understood that "first\second\third" can be interchanged in a specific order or sequence as allowed, so that the embodiments of the present application described herein can be implemented in an order other than that illustrated or described herein.

[0072] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the description herein is for describing particular embodiments only and is not intended to be limiting of the application.

[0073] Before the embodiments of the present application are further described, the terms and names involved in the embodiments of the present application are explained, and the terms and names involved in the embodiments of the present application are applicable to the following explanations.

[0074] 1) Robot: including various types of machines simulating human behavior or simulating other biological thoughts (such as robotic dogs, robotic cats, etc.). In a broad sense, some computer programs are also called robots. In the contemporary industry, robots refer to artificial robots that can automatically perform tasks to replace or assist human work, which can be electromechanical devices or controlled by computer programs or electronic circuits.

[0075] 2) Legged robot: refers to a robot with foot ends. The legged robot can be configured with one or more legs, each leg can be configured with one or more joints, and each leg corresponds to a foot end. The legged robot is, for example, a biped robot, a quadruped robot, or a hexapod robot. The quadruped robot is, for example, a robotic dog. Since the legged robot can include multiple foot ends, and the foot ends that fall at different times can be different, in order to facilitate the distinction, each foot end of the legged robot can be represented as a first foot end, a second foot end, an i-th foot end, etc. When a certain foot end falls, it also corresponds to the leg corresponding to the foot end.

[0076] 3) Foot point: refers to the position of the foot end of the robot in contact with the contact force, which is used to refer to the foot point of the robot, which can be a starting foot point or a candidate foot point. When the candidate foot point is selected as the foot point of the foot end, the candidate foot point can also be regarded as a target foot point. The starting foot point refers to the foot point of the legged robot at the starting time.

[0077] 4) Generalized coordinates: independent variables that can determine the position of a mass system (including at least one particle), such as angles, areas, etc. Polar coordinates, cylindrical coordinates, spherical coordinates, and curvilinear coordinates are all generalized coordinates. That is, generalized coordinates are nothing more than a set of coordinates. For a moving particle in a three-dimensional space, its position can be described by three rectangular coordinates or three spherical coordinates, which are all generalized coordinates. In analytical mechanics, it is customary to select a series of independent coordinates as generalized coordinates. For example, an independent coordinate: a free particle in a three-dimensional space must have three coordinates to determine its position, and the three coordinates are independent and indispensable.

[0078] In three-dimensional space, one particle has three generalized coordinates, that is, three independent variables can be used to find any position in three-dimensional space.

[0079] 5) Degree of freedom: independent coordinates that a single rigid body has, referred to as degree of freedom. In three-dimensional space, a single rigid body without constraints has six degrees of freedom, including three degrees of freedom of translation and three degrees of freedom of rotation. Among them, the three degrees of freedom of rotation include pitch, yaw and roll around the corresponding axis. Taking the right-hand Cartesian coordinate system in three-dimensional space as an example, rotation around the X-axis is called roll, and the angle of rotation around the X-axis, i.e. roll angle, can be represented by roll. The angle of rotation around the Y-axis can be represented by pitch, which is called pitch angle. Rotation around the Z-axis is also called yaw angle, which can be represented by yaw.

[0080] 6) Model predictive control (MPC): a control method for predicting the controlled object. Model predictive control is actually an optimization method to solve the control problem, or in other words, the solution to the optimization problem is used to give the action of the controller. The operation of model predictive control is described as follows: in the control process, there is always a desired reference trajectory. Taking time as the current time (the position of the longitudinal axis of the coordinate system), the controller combines the current measurement value and the prediction model to predict the output of the system in the future time domain. By solving the optimization problem that meets the objective function and various constraints, a series of control sequences in the control time domain are obtained, and the first element of the control sequence is taken as the actual control quantity of the controlled object. When it comes to the next time, repeat the above process, so as to complete one by one the optimization problem with constraints, so as to realize the continuous control of the controlled object. Model predictive control belongs to a kind of feedback control. For example, when crossing the road, you will establish a model of your speed and the speed of the car coming from the side at this moment in your mind, and then you will adjust your speed every step to determine that you will not be hit.

[0081] 7) Space landing constraint condition: used to constrain the foot end of the foot-type robot to be in the working space corresponding to the foot end after each step. The space landing constraint condition can be used to constrain the centroid position change coefficient, step sequence and landing point of the robot.

[0082] 8) Friction force constraint condition: used to constrain the contact force of the foot end to be within the friction cone to avoid slipping between the foot-type robot and the contact surface, wherein the friction cone is determined according to the normal vector of the candidate landing point of the foot-type robot and the friction coefficient between the landing foot end and the contact surface.

[0083] 9) foot end contact force constraint condition: used to constrain the contact force in the normal direction of the foot end contact force to be less than or equal to the upper limit of the contact force, so as to avoid excessive force between the foot robot and the contact surface

[0084] Based on the above explanations of the terms and nouns involved in the embodiments of the present application, the motion control system of the foot robot provided by the embodiments of the present application is described below. Referring to Figures 1A-1B , Figures 1A-1B is a schematic diagram of the architecture of the motion control system 100 of the foot robot provided by the embodiments of the present application. To realize an exemplary application, the system includes a foot robot 400 and a control device 200. The control device 200 and the foot robot 400 are two relatively independent devices. The foot robot 400 includes at least two foot ends (an exemplary foot robot including four foot ends is shown). In this case, the foot robot 400 is connected to the control device 200 through a network 300. The network 300 can be a wide area network or a local area network, or a combination of the two. The foot robot 400 and the control device 200 can be directly or indirectly connected through wired or wireless communication. The embodiments of the present application do not make any restrictions.

[0085] The foot robot 400 is configured to receive a jump instruction for the foot robot when each foot end of the foot robot stands in a single unit area independent of each other. The jump instruction is used to instruct the foot robot to jump from at least two unit areas currently located to a target unit area. In response to the jump instruction, the foot robot is controlled to jump to the target unit area, and in the target unit area, the distance between any two foot ends of the foot robot is less than the distance between the corresponding two foot ends before jumping.

[0086] The control device 200 is configured to receive a jump instruction from a host computer or obtain a jump instruction according to a user's input operation before the control device 200 controls the motion of the foot robot 400. The jump instruction can instruct the foot robot 400 to perform a corresponding jump action. The host computer can be any device wirelessly or wiredly connected to the control device 200, such as a terminal or a server, etc.

[0087] The control device 200 can be implemented by a terminal or a server. The terminal can be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart speaker, a smart watch, etc., but is not limited thereto. The server includes, but is not limited to, a stand-alone physical server, a server cluster composed of multiple physical servers, or a distributed system, and can also be a cloud server providing cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and basic cloud computing services such as big data and artificial intelligence platforms. The terminal and the server can be directly or indirectly connected through wired or wireless communication, and the present application is not limited in this embodiment.

[0088] In some embodiments, the control device 200 is part of the legged robot 400, and the control device 200 can be arranged in the body of the legged robot 400, for example, the control device 200 is an internal processor in the legged robot 400, etc.

[0089] Taking the legged robot as a quadruped robot as an example, referring to Figure 1C , Figure 1C is a structural schematic diagram of a quadruped robot provided by an embodiment of the present application. Number 1 shows a front leg, number 1-1 shows a front leg hip joint, number 1-2 shows a front thigh, number 1-3 shows a front leg knee joint, number 1-4 shows a front shank, and number 1-5 shows a front leg ankle joint. Number 2 shows a rear leg, number 2-1 shows a rear leg hip joint, number 2-2 shows a rear thigh, number 2-3 shows a rear leg knee joint, number 2-4 shows a rear shank, and number 2-5 shows a rear leg ankle joint. Number 3 shows a foot end, number 4 shows a torso, and number 5 shows a head. The angle between the thigh and the shank towards the head is the joint angle of the knee joint. In the absence of special instructions, each quantity is represented in the world coordinate system.

[0090] In order to more clearly introduce the structure of the control device 200, referring to Figure 1D , Figure 1DFig. 1 is a schematic diagram of a control device structure provided by an embodiment of the present application. The following will be described by way of example in combination with a motion control system of a legged robot shown in Fig. 1. In Fig. 1, the control device 200 includes a visual perception unit 210, a trajectory generation unit 220 and a motion control unit 230. The visual perception unit 210 can be arranged on the legged robot 400, for example, on the head of the legged robot 400. The visual perception unit 210 includes one or more of a camera, an infrared camera, for example. The camera is an RGBD camera, for example. The visual perception unit further includes a function for realizing simultaneous localization and mapping (SLAM). The visual perception unit 210 collects state data of the legged robot. The state data includes state data of the legged robot 400 at a starting time. In addition, the visual perception unit 210 can also collect an environmental image of the legged robot 400, and obtain a possible target cell region (landing point after a jumping operation) of the legged robot 400 each time it lands. After obtaining the state data and the environmental image, the visual perception unit 210 can send the state data and the environmental image to the trajectory generation unit 220. Alternatively, the trajectory generation unit 220 can obtain the state data of the legged robot 400 through internal sensors and external sensors of the legged robot 400. Alternatively, the trajectory generation unit 220 can take the expected state data at the end time of the last planning period as the state data at the starting time of the current corresponding planning period. Alternatively, the trajectory generation unit 220 obtains the state data of the legged robot 400 through a state estimator of the motion control unit 230. The trajectory generation unit 220 receives the state data and the target cell region (landing point after a jumping operation), determines the state of the legged robot 400 at multiple times (the state is determined based on state data such as joint angle, center of mass height, joint bending amplitude, joint lowering height, etc.), and obtains a center of mass motion trajectory of the legged robot 400 according to the center of mass position at the multiple times and the position of the target cell region, and further determines a jumping trajectory of the legged robot 400 according to the center of mass motion trajectory and the target cell region (landing point after a jumping operation or target landing point), and sends the jumping trajectory and the target landing point to the motion control unit 230. The motion control unit 230 can determine joint torques of each joint of the legged robot 230 according to the jumping trajectory and the target landing point, and control the rotation of each joint of the legged robot 400 according to the joint torques, thereby realizing the jumping motion of the legged robot 400.

[0091] Further, the motion control unit 230 can also monitor real-time state data (state data composed of values of various state parameters, which can also be referred to as control data) of the legged robot 400 during the jumping process, and control the jumping of the legged robot 400 according to the real-time state data, so as to ensure that the legged robot 400 can move stably.

[0092] The embodiment of the present application can also be implemented by means of cloud technology. The cloud technology refers to a kind of hosting technology that unifies series of resources such as hardware, software and network in wide area network or local area network to realize data calculation, storage, processing and sharing.

[0093] The cloud technology is a general term of network technology, information technology, integration technology, management platform technology and application technology applied based on cloud computing business model, can form resource pool, and is used on demand, flexible and convenient. Cloud computing technology will become an important support. The background service of technical network system needs a large amount of computing and storage resources.

