A quadruped robot gait adaptive control method based on active anti-interference

Abstract

This invention belongs to the field of robot motion control technology and discloses a gait adaptive control method for quadruped robots based on active anti-interference. First, a fixed trajectory of the foot landing point is achieved through a swing phase foot landing point compensation strategy. Then, an improved EID estimator is introduced to estimate the system disturbance in real time and obtain the compensation amount to the bottom force of the supporting foot in accordance with the quadruped robot's constraints, thereby improving the quadruped robot's anti-interference ability in the fixed foot landing state and ensuring the stability of the robot's posture. Only when the disturbance exceeds the body's redundancy tolerance range is the foot landing point follow-up mode temporarily switched to ensure the safety of the whole machine. Finally, stable and high-precision motion control of the quadruped robot is achieved in extreme working scenarios.

Description

Technical Field

[0001] This invention belongs to the field of robot motion control technology, specifically relating to an adaptive gait control method for quadruped robots based on active anti-interference. Background Technology

[0002] Quadruped robots, with their excellent terrain adaptability and mobility, have broad application prospects in routine inspections, field search and rescue, outdoor operations, special protection, and various unstructured extreme work scenarios. Some of these extreme work scenarios involve fixed-point operations in high-risk areas (such as mine clearance in the field, fixed-point detection in nuclear radiation areas, and sampling in toxic and hazardous environments) and passage through narrow, confined spaces (such as underground tunnels, pipeline corridors, and gaps in collapsed buildings). These scenarios place stringent demands on the motion control of quadruped robots: when subjected to sudden disturbances, the range of leg movement must be strictly controlled to prevent significant positional deviations, and the footing point must be precisely fixed at a preset location. This is to avoid robot instability, triggering dangerous scenarios, or compromising operational accuracy due to leg movement or footing point deviation, such as accidentally triggering explosives in a minefield or getting stuck and tipping over in a narrow space.

[0003] Existing quadruped robots typically employ periodic gait for motion control. The gait cycle can be divided into a support phase and a swinging phase: the support phase achieves body support and posture stability through the distribution of ground reaction forces on the feet, while the swinging phase achieves speed tracking and motion direction adjustment through foot trajectory generation and foot placement planning. Under normal operating conditions, foot placement planning in the swinging phase often employs heuristic methods or online adjustment strategies based on stability indices, allowing the foot placement to change with the center of mass state or velocity error, thereby improving the recovery capability from disturbances.

[0004] However, in the aforementioned extreme operating scenarios, the landing point usually needs to be fixed at a preset location or within a limited area. The conventional control logic of adjusting the landing point dynamically conflicts with the requirements of extreme operations: even a small disturbance may cause the existing landing point dynamic strategy to deviate from the preset location, resulting in reduced operational accuracy and inducing safety risks. On the other hand, under the constraint of landing point locking, when the support phase relies solely on conventional model predictive control for foot force distribution, it is limited by constraints such as friction cone, normal force range, and joint torque saturation. Faced with larger external force disturbances, it may experience insufficient attitude recovery, increased fuselage sway, or even instability.

[0005] Therefore, there is an urgent need for an adaptive gait control method for quadruped robots that is suitable for extreme working scenarios, in order to solve the technical defects of existing technologies that cannot guarantee fixed foot trajectory and no deviation of foot landing point, nor can they effectively resist disturbances under the premise of fixed foot trajectory, and break through the application limitations in extreme working conditions. Summary of the Invention

[0006] The purpose of this invention is to overcome the problems in existing technologies where quadruped robots cannot meet the core prerequisite of fixed landing trajectory in extreme working scenarios, and where insufficient anti-disturbance capability and easy instability of the body are easily caused when the foot trajectory is fixed. This invention provides a quadruped robot gait adaptive control method based on active anti-disturbance. First, a fixed landing trajectory is achieved by using a swing phase landing point compensation strategy. Then, an improved equivalent-input-disturbance (EID) estimator is introduced to estimate the system disturbance in real time and obtain the compensation amount to the bottom force of the supporting phase foot that meets the quadruped robot's constraints. This improves the quadruped robot's anti-disturbance capability in the fixed landing state and ensures the stability of the body posture. Only when the disturbance exceeds the body's redundancy tolerance range is the landing point follow-up mode temporarily switched to ensure the safety of the whole machine. Finally, stable and high-precision motion control of the quadruped robot is achieved in extreme working scenarios.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0008] A gait adaptive control method for a quadruped robot based on active anti-interference includes the following steps:

[0009] Collect the position of the quadruped robot's center of mass in the world coordinate system and the state of its legs at the current moment, and use the center of mass planner to obtain the expected position of the center of mass at the current moment;

[0010] For legs in the swing phase, based on the current position of the centroid in the world coordinate system, the current expected position of the centroid, and the previous expected position of the centroid, it is determined whether the quadruped robot is subjected to an external force disturbance greater than the disturbance threshold.

[0011] If the quadruped robot is subjected to an external force disturbance greater than the disturbance threshold, the current footing position of the legs is determined by the center of mass following method; otherwise, the current footing position of the legs is determined by the expected center of mass position compensation method.