[0094] Referring to Figure 2 , Figure 2 is a structural schematic diagram of an electronic device 500 for implementing the motion control method of the embodiment of the robot. In actual application, the electronic device 500 can be a server or a terminal shown in FIG. 1. Taking the electronic device 500 as the terminal shown in FIG. 1 as an example, the electronic device for implementing the marker processing method in the virtual scene is described. The electronic device 500 provided by the embodiment of the present application includes at least one processor 510, a memory 550, at least one network interface 520 and a user interface 530. Each component in the electronic device 500 is coupled together through a bus system 540. It can be understood that the bus system 540 is used to realize the connection and communication between the components. In addition to the data bus, the bus system 540 also includes power bus, control bus and status signal bus. However, in order to clearly illustrate, all kinds of buses are marked as bus system 540 in Figure 2 .

[0095] The processor 510 can be an integrated circuit chip with signal processing capability, such as a general purpose processor, a digital signal processor (DSP), or other programmable logic device, discrete gate or transistor logic device, discrete hardware component, etc. The general purpose processor can be a microprocessor or any conventional processor.

[0096] The user interface 530 includes one or more output devices 531 that enable presentation of media content, including one or more speakers and / or one or more visual display screens. The user interface 530 also includes one or more input devices 532, including user interface components that facilitate user input, such as a keyboard, mouse, microphone, touch screen display, camera, other input buttons and controls.

[0097] The memory 550 can be removable, non-removable, or a combination thereof. Exemplary hardware devices include solid-state memory, hard drives, optical drives, and the like. The memory 550 optionally includes one or more storage devices remotely located from the processor 510.

[0098] The memory 550 includes volatile memory or nonvolatile memory, and can also include both volatile and nonvolatile memory. Nonvolatile memory can be read only memory (ROM), volatile memory can be random access memory (RAM). The memory 550 described in the embodiments of the present application is intended to include any suitable type of memory.

[0099] In some embodiments, the memory 550 is capable of storing data to support various operations, examples of which include programs, modules, and data structures or subsets or supersets thereof, which are exemplarily illustrated below.

[0100] The operating system 551 includes system programs for processing various basic system services and performing hardware-related tasks, such as a framework layer, a core library layer, a driver layer, and the like, for implementing various basic services and processing hardware-based tasks;

[0101] The network communication module 552 is used to communicate with other computing devices via one or more (wired or wireless) network interfaces 520, exemplary network interfaces 520 including Bluetooth, wireless fidelity (WiFi), and universal serial bus (USB), and the like;

[0102] The presentation module 553 is used to enable the presentation of information via one or more output devices 531 associated with the user interface 530 (e.g., a display screen, a speaker, and the like) (e.g., a user interface for operating peripheral devices and displaying content and information);

[0103] The input processing module 554 is used to detect and interpret one or more user inputs or interactions from one or more input devices 532.

[0104] In some embodiments, the motion control device of the foot robot provided by the embodiments of the present application can be realized in a software manner, Figure 2 A marker processing device 555 in a virtual scene stored in the memory 550 is shown, which can be software in the form of programs and plug-ins, including the following software modules: a receiving module 5551 and a control module 5552, which are logical, and thus can be combined or further split according to the functions implemented, and the functions of each module will be described below.

[0105] In some embodiments, the motion control device of the biped robot can be implemented in a combination of software and hardware. For example, the motion control device of the biped robot can be a hardware decoding processor programmed to perform the marker processing method in the virtual scene provided by the embodiments of the present application. For example, the hardware decoding processor can be one or more application specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field programmable gate arrays (FPGAs), or other electronic elements.

[0106] The motion control method of the biped robot provided by the embodiments of the present application will be described in conjunction with the exemplary applications and implementations of the biped robot provided by the embodiments of the present application. Figures 3A-3B is a flowchart of the motion control method of the biped robot provided by the embodiments of the present application. Referring to Figure 3A , the motion control method of the biped robot provided by the embodiments of the present application will be described in conjunction with the steps shown in Figure 3A .

[0107] In step 101, when each foot end of the biped robot stands in a mutually independent cell region, the biped robot receives a jump instruction for the biped robot.

[0108] The jump instruction is used to instruct the biped robot to jump from at least two cell regions in which the biped robot is currently located to a target cell region.

[0109] The cell region will be described. In some embodiments, the cell region and the target cell region can be in the following form: the cell region and the target cell region are both the end face region of a single pile in a plum-blossom pile, and the shape and size of the cell region and the target cell region are the same.

[0110] In actual implementation, the unit area can be a series of discrete and independent landing points (unit areas) obtained by the control device of the legged robot through area segmentation and discretization of the terrain area. The unit area can be similar to the end face area of a single pile in a plum-pudding pile. The legged robot can control itself to complete a high-dynamic and high-difficulty jumping action in the plum-pudding pile or the terrain area that can be discretized into a plum-pudding pile form. The jumping action can include jumping of the legged robot from the end face area of each single pile in the plum-pudding pile to the end face area of the same single pile. Taking a quadruped robot as an example, the jumping action can be a four-pile jumping single-pile action of the quadruped robot starting from a four-foot dispersed standing state, and the four feet are gathered in a single-pile landing balance. The jumping action can also include jumping of the legged robot from the end face area of one single pile in the plum-pudding pile to the end face area of each single pile. Taking a quadruped robot as an example, the jumping action can be a single-pile jumping four-pile action of the quadruped robot starting from a four-foot gathering state. The target unit area is mainly used to represent a region in which the landing point of the legged robot is limited, that is, when the legged robot is in the target unit area (target landing point), it balances in a posture in which each foot end is gathered under the body. When each foot end is dispersed, the legged robot is easy to step on nothing or fall down.

[0111] Exemplarily, referring to Figure 4 , Figure 4 is a schematic diagram of a plum-pudding pile type terrain area provided by an embodiment of the present application. The plum-pudding pile shown in the diagram includes a plurality of independent single piles (unit areas) with the same area size. In actual application, different terrains can also be discretized into a plurality of unit areas with different heights and different area sizes.

[0112] In actual implementation, the initial state of the legged robot is that each foot end of the legged robot stands in a plurality of independent unit areas, exemplarily, referring to Figure 5 , Figure 5 is a schematic diagram of an initial state of a legged robot provided by an embodiment of the present application. Taking a quadruped robot as an example, the unit area is the end face area of a single pile in a plum-pudding pile, and the four feet of the quadruped robot stand on four single piles in the plum-pudding pile.

[0113] The way in which the foot-type robot receives the jump instruction is described. In actual implementation, the foot-type robot can receive a jump instruction for itself. The trigger mode of the jump instruction can be external triggering or internal triggering. The external triggering can be that the foot-type robot receives a jump instruction input by a user (the input form can be voice input or text input) or that the user triggers a function item for controlling the foot-type robot to jump. The internal triggering can be that the foot-type robot receives a jump instruction sent by a controller carried by the foot-type robot. In addition, the jump instruction can include at least one of a forward jump instruction and an upward jump instruction. Specifically, the jump instruction can be determined according to the position of a target unit area relative to a unit area. When the target unit area is located in front of the foot-type robot, the jump instruction is a forward jump instruction. When the target unit area is located at the center of each unit area, the jump instruction is a backward jump instruction. Subsequently, the jump instruction refers to the forward jump instruction unless otherwise specified. Different forms of jump instructions correspond to different jump processes.

[0114] In step 102, in response to the jump instruction, the foot-type robot is controlled to jump to the target unit area, so that each foot end of the foot-type robot is retracted into the target unit area, and in the target unit area, the distance between any two foot ends of the foot-type robot is less than the distance between the corresponding two foot ends before jumping.

[0115] In actual implementation, the foot-type robot performs a corresponding jump action in response to the jump instruction. Before performing the jump action, the foot-type robot can generate a reference jump trajectory corresponding to the current jump action by using a model predictive control (MPC) method. The reference jump trajectory can include state parameters of at least two key points. Each stage corresponds to corresponding control information. The foot-type robot analyzes the control information to obtain a motion trajectory corresponding to each stage.

[0116] In some embodiments, referring to Figure 6 , Figure 6 is a state parameter acquisition process schematic diagram provided by the embodiment of the present application, and the steps shown in Figure 6 illustrate the acquisition mode of the state parameters of each key point in the jump process performed based on the reference jump trajectory:

[0117] In step 201, the foot-type robot predicts a jump trajectory of the foot-type robot to obtain a reference jump trajectory.

[0118] In actual implementation, after receiving the jumping instruction, the legged robot can obtain a reference jumping trajectory corresponding to the jumping instruction by using a model predictive control (MPC) method. The reference jumping trajectory is a most energy-saving reference trajectory satisfying a constraint condition obtained after solving an optimization problem (a nonlinear optimization problem constructed to plan a jumping trajectory corresponding to the current jumping instruction to achieve an optimization goal of minimum driving force). The reference jumping trajectory can effectively ensure the feasibility of the legged robot performing each action in the jumping process, so that the reference jumping trajectory can be successfully predicted. In the jumping process, the legged robot controls its jumping action to be consistent with the reference jumping trajectory as much as possible.

[0119] In step 202, a constraint condition corresponding to the jumping instruction is obtained, and the constraint condition is used to constrain the jumping of the legged robot.

[0120] In actual implementation, different jumping instructions correspond to different reference jumping trajectories, and the constraint conditions are different when different reference jumping trajectories are obtained. The constraint condition is used to constrain the jumping of the legged robot. The constraint condition for the jumping instruction can be a spatial landing point constraint condition, a friction force constraint condition, a foot end constraint force constraint condition, etc. For example, for the jumping action of the legged robot on the plum-blossom stake in the embodiment of the present application, the constraint conditions include joint torque limiting, foot bottom friction cone, joint drop height threshold, dynamics constraint, and foot end non-slip, etc.

[0121] In step 203, based on the reference jumping trajectory and the constraint condition, the state parameters of at least two key points of the reference jumping trajectory are determined.

[0122] In some embodiments, when determining the state parameters of the at least two key points, the legged robot first obtains at least two key points in the reference jumping trajectory, wherein the key points include at least one of the following: a starting point of the reference jumping trajectory, a vertex of the reference jumping trajectory, and an ending point of the reference jumping trajectory.

[0123] In actual implementation, the key points can be sampling points at each operation segment in the jumping process. The parameter values of the state parameters corresponding to the legged robot at the sampling points can be represented in the form of generalized coordinates. In order to facilitate description, the parameter values of the state parameters corresponding to each key point are referred to as control information, that is, one key point corresponds to one piece of control information. The corresponding operation of the legged robot controlled according to the obtained control information of each key point can be as close as possible to the reference jumping trajectory. The state parameters can include the center of mass position of the legged robot, the positions of each foot end, the joint angles of each joint, the joint angular velocity, the joint angular acceleration, the joint torque, the joint degrees of freedom, etc. The control data represented by the generalized coordinates will be described in detail below taking a quadruped robot as an example.

[0124] In some embodiments, the foot-type robot can jump to the target cell region located in front of the foot-type robot in the following stages: the foot-type robot controls the foot-type robot to jump forward to the target cell region in a take-off stage, a flight stage, and a landing stage in response to the jump instruction.

[0125] In actual implementation, the entire process of jumping of each foot end of the foot-type robot from the corresponding cell region to the target cell region can be divided into a take-off stage, a flight stage, and a landing stage. The foot-type robot obtains control information of each key point in the reference jump trajectory corresponding to the current jump instruction, and the foot-type robot analyzes the control information to obtain parameter values of state parameters corresponding to each operation, thereby presenting a corresponding form.

[0126] In some embodiments, the foot-type robot can perform the following operations in the take-off stage of the jump process: in the take-off stage, the foot-type robot sequentially performs squatting to accumulate power, foot end kicking the ground, and starting flight.

[0127] In actual implementation, in the take-off stage, the foot-type robot can at least obtain control information of the following key points in the reference jump trajectory: squatting to accumulate power, foot end kicking the ground, and starting flight. In the take-off stage of the entire jump process, the foot-type robot starts from an initial state of normally standing on mutually independent cell regions, sequentially performs squatting to accumulate power, foot end kicking the ground, and starting flight, and the foot-type robot controls itself to be in a state corresponding to each operation according to the corresponding control information.

[0128] Exemplarily, referring to Figure 7 , Figure 7 is a schematic diagram of the foot-type robot in the take-off stage provided by the embodiments of the present application, the foot-type robot in the diagram is a quadruped robot, and the cell region is a single stake in a plum-blossom stake. Number 1 in the diagram shows an initial state of the quadruped robot, number 2 in the diagram shows squatting to accumulate power of the foot-type robot in the take-off stage of the jump process, number 3 in the diagram shows foot end kicking the ground of the foot-type robot in the take-off stage of the jump process, and number 4 in the diagram shows starting flight of the foot-type robot in the take-off stage of the jump process.

[0129] In some embodiments, in the take-off stage, each body part of the foot-type robot can present the following form changes: in the process of squatting to accumulate power of the foot-type robot, the leg of the foot-type robot is bent to lower the center of mass of the foot-type robot from an initial height to a take-off height; in the process of foot end kicking the ground of the foot-type robot, the foot-type robot increases the torque of the leg joint to increase the friction force of each foot end and the foot end contact surface; and in the process of starting flight of the foot-type robot, the foot-type robot performs a take-off action forward and upward at a target pitch angle to increase the height of the knee joint and the height of the center of mass, and the height of the center of mass is higher than the initial height.

[0130] In actual implementation, the foot-type robot obtains first control information after the squatting power accumulation operation is completed, and based on the control information, the joint angles of the joints of the foot-type robot are bent from the joint angles at the initial state to the joint angles specified in the control information, and the center of mass is lowered from the initial height to the take-off height, where the take-off height is the center of mass height specified in the control information, and the center of mass height specified in the first control information is determined based on the constraint conditions that the joints cannot touch the contact surface of the unit area and the shutdown torque is limited. After the squatting power accumulation operation is completed, the foot-type robot starts the foot end ground-pushing operation, and in the process of ground-pushing, the corresponding second control information is analyzed, the joint torque of each joint is increased to the value of the corresponding joint torque in the second control information, the friction between the foot end and the single pile end surface is increased, at the same time, the head of the foot-type robot is raised (the pitch angle of the foot end robot is adjusted to increase to the pitch angle in the second control information), the height of the hip joint is lowered, the angle between the body trunk and the thigh joint is increased, such as 180 degrees, the angle of the front leg joint is increased, and the center of mass is migrated. The second control information here is determined at least based on the constraint condition that the foot end does not slide in the process of ground-pushing (friction constraint condition); when the foot end ground-pushing operation is completed, the start floating operation is performed, at this time the foot end robot obtains the third control information corresponding to this time point, adjusts the position of the center of mass to the center of mass position in the third control information, at this time the height of the center of mass is higher than the initial height, the pitch angle is reduced relative to the foot end ground-pushing to the pitch angle in the third control information, the joint angle of each joint is increased to the joint angle in the third control information, and the change speed of the joint angle is consistent with the joint angle speed of the third control information.

[0131] Taking the foot-type robot as a four-legged robot as an example, referring to the squatting power accumulation operation of the foot-type robot shown in FIG. 2 in the foregoing embodiment, compared with the initial state shown in FIG. 1, the joint angles of the joints of the foot-type robot are bent, and the center of mass is lowered from the initial height to the take-off height. Referring to the foot end ground-pushing operation of the foot-type robot shown in FIG. 3, compared with the state shown in FIG. 2 in the squatting power accumulation, the head of the foot-type robot is raised, the height of the hip joint of the rear leg is lowered, and the body trunk and the rear thigh are in a straight line. Referring to the start floating operation of the foot-type robot shown in FIG. 4, compared with the state shown in FIG. 3 in the foot end ground-pushing, the joint height of each joint is increased, the height of the center of mass is higher than the initial height, the center of mass moves upward and forward, the joint angle of each joint is increased, and the trunk is parallel to the pile surface. Figure 7 Figure 7 Figure 7 Figure 7

[0132] ​​​​In some embodiments, the foot-type robot can perform the following operations in the flight phase of the jumping process: in the flight phase, the foot-type robot sequentially performs the operations of in-air foot retraction and leg retraction and in-air leg extension.

[0133] In actual implementation, in the flight phase of the entire jumping process, the foot-type robot sequentially performs the operations of in-air foot retraction and leg retraction and in-air leg extension from the posture corresponding to the start of the flight operation in the take-off phase, and dynamically changes the state change (also referred to as posture change) of the foot-type robot corresponding to each operation according to the control information corresponding to each operation.

[0134] Exemplarily, referring to Figure 8 , Figure 8 is a schematic diagram of the flight phase of the foot-type robot provided by the embodiments of the present application, the foot-type robot in the diagram is a quadruped robot, and the unit area is a single pile in a plum-pudding pile. The number 1 in the diagram shows the in-air foot retraction and leg retraction operation of the foot-type robot in the flight phase of the jumping process. The number 2 in the diagram shows the foot end ground kicking operation of the foot-type robot in the flight phase of the jumping process. The number 4 in the diagram shows the in-air leg extension operation of the foot-type robot in the take-off phase of the jumping process.

[0135] In some embodiments, in the flight phase, each body part of the foot-type robot can present the following morphological changes: in the process of in-air foot retraction and leg retraction of the foot-type robot, the foot-type robot controls each leg to drive the corresponding foot end to be retracted upwards to below the body center; in the process of in-air leg extension of the foot-type robot, the foot-type robot controls the front leg to stretch towards the target unit area, so that each foot end of the foot-type robot can land on the target unit area.

[0136] In actual implementation, the foot-type robot obtains fourth control information corresponding to in-air foot retraction and leg retraction, and based on the fourth control information, controls the foot-type robot to gradually move each foot end towards below the body center in the air, that is, controls the joint angle of each joint of the foot-type robot to increase from the joint angle at the time when the start of the flight operation is completed to the joint angle specified in the fourth control information according to the joint angular velocity and the joint acceleration in the fourth control information, and increases the height of the center of mass and the component of the center of mass along the X axis (that is, the center of mass moves forward); when the height of the center of mass starts to decrease, fifth control information for in-air foot retraction and leg extension is obtained, based on the fifth control information, the component of the center of mass along the X axis is continuously controlled to increase, the height of the center of mass is decreased, the body is controlled to have a target pitch angle (the pitch angle in the pitch angle), and the joint angle of each joint of the front leg is continuously reduced to the joint angle in the fourth control information to prepare for landing, and at the same time, since the number of target unit areas is one, the distance of the rear leg stretching forward is far, at this time, according to the joint angle of the knee joint and the joint angle of the hip joint of the rear leg in the fifth control information, the rear leg is controlled to stretch towards the target unit area.

[0137] Taking the four-legged robot as an example, see Figure 8 The air foot-receiving and leg-receiving operation of the foot-type robot shown in FIG. 1, see Figure 7 The state of the start of the take-off operation shown in FIG. 4, the foot-type robot is controlled to gradually move each foot end toward below the body mass center in the air, and the height of the mass center, the component of the mass center along the X axis is increased (that is, the mass center moves forward). The air leg-out operation of the foot-type robot shown in FIG. 2, see Figure 8 The state of the air foot-receiving and leg-receiving shown in FIG. 1, continue to control the X axis component of the mass center to increase, control the mass center to move forward and downward, reduce the height of the mass center, and control the head to have a target pitch angle.

[0138] In some embodiments, the foot-type robot can perform the following operations in the landing phase of the jumping process: in the landing phase, the foot-type robot sequentially performs the foot end landing, the leg bending and buffering, and the body balancing operations.

[0139] In actual implementation, in the landing phase of the entire jumping process, the foot-type robot starts to sequentially perform the foot end landing, the leg bending and buffering, and the body balancing operations from the posture of the stretching operation to the target unit area in the take-off phase, and dynamically changes the state changes (also referred to as posture changes) of the foot-type robot corresponding to each operation according to the control information corresponding to each operation.

[0140] Exemplarily, see Figure 9 , Figure 9 is a schematic diagram of the landing phase of the foot-type robot provided by the embodiments of the present application, the foot-type robot in the figure is a four-legged robot, and the unit area is a single pile in the plum-blossom pile. FIG. 1 shows the foot end landing operation of the foot-type robot in the landing phase of the jumping process, FIG. 2 shows the leg bending and buffering operation of the foot-type robot in the landing phase of the jumping process, and FIG. 4 shows the body balancing operation of the foot-type robot in the landing phase of the jumping process.

[0141] In some embodiments, in the landing phase, each body part of the foot-type robot can present the following morphological changes: in the process of the foot end landing of the foot-type robot, the foot-type robot controls each foot end to simultaneously contact the contact surface in the target unit area; in the process of the leg bending and buffering of the foot-type robot, the foot-type robot performs the leg bending action of bending and contracting each leg, and increases the torque of the leg joint to reduce the speed of the foot-type robot; in the process of the body balancing of the foot-type robot, the foot-type robot adjusts the body posture to move the mass center of the foot-type robot to the target unit area.

[0142] In actual implementation, the legged robot obtains the sixth control information when the foot ends touch the ground. Each foot end of the legged robot simultaneously contacts the contact surface of the target unit area. At this time, the knee joints of the legged robot's hind legs are still at a large joint angle. This is because the distance the hind legs extend toward the target unit area is longer than the distance the front legs extend toward the target unit area. To ensure simultaneous contact of the foot ends, the joint angles are larger than those during the leg extension operation in the air. After the foot ends simultaneously contact the contact surface of the target unit area, the legged robot begins to control each leg to perform a flexion and contraction flexion movement according to the seventh control information corresponding to the flexion leg buffer. At the same time, torque control is applied to each joint to buffer and decelerate the legged robot, smoothly moving the body's center of mass projection into the target unit area, thereby ultimately achieving balance. Because the hind legs extend forward a long distance when contacting the pile surface (the backward knee and hip joints extend a large distance), the hind leg knee joints tend to move downward during the buffering process after landing. At the same time, in order to avoid the hind leg calf colliding with the contact surface of the target unit area after the knee joint is lower than the height of the target unit area, the seventh control information is determined based on the friction constraint conditions, the spatial landing constraint conditions, and the foot end contact force constraint conditions. According to the joint angles and center of mass positions of each joint in the seventh control information, the contraction amplitude of the hind leg bending during the leg bending buffering of the foot-type robot is controlled to not exceed the contraction amplitude threshold, and the descent amplitude of the hind leg knee joint does not exceed the descent amplitude threshold. In addition, it can also ensure that the foot end does not rebound due to impact after contacting the pile surface, nor does it slide due to the leg posture and force (the seventh control information is determined based on the friction constraint conditions).