[0012] For legs in the support phase, an improved equivalent input perturbation estimator is used to estimate external perturbations, and the estimated external perturbation values ​​are applied to the foot contact force of the leg to compensate for the external perturbation. The improved equivalent input perturbation estimator is obtained by replacing the generalized inverse substitution of the B matrix in the equivalent input perturbation estimator with a quadratic programming method.

[0013] Several alternative methods are provided below, but they are not intended as additional limitations on the overall solution above. They are merely further additions or optimizations. Provided there are no technical or logical contradictions, each alternative method can be combined individually with respect to the overall solution above, or multiple alternative methods can be combined with each other.

[0014] Preferably, the step of determining whether the quadruped robot is subjected to an external force disturbance greater than the disturbance threshold based on the current position of the centroid in the world coordinate system, the expected position of the centroid at the current moment, and the expected position of the centroid at the previous moment includes:

[0015] If the expected centroid position at the current moment is equal to the expected centroid position at the previous moment, and the absolute value of the difference between the current centroid position in the world coordinate system and the expected centroid position at the current moment is less than the disturbance threshold, then the quadruped robot is not subjected to an external force disturbance greater than the disturbance threshold; otherwise, it means that the quadruped robot is subjected to an external force disturbance greater than the disturbance threshold.

[0016] Preferably, the method of determining the current foot placement position of the leg using a center-of-mass tracking approach includes:

[0017] Based on the current position of the center of mass in the world coordinate system, the velocity of the center of mass, the expected velocity of the center of mass, and the phase of the leg swing at the current moment, the position of the leg landing point at the current moment is calculated using a heuristic landing point planning method.

[0018] Preferably, the method of determining the current foot placement position of the leg using the desired centroid position compensation method includes:

[0019] Based on the radius and azimuth of the circular motion of the leg and foot on the plane, determine the position of the leg's footing point in the world coordinate system. shaft and Projection components along the axial direction;

[0020] Using the world coordinate system shaft and The desired centroid position along the axis, respectively, for the leg footing point in the world coordinate system. shaft and Compensation is performed on the projection components along the axis to obtain the current foot placement point of the leg in the world coordinate system. shaft and Position along the axis.

[0021] As a preferred method, the foot trajectory is planned using Bézier curves based on the current foot landing point position, thus obtaining the continuous foot position within each control cycle.

[0022] Preferably, the method of estimating external disturbances using an improved equivalent input disturbance estimator includes:

[0023] Construct a three-dimensional single rigid body model of the quadruped robot and obtain its state space model.

[0024] Define the observation residuals and design the state observer based on the system state, controller inputs, and observation outputs;

[0025] The observation residual is amplified by the state observer gain, and the dimensions of the amplified observation residual are converted into forces and torques by the transformation matrix to obtain the perturbation force acting on the center of mass.

[0026] Based on the dynamics of the center of mass, a mapping relationship is established between the foot compensation force and the disturbance force acting on the center of mass. Based on the mapping relationship, a quadratic programming problem is constructed, and the foot compensation force is obtained by solving the quadratic programming problem.

[0027] The foot compensation force is passed through a filter, and the external disturbance estimate is obtained based on the filter output.

[0028] Preferably, the transformation matrix contains two rows and two columns of elements, wherein the elements in the first row and first column are the ratio of the inertial tensor to the control period in the world coordinate system of the quadruped robot, the elements in the first row and second column and the elements in the second row and first column are both zero matrices, and the elements in the second row and second column are the ratio of the product of the mass of the quadruped robot and the identity matrix to the control period.

[0029] Preferably, the construction of the quadratic programming problem is as follows:

[0030]

[0031] In the formula, For parameter matrices, for The compensating force at the toes at all times, for The disturbance force that constantly acts on the center of mass. Indexing the legs of a quadruped robot. This is the set of legs of the quadruped robot that are currently in a swinging phase. This is the set of legs of the quadruped robot in the current supporting phase. This represents the friction cone constraint condition. Represents the friction cone constraint function. Indicates the friction cone constraint threshold. , The upper and lower bounds of the normal force; Indicates in Time of the first The contact force at the foot of one leg. for Time of the first The compensating force at the foot of each leg, Indicates the contact force at the foot. In the normal component, For foot compensation force In the normal direction of the component.

[0032] The present invention provides a gait adaptive control method for quadruped robots based on active anti-interference, which has the following advantages compared with the prior art:

[0033] This invention addresses the core needs of extreme work scenarios (fixed-point operation in high-risk areas, passage through narrow and confined spaces, etc.) by using a swing phase foot landing point position compensation strategy to ensure that the trajectory of the quadruped robot's feet does not change when disturbed. This solves the defects of existing technologies where the foot landing point moves with the center of mass and cannot meet the requirements of extreme scenarios, thus avoiding safety accidents and work failures caused by leg movement and foot landing point deviation.