[0143] Continuing from the above example, let's take the legged robot as a quadruped robot, see Figure 9 The foot-end landing operation of the foot-type robot shown in the number 1 is compared with Figure 8 The state of the mid-air leg extension operation shown in the figure, numbered 2, controls the foot ends of the leg robot to contact the pile surface at the same time, the center of mass descends, the torso is parallel to the pile surface, the joints are bent, and the front leg hip joint and the torso are approximately in the same straight line. Compared with the state of the foot end landing shown in number 1, the leg bending and buffering operation of the leg robot shown in the figure, numbered 2, continues to control the knee joint to bend, the center of mass to move backward, the rear thigh to be perpendicular to the torso and rear calf, and the rear calf to be parallel to the pile surface. Compared with the state of the bent leg buffering shown in number 2, the body balancing operation of the leg robot shown in the figure, numbered 3, smoothly moves the projection of the body center of mass onto the single pile.

[0144] In some embodiments, when the target unit area is located at the center of each unit area where the foot end is located, the legged robot can achieve upward jumping in the following way: the legged robot responds to the jumping instruction, controls the legged robot to jump upward to the target unit area through the take-off stage, the flight stage and the landing stage.

[0145] In actual implementation, the legged robot obtains a reference jumping trajectory for upward jumping and control information (including reference values of various state parameters) of each key point on the reference jumping trajectory according to the MPC prediction, so as to realize upward jumping of the legged robot to the target unit region. The upward jumping process can be divided into a take-off phase, a flight phase and a landing phase.

[0146] In some embodiments, the legged robot can sequentially perform the following operations in each phase of the upward jumping process: in the take-off phase, the legged robot sequentially performs the operations of squatting to accumulate power and starting flight; in the flight phase, the legged robot performs the operation of folding legs in the air; and in the landing phase, the legged robot sequentially performs the operations of landing of foot ends, bending of legs and balance of the body. In the process of squatting to accumulate power of the legged robot, the legged robot bends the legs to lower the center of mass from an initial height to a take-off height; in the process of starting flight of the legged robot, the legged robot performs upward bouncing to increase the height of the knee joint and the height of the center of mass, and the height of the center of mass is higher than the initial height; in the process of folding legs in the air of the legged robot, the legged robot controls the trunk to be in an upright state, and each leg of the legged robot drives the corresponding foot end to be folded upward below the trunk.

[0147] In some embodiments, when the legged robot is a quadruped robot, the legged robot can also implement the jumping operation in the following manner: the legged robot responds to the jumping instruction, and when the jumping instruction indicates two-foot jumping, controls the legged robot to alternately use two front legs and two rear legs to jump to the target unit region.

[0148] In actual implementation, the jumping gait of the quadruped robot for performing the jumping operation can include two-foot jumping and four-foot jumping, wherein the two-foot jumping is used to indicate a jumping manner in which the legged robot alternately uses two front legs and two rear legs to jump.

[0149] In some embodiments, the legged robot can implement the jumping process of two-foot jumping through the following phases: the process in which the legged robot alternately uses two front legs and two rear legs to jump includes a power accumulation phase, a four-foot ground-pushing phase, a front-leg-off-ground-rear-leg-ground-pushing phase, a flight phase and a landing phase.

[0150] In some embodiments, referring to Figure 3B After each foot end of the legged robot stands in the target unit region and keeps balance, the legged robot can further perform step 103 of controlling the posture of the legged robot:

[0151] In step 103, the legged robot continuously performs at least one of the following operations in the target region based on at least one degree of freedom: a head shaking action and a tail swinging action.

[0152] In actual implementation, when the legged robot is in a target area and receives an instruction of a target action, the robot is controlled to complete the target action based on at least one degree of freedom, wherein the target action includes at least one of a head shaking action and a tail swinging action. In actual application, the legged robot can further include a tail connected to a body part of the legged robot through at least one joint, each joint can correspond to at least one degree of freedom, and the degree of freedom includes at least one of roll, pitch and yaw. The legged robot can realize the head shaking action, the tail swinging action or a coherent action of head shaking and tail swinging based on the degree of freedom. The degree of freedom of the legged robot is relative to a spatial coordinate system, which can be a world coordinate system or a body coordinate system with a body center of mass of the legged robot as an origin. In actual application, in the world coordinate system, a right-hand rule coordinate system is adopted, a jumping direction of the legged robot (a front direction of the body of the legged robot) is taken as an X axis, a direction perpendicular to the X axis on the left side of the legged robot is taken as a Y axis, and a direction perpendicular to the ground is taken as a Z axis. When the legged robot performs the head shaking and tail swinging action, the legged robot can be controlled to rotate around the X axis (X roll) to obtain a corresponding roll angle, to rotate around the Y axis (Y pitch) to obtain a corresponding pitch angle, and to rotate around the Z axis (Z yaw) to obtain a corresponding yaw angle, and then the head shaking action, the tail swinging action or the head shaking and tail swinging action is realized based on at least one of the pitch angle, the yaw angle and the roll angle. In terms of technical implementation, the action reference trajectory of the legged robot when performing the head shaking action (or the tail swinging action) can be planned according to an MPC controller, and the optimal target state data of the legged robot when performing the head shaking action (or the tail swinging action) is predicted based on the MPC and in combination with state data of the legged robot in a current state, the state data including at least one of a center of mass position of the body center of mass, a joint torque of each joint and a joint angle of each joint. The legged robot receives the target action instruction and the target state data matched with the target action, and completes the corresponding target action (at least one of the head shaking action and the tail swinging action).

[0153] Exemplarily, referring to Figure 10 , Figure 10 is an angle implementation schematic diagram of a legged robot based on a degree of freedom provided in the embodiments of the present application. In the diagram, a quadruped robot is taken as an example, the quadruped robot is retracted on a single plum blossom pile (at this time, in a balanced state), is rotated along an X axis of a coordinate system shown in the diagram to realize pile roll angle control of the quadruped robot, is rotated along a Y axis of the coordinate system shown in the diagram to realize pile pitch angle control of the quadruped robot, and is rotated along a Z axis of the coordinate system shown in the diagram to realize pile yaw angle control of the quadruped robot.

[0154] The above angle control method for the legged robot based on the degree of freedom can expand the motion ability of the robot and enrich the diversity of the action of the legged robot.

[0155] By applying the embodiment of the application, the jumping action completed by the legged robot improves the dynamicity, difficulty and ornamental of the legged robot movement, expands the diversity of the legged robot movement, and makes the legged robot not only be able to complete high-difficulty action display on the flat ground, but also be able to complete high-difficulty action display in scenes such as plum-blossom stumps and small stone piers.

[0156] Next, the movement control method of the legged robot provided by the embodiment of the application is continuously described. At this time, the initial state of the legged robot is that each foot end stands in the same initial unit area in a folded posture, and after the jumping action is completed, each foot end stands in a different target unit area. Referring to Figure 11 , Figure 11 is another flowchart of the movement control method of the legged robot provided by the embodiment of the application, which will be described in combination with the steps shown in Figure 11 .

[0157] In step 401, when each foot end of the legged robot stands in the same initial unit area and each foot end is folded in the initial unit area, the legged robot receives a jumping instruction for the legged robot.

[0158] The jumping instruction is used to instruct the legged robot to jump from the current initial unit area to a first target unit area (distinguished from the foregoing) which is independent of each other, and the foot end has a one-to-one corresponding relationship with the target unit area; the distance between any two foot ends of the legged robot in the target unit area is greater than the distance between the corresponding two foot ends of the legged robot in the initial unit area. The target unit area can be located in front of the legged robot, or can be scattered around the initial unit area.

[0159] In actual implementation, the terrain area where the legged robot is located includes at least two cell areas (which is consistent with the cell area in step 101). The initial state of the legged robot is that each foot end of the legged robot stands in the same initial cell area, and each foot end is retracted in the initial cell area. It can be understood that the initial cell area can be a restricted terrain area, and when each foot end of the legged robot stands in the area, it is usually in a posture where each foot end is retracted downward to the center of the body; the distance between any two foot ends of the legged robot when the legged robot is in the target cell area is greater than the distance between the corresponding two foot ends when the legged robot is in the initial cell area. According to the received jump instruction (the way in which the legged robot receives the jump instruction is as described above), the legged robot can control itself to complete a high-dynamic and high-difficulty first jump action (distinguished from the jump action described above) in the Chinese plum-blossom stake or the terrain area capable of being discretized into the form of the Chinese plum-blossom stake. Taking a quadruped robot as an example, the jump action can be a single-stake-to-four-stake action of the quadruped robot starting from a four-foot retracted state.

[0160] In step 402, in response to the jump instruction, the legged robot is controlled to jump to each target cell area.

[0161] In actual implementation, the legged robot performs a corresponding jump action in response to the jump instruction. Before performing the jump action, the legged robot can generate a reference jump trajectory corresponding to the current jump action by using a model predictive control (MPC) method. The reference jump trajectory can include state parameters of at least two key points, and each stage corresponds to corresponding control information. The legged robot analyzes the control information to obtain a motion trajectory corresponding to each stage.

[0162] In some embodiments, the legged robot can jump to each target cell area located in front of the legged robot in the following stages: when each first target cell area is located in front of the legged robot, the legged robot is controlled to jump from the initial cell area to each target cell area through a take-off stage, a flight stage, and a landing stage in response to the jump instruction.

[0163] In actual implementation, the process in which the legged robot jumps from the initial cell area to the target cell area independent of each other can at least include at least one of a take-off stage, a flight stage, and a landing stage. The legged robot obtains control information of each key point in the reference jump trajectory corresponding to the current jump instruction, and the legged robot analyzes the control information to obtain parameter values of state parameters corresponding to each operation, so as to present a corresponding form.

[0164] The operations performed by the legged robot in the take-off phase of the jumping process are described in sequence. In some embodiments, the legged robot can perform the following operations in the take-off phase of the jumping process: in the take-off phase, the legged robot performs the operations of foot end ground pushing, front foot end take-off in sequence.

[0165] In actual implementation, in the take-off phase, the legged robot can obtain control information of at least the following key points in the reference jumping trajectory: foot end ground pushing, front foot end take-off (at this time, the rear foot end continues to push the ground). In the take-off phase of the entire jumping process, the legged robot starts from the initial state and performs the operations of foot end ground pushing (each foot end performs the ground pushing action in the initial cell area) and front foot end take-off (after the front foot takes off, the rear foot continues to push the ground) in sequence. For each operation, the legged robot controls itself to be in a state corresponding to the operation according to the corresponding control information.

[0166] For example, referring to Figure 12 , Figure 12 is another schematic diagram of the take-off phase of the legged robot provided by the embodiments of the present application. In the diagram, the legged robot is a quadruped robot, and the cell area is a single stake in the plum blossom stake. In the diagram, number 1 shows the initial state of the quadruped robot, number 2 shows the foot end ground pushing operation of the legged robot in the take-off phase of the jumping process, and number 3 shows the front foot end take-off operation of the legged robot in the take-off phase of the jumping process.

[0167] In some embodiments, in the take-off phase, each body part of the legged robot can present the following morphological changes: in the process of foot end ground pushing by the legged robot, the legged robot reduces the joint angle of the rear leg knee joint and increases the joint angle of the front leg knee joint, so that the rear upper leg is in a straight line with the torso, the torso is raised at a target pitch angle, and the center of mass is shifted forward and upward.

[0168] In actual implementation, the legged robot obtains the first control information corresponding to the foot end ground pushing key point in the reference jumping trajectory, analyzes the first control information, adjusts the values of the state parameters to the reference values of the state parameters in the first control information, and makes the legged robot accelerate forward and upward at a target body pitch angle and center of mass position by simultaneously pushing the ground according to the reference jumping trajectory.

[0169] Taking the legged robot as a quadruped robot as an example, referring to the foot end ground pushing operation of the legged robot shown in number 2 in Figure 12 , compared with the initial state shown in number 1 in Figure 12 , the knee joint is raised in height during foot end ground pushing, the torso of the quadruped robot is in a straight line with the rear upper leg, the joint angle of the front leg knee joint is close to 180°, and the torso is raised at a target pitch angle.

[0170] In some embodiments, during the process of the legged robot lifting its front foot off the ground (the front foot lifts off the ground and the hind foot continues to push off the ground), the legged robot reduces the joint angle of the front leg knee joint to a first target angle, and increases the joint angle of the hind leg knee joint to a second target angle, so as to increase the height of each knee joint and the height of the center of mass; wherein the height of the front leg knee joint is higher than the height of the center of mass, and the height of the hind leg knee joint is lower than the height of the center of mass; the first target angle is used to represent the angle of the front leg knee joint when the front thigh and torso of the legged robot are in the same straight line, and the second target angle is used to represent the angle of the hind leg knee joint when the hind calf of the legged robot is perpendicular to the contact surface and the hind thigh and torso are in the same straight line.

[0171] In actual implementation, after the legged robot performs the foot-end pushing off the ground operation according to the reference jumping trajectory, it then performs the front foot-end lifting off the ground operation. At this time, each rear foot continues to push off the ground, and the legged robot obtains the second control information corresponding to the key point of the front foot lifting off the ground in the reference jumping trajectory, and parses the second control information, adjusts the values ​​of each state parameter of the legged robot to the reference values ​​of each state parameter in the second control information, and the legged robot continues to accelerate forward and upward according to the corresponding reference angle and position to take off, while the front legs bend and contract.

[0172] Continuing from the above example, let's take the legged robot as a quadruped robot, see Figure 12 The front foot of the foot-type robot shown in number 3 is off the ground, compared with Figure 12 In the state shown by number 2 when the foot pushes off the ground, the center of mass moves forward and upward, the front leg bends and contracts, the rear calf is perpendicular to the pile surface, the rear thigh and the torso are in the same straight line, and the torso is raised at the target pitch angle.

[0173] The operations performed sequentially by the legged robot during the take-off phase of the jumping process are described. In some embodiments, during the take-off phase, the legged robot performs an operation of lifting the rear end of the foot off the ground. During the process of lifting the rear end of the foot off the ground, the legged robot controls each front leg to extend toward the corresponding target unit area, and controls each hind leg to first contract and then extend toward the corresponding target unit area, so that each foot end of the legged robot can land on the corresponding target unit area.

[0174] In actual implementation, after the foot robot performs the forefoot end take-off operation according to the reference jump trajectory, the foot robot then performs the hindfoot end take-off operation, at which time the foot robot starts to take off, the foot robot obtains third control information corresponding to the hindfoot end take-off key point in the reference jump trajectory, and analyzes the third control information, adjusts the values of the state parameters of the foot robot to the reference values of the state parameters in the third control information, the foot robot completely takes off, the foreleg starts to stretch to the position of the corresponding target unit area of the corresponding foot end, and the hindleg first contracts to provide a swing foot landing height and then starts to stretch to the position of the corresponding target unit area of the corresponding foot end, and prepares for landing.

[0175] Taking the foot robot as a four-legged robot, referring to Figure 13 , Figure 13 is an operation schematic diagram provided by the embodiment of the application and executed by the foot robot in the take-off phase. In the take-off phase, the foot robot performs the operation of taking off the hindfoot end. Compared with the state shown in No. 3 in Figure 12 , the foot robot controls the hindfoot end to stand, and the foot robot completely takes off, wherein the foreleg starts to stretch to the position of the corresponding target single stake (the single stake shown in No. 1 in the figure), and the hindleg first contracts to provide a swing foot landing height and then stretches to the position of the corresponding target single stake (the single stake shown in No. 2 in the figure).

[0176] The operations performed by the foot robot in the landing phase of the jump process are described. In some embodiments, in the landing phase, the foot robot performs the operations of foot end landing, leg bending and body balancing in sequence.

[0177] In actual implementation, in the landing phase, the foot robot can at least obtain the control information of the following key points in the reference jump trajectory: foot end landing, leg bending and body balancing. In the landing phase of the entire jump process, the foot robot performs the operations of foot end landing, leg bending and body balancing in sequence, and the foot robot controls itself to be in a state corresponding to the operation according to the corresponding control information.

[0178] Exemplarily, referring to Figure 14 , Figure 14 is another schematic diagram provided by the embodiment of the application and executed by the foot robot in the landing phase. The foot robot in the figure is a four-legged robot, and the unit area is a single stake in the plum blossom stake. No. 1 in the figure shows that the foot robot performs the foot end landing operation, No. 2 in the figure shows that the foot robot performs the leg bending operation, and No. 3 in the figure shows the body balancing operation of the foot robot.

[0179] In some embodiments, during the landing phase, the various body parts of the legged robot may exhibit the following morphological changes: during the process of the foot-end of the legged robot landing, the legged robot controls each foot-end to contact the contact surface within the corresponding target unit area at the same time; during the process of the legged robot bending leg to cushion, the legged robot performs the bending leg action of bending and contracting each leg, and increases the torque of the leg joints to reduce the speed of the legged robot; during the process of the legged robot balancing the body, the legged robot adjusts its body posture to move the center of mass of the legged robot to the area composed of each target unit area.

[0180] In actual implementation, the legged robot obtains fourth control information corresponding to the key point of landing of the rear foot in the reference jump trajectory, analyzes the fourth control information, and adjusts the values ​​of various state parameters of the legged robot to the reference values ​​of each state parameter in the fourth control information. Each foot of the legged robot simultaneously contacts the contact surface of the corresponding target unit area. At this time, the knee joints of the front legs of the legged robot are still at a large joint angle. The body of the legged robot is raised at the target pitch angle. The knee joints of the rear legs move downward until the rear shanks are nearly parallel to the contact surface. The center of mass shifts toward the rear legs, preparing for subsequent leg bending and cushioning. The legged robot obtains fifth control information for the key point of leg bending and cushioning and executes the leg bending and cushioning operation. The legged robot controls the bending and contraction of each leg while applying torque control to each joint, smoothly moving the projection of the body's center of mass into the four-leg support area. At this time, the torso is raised at the target pitch angle, the rear thighs extend backward to align with the torso, and the knee joints descend until they are aligned with the contact surface of the corresponding target area, achieving balance within each target unit area.

[0181] Continuing from the above example, let's take the legged robot as a quadruped robot, see Figure 14 The foot-end landing operation of the foot-type robot shown in the number 1 is compared with Figure 13 In the state of a mid-legged robot in the air, the four foot ends of the quadruped robot contact the pile surface at the same time, the knee joints of the hind legs move downward until the hind legs are nearly parallel to the contact surface, the center of mass migrates toward the hind legs, preparing for subsequent bent-leg buffering, and the trunk is raised at the target pitch angle. In the bent-leg buffering operation of the leg-type robot shown in the figure with number 2, compared with the state shown in the figure with number 1, the knee joints of the hind legs of the quadruped robot continue to move downward until the hind thigh and the trunk are in the same straight line, the hind shank is nearly parallel to the contact surface, and the trunk is raised at the target pitch angle. In the foot-end landing operation of the leg-type robot shown in the figure with number 3, compared with the state shown in the figure with number 2, the trunk of the quadruped robot is parallel to the pile surface, the center of mass migrates to the center of the trunk, and the knee joints of the hind legs are raised to the same height as the knee joints of the front legs.

[0182] In actual implementation, the foot-type robot can perform another continuous jumping action, i.e., when the foot-type robot stands on mutually independent unit areas from each foot end, the continuous stable jumping action of the foot-type robot can be realized by combining steps 101-102 and steps 401-402. That is, from the initial state that the foot-type robot stands on mutually independent unit areas from each foot end, the foot-type robot jumps to the intermediate unit area (each foot end of the foot-type robot is collected in the intermediate unit area) through the above steps 101-102, and then jumps from the single intermediate unit area to the plurality of intermediate unit areas with the same number of foot ends through the above steps 401-402.

[0183] Exemplarily, referring to Figure 15 , Figure 15 is a schematic diagram of a continuous jumping process of a foot-type robot provided by the embodiment of the present application, and the continuous jumping process includes first jumping and second jumping, wherein the first jumping is that the four-foot robot jumps from the four single piles of the plum-blossom pile to one single pile in front, and the second jumping is that the four-foot robot jumps from one single pile to another four single piles on the basis of the first jumping. The jumping process of the first jumping is consistent with the jumping process shown in steps 101-102, and the jumping process of the second jumping is consistent with the jumping process shown in steps 401-402. It should be noted that the two jumping processes can be continuously and alternately performed, and the embodiment of the present application does not make any limitation.

[0184] In actual implementation, the foot-type robot can also continuously perform the jumping action, i.e., when the foot-type robot jumps from the initial unit area to the target unit area, the continuous stable jumping action of the foot-type robot can be realized by combining steps 401-402 and steps 101-102. That is, from the initial state that the foot-type robot stands on the same initial unit area from each foot end and each foot end is collected in the initial unit area, the foot-type robot jumps to the plurality of intermediate unit areas with the same number of foot ends through the above steps 401-402, and then jumps to one target unit area from each intermediate unit area (at this time, each foot end corresponds to one intermediate unit area) through the above steps 101-102.

[0185] Exemplarily, referring to Figure 16 , Figure 16is another schematic diagram of a continuous jumping process of the legged robot provided by the embodiment of the present application, and the continuous jumping process includes a first jump and a second jump, wherein the first jump is a jump of the quadruped robot from one single stake on a plum stake to four single stakes in front, and the second jump is a jump from the four single stakes to another single stake on the basis of the first jump. The jumping process of the first jump is consistent with the jumping process shown in steps 401-402, and the jumping process of the second jump is consistent with the jumping process shown in steps 101-102. It should be noted that the two jumping processes can be continuously and alternately performed, and the embodiment of the present application does not make any limitation.

[0186] The jumping action of the legged robot implemented by the embodiment of the present application can realize the action of high and far take-off, flight, accurate landing, smooth completion of buffering, balance and posture control while combining the constraint condition of limited landing position, so that the legged robot can complete the set of jumping actions on various terrain areas such as plum stake arrays, block-shaped cushioning stakes and flat ground, and the motion ability of the robot can be expanded to show the excellent motion planning and control technology level of the robot. Meanwhile, the jumping action of the legged robot implemented in the embodiment of the present application has the advantages of high dynamics and high difficulty, and has the characteristics of bionics, which can improve the dynamics, difficulty and ornamental nature of the motion of the legged robot and expand the diversity of the motion of the legged robot.

[0187] In the following, an exemplary application of the embodiment of the present application in an actual application scenario will be described. In the application scenario, the legged robot is a quadruped robot, and the quadruped robot is located in a plum stake. In the application scenario, the execution process of the jumping action of the quadruped robot is as follows: the quadruped robot jumps from a high stake with a small stake surface of each of the four foot ends to a target stake surface with a limited stake surface in front of the body of the quadruped robot, and the quadruped robot stands on the target stake surface with the four foot ends collected downward and keeps balance.

[0188] In the related art, the jumping action that can be realized by the quadruped robot is mostly the continuous gait (such as bound or pronk) motion on flat ground or terrain with a large landing area, and the jumping action under the constraint condition of limited landing point is not realized.

[0189] The jump action realized on flat ground or terrain with large foot placement area is easier to balance because the four foot placement positions are dispersed, increasing the support projection area. However, for the case where the four foot placement positions are limited, especially when the four foot placement positions need to be brought close to the body directly below, the implementation in the related art is realized in the gait of continuous Trot in place, and the dynamic is relatively low, and there is no action after standing. For the combination of high dynamic jump action (four-foot jump or two-foot jump), there is no related technical solution for the action of bringing the four feet together to complete smooth buffering, balancing and posture control under the condition of high landing impact and limited foot placement position.

[0190] Based on this, the embodiment of the present application provides a motion control method of a legged robot, which can be applied to the scene of the legged robot jumping on the plum-pudding pile, and can realize the four-legged robot to take off in the four-legged jump mode from the normal standing posture, then bring the four feet together and land smoothly in a specific small area (a single plum-pudding pile surface) and keep balance, and complete the in-place head shaking and tail wagging action in the target posture.

[0191] In actual implementation, the action is based on the design of the four-legged robot, and the action is described by taking the plum-pudding pile scene with the highest difficulty as an example, as shown in Figure 4 , which shows that the terrain area is a plum-pudding pile, and each pile surface in the plum-pudding pile can be regarded as a mutually independent unit area in the terrain area. The four-legged robot is controlled to realize the jump action in the current scene on the plum-pudding pile, and the whole jump process can be divided into a take-off phase, a take-off phase, and a landing phase. Next, the three phases are described in turn:

[0192] First, in the take-off phase of the jump process, the four-legged robot experiences the following state changes in turn: initial state -> squatting and accumulating force -> starting four-foot ground kicking -> starting take-off. When the four-legged robot is in the initial state, the four feet of the four-legged robot stand on four pile surfaces in the normal standing posture (see Figure 7 No. 1 shows the state of the four-legged robot), and when the jump instruction is received, the body height is actively lowered to the planned take-off height (i.e., the squatting action is performed) (see Figure 7 No. 2 shows the state of the four-legged robot); when the four-legged robot is in the squatting and accumulating force state, the four feet of the four-legged robot kick the ground according to the joint reference trajectory (pre-planned motion trajectory of the hip joint and the knee joint) output by the planning phase at the same time, the four-legged robot has a specific body pitch angle (see Figure 7 No. 3 shows the squatting and accumulating force state), then the joint angle at the knee joint is controlled to increase (the acute angle forward gradually increases to the obtuse angle forward), the center of mass height of the body center of mass is raised, and the body center of mass is accelerated upward and forward to take off (i.e., start take-off), (see Figure 7 No. 4 shows the start of take-off state).

[0193] Secondly, in the take-off phase of the jumping process, the quadruped robot experiences the following state changes in turn: legs retracted and folded -> legs retracted and unfolded. The quadruped robot's four legs are completely off the ground, and the robot is completely in the air. In the air, the quadruped robot retracts the four legs and folds the legs (see the state shown in Figure 8 No. 1), that is, the control of retracting the four legs towards the center of the body, while also folding the legs upwards (the joint angle of the knee joint is reduced) first, then the quadruped robot retracts the four legs and unfolds the legs (see the state shown in Figure 8 No. 2), that is, in the case of retracting the four legs, continue to stretch forward and downward towards the (planned) landing point to prepare for landing.

[0194] Thirdly, in the landing phase of the jumping process, the quadruped robot experiences the following state changes in turn: contact with the pile surface -> leg bending and buffering -> keep balance. The four legs of the quadruped robot contact the pile surface at the same time (see the state shown in Figure 9 No. 1), then the four legs begin to bend and contract while applying torque control to each joint for buffering and deceleration (see the state shown in Figure 9 No. 2), and the body's center of mass is smoothly moved into the four-legged support area (see the state shown in Figure 9 No. 3), so as to finally achieve balance. Since the distance of the rear legs stretching forward is far (the extension distance of the knee joint and hip joint of the rear leg), and due to the configuration of the degrees of freedom of the legs of the quadruped robot, the knee joint of the rear leg will tend to move downward during the buffering process after landing. At the same time, in order to avoid the knee joint being lower than the height of the target pile surface, causing the collision between the shank of the rear leg and the edge of the pile surface, or the collision between the knee joint and the ground in other terrains such as flat ground, the amplitude of the bending and contraction of the rear leg and the downward amplitude of the knee joint of the rear leg need to be limited during the buffering motion control. In addition, it is also necessary to ensure that the foot end does not bounce due to impact or slide due to leg posture and force after contacting the pile surface.

[0195] In actual application, the quadruped robot completes the jumping action towards the target pile surface through the above three phases of the jumping process, and maintains balance on the target pile surface in the posture of the four legs retracted under the body.

[0196] In actual implementation, when the quadruped robot is in the balance state on the target pile surface (at this time it is in the state of standing with small landing point spacing), it can also receive instructions for controlling the quadruped robot to swing its head and tail, and perform posture control of the body of the quadruped robot on the target pile surface in the current balance posture, to move along at least one of the three degrees of freedom of roll (see Figure 10 No. 1), pitch (see Figure 10 No. 3), and yaw (see Figure 10 No. 2), to realize the bionic action of swinging the head and tail.

[0197] Finally, it needs to be noted that the quadruped robot can complete the whole set of jumping actions on various terrains such as plum-pudding piles, which can expand the motion capability of the robot and demonstrate the excellent motion planning and control technology level of the robot.

[0198] Next, the motion control method of the legged robot provided by the embodiments of the present application is described from the technical side. In terms of technical implementation, the embodiments of the present application plan the most economical jumping trajectory by a nonlinear optimization method, and perform motion control based on MPC, so as to realize a whole set of high-dynamic and high-difficult jumping actions of the quadruped robot, including starting from a normal standing posture, accumulating force, kicking the ground, taking off, folding the four legs, landing with a small foot spacing, buffering, balancing, and shaking the head and tail. During the whole jumping process in the current scene, the functional modules of the quadruped robot for realizing the current jumping action mainly include a jumping trajectory planning module, a jumping control module, and a terrain perception module, etc.

[0199] First, the operations performed by the jumping trajectory planning module are described in detail. The whole set of jumping actions needs to ensure the forward jumping distance and the upward jumping height, and at the same time, it needs to realize buffering and balancing in a very short time after landing, so the output torque of the joint motor is required to be very high. Referring to Figure 17 , Figure 17 is a schematic diagram of a planar model of the quadruped robot provided by the embodiments of the present application, based on the planar model of the quadruped robot shown in the figure, the generalized coordinates are defined as:

[0200] q = [x, z, pitch, q hip,front , q knee,front , q hip,hind , q knee,hind ] T

[0201] Among them, x represents the X-axis component of the mass center, z represents the Z-axis component of the mass center, pitch represents the pitch angle of the mass center, q hip,front represents the joint angle of the front leg hip joint, q knee,front represents the joint angle of the front leg knee joint, q hip,hind represents the joint angle of the rear leg hip joint, and q knee,hind represents the joint angle of the rear leg knee joint. The following nonlinear optimization problem is constructed to plan the action trajectory, and the optimization objective is to minimize the total driving force required to complete the whole set of actions:

[0202]

[0203] s.t.

[0204] f min ≤ f i ≤ f max , (1)

[0205] |f i,x |≤μf i,z , (2)

[0206] 0≤p hip,z ,0≤p knee,z , (3)

[0207]

[0208] where f i is the contact force vector of all feet at the i-th sampling point, W is the weight parameter matrix, and k is the number of sampling points of the motion. The constraints include (1) joint torque and contact force limits, (2) foot-ground friction cone, (3) hip and knee heights above the ground, and (4) dynamics constraints and no-slip at the feet, where f min ,f max are the minimum and maximum values of the contact force, f i,x ,f i,z are the x and z components of the contact force at the i-th sampling point, μ is the friction coefficient, p hip,z ,p knee,z represent the heights of the hip and knee joints, H, S, C, and G are the generalized mass matrix, selection matrix, friction and Coriolis force related terms matrix, and gravity terms matrix, respectively, J front , J hind are the Jacobian matrices of the front and rear legs, respectively. The first row in constraint (4) represents the dynamics constraints, and the second and third rows represent no-slip at the front and rear feet, respectively.

[0209] To accelerate the solution of the optimization problem and obtain trajectories that are as close as possible to the designed motion on the product side, the generalized coordinate sampling points at the motion segmentation points need to be given, in order of the initial state q start , the four-foot disengagement from the pile state q lift-off , the four-foot landing on the single pile state q lift-off , the four-foot landing on the single pile state q touchdown , and the single pile balance standing state q end . The unique point in this motion is the four-foot folding landing on the single pile, and the corresponding sampling points are given as follows:

[0210] q touchdown = [x pole +x offset , z pole +z offset , pitch touchdown , IK (p pole +p offset,front ), IK (p pole +p offset,hind )]T ,

[0211] where the center of mass reference position is given based on the position of the single pole (x pole ,z pole ) plus a specific offset (x offset ,z offset ) according to the action design, the reference pitch angle of the body is also given according to the action design, p touchdown +p pole +p offset,front / hind represent the expected foot landing positions on the pole, i.e. based on the pole surface center position p pole plus the foot retraction offset p offset,front / hind , while the reference joint angles of the legs are calculated according to the inverse kinematics (IK(·)) from the expected foot landing positions on the pole and the center of mass reference position, through p offset,front and p offset,hind . The settings of p offset,front and p offset,hind can adjust the foot landing positions when retraction.

[0212] Based on the above method, the most economical reference jumping trajectory that satisfies the constraint condition is obtained after solving the optimization problem. The reference jumping trajectory can effectively ensure the feasibility of the whole jumping action, so that the control module can smoothly predict the reference trajectory. For the head shaking and tail wagging action after balance, it can be realized by pre-programming or online remote control instruction method.

[0213] Secondly, the technical implementation of the jumping control module is described. For the take-off stage, since the real machine conforms to the theoretical plane model, the joint reference torque τ ref in the optimal reference trajectory can be directly used as the feedforward torque τ feedforward , while the PD control feedback torque τ feedback is calculated based on the joint reference angle q ref and angular velocity , so as to realize the determination and prediction of the take-off action, i.e.:

[0214]

[0215] where k p and k d are the PD gains, respectively.

[0216] The leg action control is realized only by joint PD control in the emptying phase. After landing, the four legs on the real machine will also converge to the body sagittal plane in the lateral direction of the body, so it does not conform to the plane model and the optimal joint reference torque cannot be directly obtained. In addition, due to the errors between the body posture, position and landing time and the reference trajectory in the emptying phase, the controller based on MPC is used to realize the buffering, balancing and subsequent head shaking and tail wagging. The MPC optimization problem is constructed based on the optimal reference trajectory obtained by planning and the current state of the robot, and the optimal joint torque is obtained by solving.

[0217] Finally, the operations performed by the terrain perception module are described. The entire set of actions can be realized with or without the terrain perception module. When the perception module is provided, the independent single pile is identified and positioned according to the information, i.e., p pole online trajectory planning is performed; when the perception module is not provided, the prediction and determination can be based on the action trajectory planned offline, provided that prior knowledge about the terrain is available.

[0218] Due to the existence of the emptying phase, the actual emptying time and the emptying trajectory will have certain errors with the planned reference trajectory, which leads to the difference between the actual landing time and the planned time. If the foot end ground contact sensing capability is not available, the quadruped robot cannot timely switch the MPC force control, thereby causing problems such as foot end collision rebound and sliding on the pile surface. In the embodiment of the present application, the foot end contact force is calculated based on the joint torque, and then it is judged whether the pile surface is contacted through the foot bottom force sensing, thereby providing timely sensing information for the control algorithm switching. The specific calculation formula is as follows.

[0219] f = J -T τ

[0220]

[0221] where b contact represents the Boolean type of foot bottom contact state, f normal is the component of the foot bottom contact force in the normal direction of the contact surface, and f threshold is the threshold for judging contact or not. The judgment logic of the foot bottom force sensing is that if the component of the foot bottom contact force in the normal direction of the contact surface is greater than the given threshold, it is determined as the contact state (1), otherwise it is determined as the non-contact state (0).

[0222] It should be noted that the jumping action can be quadruped jumping or biped jumping. For biped jumping, it can be divided into four phases of energy storage, quadruped ground lifting, front leg lifting and rear leg ground lifting, and emptying. For the trajectory planning algorithm, in addition to nonlinear optimization, other simple methods such as heuristic, spline interpolation, etc. can also be used, although the optimality of the trajectory cannot be guaranteed, but similar actions can also be realized. In addition to MPC, the motion prediction can also be realized based on the QP controller.

[0223] The application embodiment is applied to complete the jumping action, improve the dynamic, difficulty and ornamental of the quadruped robot movement, and expand the diversity of the quadruped robot movement, so that the quadruped robot can not only move on the flat ground, but also can complete high-difficulty action display in scenes such as plum-blossom staking and small stone piers.

[0224] The following continues to describe an exemplary structure of the embodiment of the motion control device 555 of the legged robot provided by the application as a software module. In some embodiments, as shown in FIG. 5, the software module stored in the motion control device 555 of the legged robot in the memory 550 can include: Figure 2

[0225] In some embodiments, the target unit area is located in front of the legged robot, and the control module is further configured to control the legged robot to jump forward to the target unit area through a take-off stage, a flight stage and a landing stage in response to the jumping instruction.

[0226] In some embodiments, the control module is further configured to control the legged robot to sequentially perform the following operations in the take-off stage: squatting to accumulate power, foot-end ground pushing and starting flight; in the flight stage, the legged robot sequentially performs the following operations: in-air foot and leg folding and in-air leg stretching; and in the landing stage, the legged robot sequentially performs the following operations: foot-end landing, leg bending buffering and body balancing.

[0227] In some embodiments, the control module is further configured to, in the process of the legged robot performing the squatting to accumulate power, bend the leg of the legged robot to lower the center of mass of the legged robot from an initial height to a take-off height; in the process of the legged robot performing the foot-end ground pushing, increase the torque of the leg joint to increase the friction between each foot end and the foot-end contact surface; and in the process of the legged robot performing the starting flight, perform a take-off action in a target pitch angle forward and upward to increase the height of the knee joint and the height of the center of mass, the height of the center of mass being higher than the initial height.

[0228] In some embodiments, the control module is further configured to, in the process of the legged robot performing the in-air foot and leg folding, control each leg of the legged robot to fold the corresponding foot end upward to below the body center; and in the process of the legged robot performing the in-air leg stretching, control the front leg to stretch toward the target unit area, so that each foot end of the legged robot can land on the target unit area.

[0229] ​In some embodiments, the control module is also used to control each foot end of the leg robot to contact the contact surface within the target unit area at the same time during the process of the leg robot landing the foot end; during the process of the leg robot performing the bent leg buffering, the leg robot performs the bent leg action of bending and contracting each leg, and increases the torque of the leg joints to reduce the speed of the leg robot; during the process of the leg robot performing the body balance, the leg robot adjusts its body posture to move the center of mass of the leg robot to the target unit area.

[0230] In some embodiments, the legged robot includes a tail, and the control module is further used to control the legged robot to continuously perform at least one of the following operations within the target area based on at least one degree of freedom: shaking head and tail; wherein the degree of freedom includes at least one of roll, pitch, and yaw.

[0231] In some embodiments, the target unit area is located at the center of each of the unit areas where the foot end is located, and the control module is also used to respond to the jumping instruction to control the foot-type robot to jump upward to the target unit area through the take-off stage, the flight stage and the landing stage.

[0232] In some embodiments, the control module is also used to, during the take-off phase, enable the legged robot to sequentially perform the operations of squatting to accumulate strength and starting to take off; during the take-off phase, enable the legged robot to sequentially perform the operations of retracting the feet and legs in the air; during the landing phase, enable the legged robot to sequentially perform the operations of landing with the foot, bending the legs for cushioning, and balancing the body.

[0233] In some embodiments, the control module is also used to bend the legs of the legged robot during the process of the legged robot performing the squatting and accumulating power, so as to lower the center of mass of the legged robot from an initial height to a take-off height; during the process of the legged robot starting to take off, the legged robot bounces upward to increase the height of the knee joint and the height of the center of mass, and the height of the center of mass is higher than the initial height; during the process of the legged robot retracting its feet and legs in the air, the legged robot controls the torso to be in an upright state, and each leg of the legged robot drives the corresponding foot end to retract upward to below the torso.

[0234] In some embodiments, the legged robot is a quadruped robot, and the control module is further used to respond to the jumping instruction. When the jumping instruction indicates a two-legged jump, the control module controls the legged robot to alternately use two front legs and two hind legs to jump to the target unit area.

[0235] In some embodiments, the control module is further configured to control the hopping process of the legged robot using two front legs and two back legs alternately, including: a power accumulation stage, a quadruped ground-pushing stage, a front leg take-off and back leg ground-pushing stage, a flight stage, and a landing stage.

[0236] In some embodiments, the unit area and the target unit area are both end surface areas of a single stake in the plum-blossom stake, and the unit area and the target unit area have the same shape and size.

[0237] In some embodiments, the control module is further configured to predict a hopping trajectory of the legged robot to obtain a reference hopping trajectory, obtain a constraint condition corresponding to the hopping instruction, the constraint condition being used to constrain the hopping of the legged robot, and determine state parameters of at least two key points of the reference hopping trajectory based on the reference hopping trajectory and the constraint condition. Correspondingly, the control module is further configured to control the legged robot to hop to the target unit area based on the state parameters of the key points in response to the hopping instruction.

[0238] In some embodiments, the control module is further configured to obtain at least two key points in the reference hopping trajectory, wherein the key points include at least one of the following: a starting point of the reference hopping trajectory, a vertex of the reference hopping trajectory, and an ending point of the reference hopping trajectory.

[0239] Embodiments of the present application provide a computer program product or a computer program, which includes computer instructions stored in a computer readable storage medium. A processor of a computer device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions to cause the computer device to perform the motion control method of the legged robot provided in the embodiments of the present application.

[0240] Embodiments of the present application provide a computer readable storage medium storing executable instructions, wherein the executable instructions, when executed by a processor, cause the processor to perform the motion control method of the legged robot provided in the embodiments of the present application, for example, as shown in the motion control method of the legged robot. Figures 3A-3B Embodiments of the present application provide a computer readable storage medium storing executable instructions, wherein the executable instructions, when executed by a processor, cause the processor to perform the motion control method of the legged robot provided in the embodiments of the present application, for example, as shown in the motion control method of the legged robot.

[0241] In some embodiments, the computer readable storage medium can be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, flash memory, magnetic surface memory, optical disc, or CD-ROM; or various devices including one or any combination of the above memories.

[0242] In some embodiments, executable instructions can be in the form of programs, software, software modules, scripts, or code, written in any programming language, including compiled or interpreted languages, or declarative or procedural languages; and can be deployed in any form, including as a stand-alone program or as a module, component, subroutine, or other unit suitable for use in a computing environment.

[0243] By way of example, an executable instruction can be, but is not limited to, a file in a file system, can be stored in a part of a file that holds other programs or data, for example, one or more scripts stored in a markup language document, in a single file dedicated to the program in question, or in multiple coordinated files, for example, files that store one or more modules, sub programs, or portions of code.

[0244] By way of example, an executable instruction can be, but is not limited to, a file in a file system, can be stored in a part of a file that holds other programs or data, for example, one or more scripts stored in a markup language document, in a single file dedicated to the program in question, or in multiple coordinated files, for example, files that store one or more modules, sub programs, or portions of code.

[0245] In summary, the embodiments of the present application improve the dynamics, difficulty and watchability of the quadruped robot movement, expand the diversity of the quadruped robot movement, so that the quadruped robot can not only move on the flat ground, but also complete high-difficulty action demonstration in scenes such as plum-blossom stake and small stone piers.

[0246] The above merely describes the embodiments of the present application, but is not used to limit the protection scope of the present application. Any modification, equivalent replacement and improvement made within the spirit and scope of the present application shall be included in the protection scope of the present application.

Claims

1. A motion control method for a legged robot, characterized in that: The legged robot includes at least two foot ends, and the method includes: When the respective foot ends of the legged robot are respectively positioned in independent unit areas, a jump instruction for the legged robot is received; The jump instruction is used to instruct the legged robot to jump from at least two unit areas currently located to a target unit area; the unit areas and the target unit area are both end surface areas of a single pile in a plum blossom pile, and the unit area and the target unit area have the same shape and size; In response to the jump instruction, controlling the legged robot to jump to the target unit area so that each foot end of the legged robot is retracted into the target unit area, and in the target unit area, the distance between any two foot ends of the legged robot is smaller than the distance between the corresponding two foot ends before the jump; Wherein, in response to the jump instruction, controlling the legged robot to jump to the target unit area includes: In response to the jump instruction, a jump trajectory of the legged robot is predicted to obtain a reference jump trajectory; constraints corresponding to the jump instruction are obtained, wherein the constraints are used to constrain the jump of the legged robot; state parameters of at least two key points of the reference jump trajectory are determined based on the reference jump trajectory and the constraints; and based on the state parameters of each key point, the legged robot is controlled to jump to the target unit area; Alternatively, when the target unit area is located in front of the legged robot, in response to the jump instruction, the legged robot is controlled to jump forward to the target unit area through a take-off phase, a flying phase, and a landing phase; In the take-off phase, the legged robot sequentially performs the operations of squatting to accumulate power, pushing off the ground with the foot end, and starting to take off; in the take-off phase, the legged robot sequentially performs the operations of retracting the foot and leg in the air, and extending the leg in the air; in the landing phase, the legged robot sequentially performs the operations of landing with the foot end, bending the leg for cushioning, and balancing the body; In the process of squatting and accumulating power, the leg of the leg robot bends to lower the center of mass of the leg robot from an initial height to a take-off height; in the process of pushing off the ground with the foot end, the leg robot increases the torque of the leg joint to increase the friction force of the contact surface between each foot end; in the process of starting to take off, the leg robot performs a take-off action forward and upward at a target pitch angle to increase the height of the knee joint and the height of the center of mass, and the height of the center of mass is higher than the initial height; Alternatively, when the target unit area is located at the center of each of the unit areas where the foot ends are located, in response to the jump instruction, the legged robot is controlled to jump upward to the target unit area through a take-off phase, a soaring phase, and a landing phase, wherein in the take-off phase, the legged robot sequentially performs the operations of squatting to accumulate power and starting to soar; in the soaring phase, the legged robot performs the operations of retracting the feet and legs in the air; in the landing phase, the legged robot sequentially performs the operations of landing the foot ends, bending the legs for cushioning, and balancing the body; during the squatting and power accumulation process of the legged robot, the legs of the legged robot are bent to lower the center of mass of the legged robot from an initial height to a take-off height; during the soaring phase of the legged robot, the legged robot bounces upward to increase the height of the knee joint and the height of the center of mass, the height of the center of mass being higher than the initial height; during the soaring phase of the legged robot, the trunk of the legged robot is controlled to be in an upright state, and each leg of the legged robot drives the corresponding foot ends to be retracted upward to below the trunk; When the legged robot includes a tail, after controlling the legged robot to jump to the target unit area, the legged robot is controlled to continuously perform at least one of the following operations within the target unit area based on at least one degree of freedom: shaking head and tail; wherein the degree of freedom includes at least one of roll, pitch, and yaw.

2. The method according to claim 1, wherein The legged robot is a quadruped robot; in the airborne phase, the legged robot sequentially performs the operations of retracting the legs and feet in the air and extending the legs in the air; Wherein, during the process of the legged robot extending its legs in the air, the legged robot controls its front legs to extend toward the target unit area, so that each foot end of the legged robot can land on the target unit area.

3. The method according to claim 1, wherein During the landing phase, the legged robot sequentially performs foot landing, leg bending for cushioning, and body balancing operations; Wherein, during the process of the foot end of the legged robot landing, the legged robot controls each foot end to contact the contact surface in the target unit area at the same time; During the leg-bending buffering process, the leg-bending robot performs a leg-bending action in which each leg is bent and contracted, and increases the torque of the leg joints to reduce the speed of the leg-bending robot. During the process of the legged robot performing the body balancing, the legged robot adjusts its body posture to move the center of mass of the legged robot into the target unit area.

4. The method according to claim 1, wherein The legged robot is a quadruped robot, and in response to the jump instruction, controlling the legged robot to jump to the target unit area includes: In response to the jump instruction, when the jump instruction indicates a two-legged jump, the legged robot is controlled to alternately use two front legs and two hind legs to jump to the target unit area.

5. The method according to claim 4, wherein The process of the legged robot using two front legs and two hind legs alternately to jump includes: a power accumulation stage, a four-legged pushing-off stage, a front leg leaving the ground and a hind leg pushing-off stage, a flying stage and a landing stage.

6. The method according to claim 1, wherein Before determining the state parameters of at least two key points of the reference jump trajectory based on the reference jump trajectory and the constraint conditions, the method further includes: Obtaining at least two key points in the reference jump trajectory; The key points include at least one of the following: The starting point of the reference jump trajectory, the vertex of the reference jump trajectory, and the ending point of the reference jump trajectory.

7. A motion control device for a legged robot, characterized in that: The device comprises: a receiving module, configured to receive a jump instruction for the legged robot while each foot of the legged robot is positioned in a mutually independent unit area; wherein the legged robot includes at least two foot ends, and the jump instruction is configured to instruct the legged robot to jump from the at least two unit areas currently located to a target unit area; the unit areas and the target unit area are both end surface areas of a single pile in a plum blossom pile, and the unit areas and the target unit area have the same shape and size; a control module, configured to control the legged robot to jump to the target unit area in response to the jump instruction, so that each foot end of the legged robot is retracted into the target unit area, and in the target unit area, the distance between any two foot ends of the legged robot is smaller than the distance between the corresponding two foot ends before the jump; The control module is further configured to predict a jumping trajectory of the legged robot in response to the jumping instruction to obtain a reference jumping trajectory; obtain constraint conditions corresponding to the jumping instruction, wherein the constraint conditions are used to constrain the jumping of the legged robot; determine state parameters of at least two key points of the reference jumping trajectory based on the reference jumping trajectory and the constraint conditions; and control the legged robot to jump to the target unit area based on the state parameters of each of the key points; Alternatively, when the target unit area is located in front of the legged robot, in response to the jump instruction, the legged robot is controlled to jump forward to the target unit area through a take-off phase, a flying phase, and a landing phase; In the take-off phase, the legged robot sequentially performs the operations of squatting to accumulate power, pushing off the ground with the foot end, and starting to take off; in the take-off phase, the legged robot sequentially performs the operations of retracting the foot and leg in the air, and extending the leg in the air; in the landing phase, the legged robot sequentially performs the operations of landing with the foot end, bending the leg for cushioning, and balancing the body; In the process of squatting and accumulating power, the leg of the leg robot bends to lower the center of mass of the leg robot from an initial height to a take-off height; in the process of pushing off the ground with the foot end, the leg robot increases the torque of the leg joint to increase the friction force of the contact surface between each foot end; in the process of starting to take off, the leg robot performs a take-off action forward and upward at a target pitch angle to increase the height of the knee joint and the height of the center of mass, and the height of the center of mass is higher than the initial height; Alternatively, when the target unit area is located at the center of each of the unit areas where the foot ends are located, in response to the jump instruction, the legged robot is controlled to jump upward to the target unit area through a take-off phase, a soaring phase, and a landing phase, wherein in the take-off phase, the legged robot sequentially performs the operations of squatting to accumulate power and starting to soar; in the soaring phase, the legged robot performs the operations of retracting the feet and legs in the air; in the landing phase, the legged robot sequentially performs the operations of landing the foot ends, bending the legs for cushioning, and balancing the body; during the squatting and power accumulation process of the legged robot, the legs of the legged robot are bent to lower the center of mass of the legged robot from an initial height to a take-off height; during the soaring phase of the legged robot, the legged robot bounces upward to increase the height of the knee joint and the height of the center of mass, the height of the center of mass being higher than the initial height; during the soaring phase of the legged robot, the trunk of the legged robot is controlled to be in an upright state, and each leg of the legged robot drives the corresponding foot ends to be retracted upward to below the trunk; When the legged robot includes a tail, after controlling the legged robot to jump to the target unit area, the legged robot is controlled to continuously perform at least one of the following operations within the target unit area based on at least one degree of freedom: shaking head and tail; wherein the degree of freedom includes at least one of roll, pitch, and yaw.

8. The device according to claim 7, characterized in that The legged robot is a quadruped robot; the control module is also used for the legged robot to sequentially perform the operations of retracting the feet and legs in the air and extending the legs in the air during the take-off stage; wherein, during the process of the legged robot extending its legs in the air, the legged robot controls the front legs to extend toward the target unit area so that each foot end of the legged robot can land on the target unit area.

9. The device according to claim 7, characterized in that The control module is also used for the legged robot to sequentially perform foot landing, leg bending buffering and body balancing operations during the landing phase; wherein, during the foot landing process of the legged robot, the legged robot controls each foot end to contact the contact surface within the target unit area at the same time; during the leg bending buffering process of the legged robot, the legged robot performs a leg bending action of bending and contracting each leg, and increases the torque of the leg joints to reduce the speed of the legged robot; during the body balancing process of the legged robot, the legged robot adjusts its body posture to move the center of mass of the legged robot to the target unit area.

10. The device according to claim 7, characterized in that The legged robot is a quadruped robot, and the control module is further used to respond to the jumping instruction. When the jumping instruction indicates a two-legged jump, the control module controls the legged robot to alternately use two front legs and two hind legs to jump to the target unit area.

11. The device according to claim 10, characterized in that The control module is also used for the process in which the legged robot alternately uses two front legs and two hind legs to jump, including: a power accumulation stage, a four-legged pushing-off stage, a front leg leaving the ground and a hind leg pushing-off stage, a take-off stage and a landing stage.

12. The device according to claim 7, characterized in that The control module is further configured to obtain at least two key points in the reference jump trajectory before determining the state parameters of the at least two key points in the reference jump trajectory based on the reference jump trajectory and the constraint conditions; wherein the key points include at least one of the following: a starting point of the reference jump trajectory, a vertex of the reference jump trajectory, and an end point of the reference jump trajectory.

13. A legged robot, characterized in that: The foot-type robot comprises: A robot body having at least two foot ends; A controller is provided on the robot body and is used to execute the motion control method of the legged robot according to any one of claims 1 to 6.

14. A computer-readable storage medium storing executable instructions, characterized in that: When the executable instructions are executed by the processor, the motion control method of the legged robot according to any one of claims 1 to 6 is implemented.

15. A computer program product comprising computer executable instructions, characterized in that When the computer executable instructions are executed by a processor, the motion control method of the legged robot according to any one of claims 1 to 6 is implemented.

Images