[0034] To address the multi-contact constraint characteristics of quadruped robots, this invention designs an improved equivalent input perturbation estimator. This avoids the violation of foot constraints that occurs when traditional equivalent input perturbation estimators equate perturbations to input channels. This invention employs Quadratic Programming (QP) to solve for the foot force compensation under constraints, thus preventing such violations. This method can estimate and compensate for sudden perturbations in real time, thereby ensuring stronger body anti-interference capabilities and improving the quadruped robot's motion control capabilities without altering the foot landing trajectory. Attached Figure Description

[0035] Figure 1 This is a flowchart of a gait adaptive control method for a quadruped robot based on active anti-interference, according to the present invention.

[0036] Figure 2 This is a block diagram of the compensation for the support phase based on the improved EID estimator of the present invention;

[0037] Figure 3 This is a comparison diagram of the trajectory of the quadrilateral projection center of the landing point when the robot of the present invention is subjected to lateral disturbance during movement, with and without landing point position compensation;

[0038] Figure 4 This is a comparison diagram showing the centroid deviation of the robot of the present invention when its landing point trajectory is fixed and subjected to lateral disturbance. Detailed Implementation

[0039] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0040] It should be noted that when a component is said to be "connected" to another component, it can be directly connected to the other component or it can be connected to a component in between; when a component is said to be "fixed" to another component, it can be directly fixed to the other component or it can be connected to a component in between.

[0041] 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 invention pertains. The terminology used herein in the description of the invention is for the purpose of describing particular embodiments only and is not intended to limit the invention.

[0042] Example 1:

[0043] To overcome the shortcomings of existing technologies, this embodiment provides a gait adaptive control method for quadruped robots based on active anti-interference, such as... Figure 1 As shown, it includes the following steps:

[0044] Step 1: Collect the position of the quadruped robot's center of mass in the world coordinate system and the state of its legs at the current moment, and use the center of mass planner to obtain the desired center of mass position at the current moment.

[0045] This embodiment performs gait adaptive control based on the real-time state of the quadruped robot. At each moment, the current state data of the quadruped robot is collected, including but not limited to the body posture, angular velocity, and linear acceleration collected by the inertial measurement unit; the joint angles, angular velocities, and torques collected by the joint encoders; the foot contact state collected by the foot sensors. If there are no foot sensors, the foot contact state is determined by the timing of the support phase and the swing phase in a control cycle. If the leg is in the support phase, the foot of that leg is in contact with the ground, i.e., the leg state is in the support phase; if the leg is in the swing phase, the foot of that leg is not in contact with the ground, i.e., the leg state is in the swing phase; and the desired motion speed and gait parameters input from the outside.

[0046] The input to the centroid planner is the desired linear velocity of the quadruped robot. shaft and (axis direction), and the desired rotational angular velocity, and plan the desired center of mass position of the quadruped robot ( shaft and Axial direction, The axis also has an output, a fixed value, and is not required in this embodiment.

[0047] Step 2: For legs in the swing phase, determine whether the quadruped robot is subjected to an external force disturbance greater than the disturbance threshold based on the current position of the center of mass in the world coordinate system, the current expected position of the center of mass, and the previous expected position of the center of mass. If the quadruped robot is subjected to an external force disturbance greater than the disturbance threshold, the current footing position of the leg is determined by the center of mass follow-up method; otherwise, the current footing position of the leg is determined by the expected center of mass position compensation method.

[0048] Swing phase landing point position compensation: When the quadruped robot is performing translational movements in extreme operating scenarios (such as fixed-point operation in high-risk areas, passage through narrow and confined spaces, and high-precision attitude maintenance operations), the traditional heuristic landing point planning method needs to be optimized. The logic of landing point following the center of mass should be abandoned, and the landing point should be fixed at the preset position first. Only when the disturbance exceeds the redundancy tolerance range of the body should the follow mode be temporarily switched to ensure the safety of the whole machine.

[0049] When the quadruped robot is translating, according to the heuristic foot placement planning method, the foot placement follows the center of mass and is affected by user input. Specifically, based on the current position of the center of mass in the world coordinate system, the velocity of the center of mass, the expected velocity of the center of mass, and the phase of the leg swing at the current moment, the foot placement position of the leg at the current moment is calculated according to the heuristic foot placement planning method, as shown in formula (1):

[0050] (1)

[0051] In the formula, For the first The position of the foot of one leg in the world coordinate system; This represents the current position of the center of mass in the world coordinate system. When rotating in place, the radius of rotation of the neutral foot point around the center of the robot represents the radius of the circular motion of the foot tip on the plane, and is a fixed value; During rotation, the line connecting the target landing point and the fuselage centerline is drawn to the world coordinate system. Axis in The included angle of the plane represents the azimuth angle of the foot on the plane, and is a real-time changing value; For the present The velocity of the center of mass in the direction; The current phase of the leg swing is 0 when the swing begins and 1 when the swing ends, which is determined by the swing phase. These are the ground contact time and swing time within one gait cycle, respectively, and are fixed values. They are respectively The directional velocity deviation gain is a given fixed value; They are respectively The desired velocity of the centroid in the direction is the input value.

[0052] As shown in formula (1), the landing point of the quadruped robot is dynamic relative to the body. When subjected to external impact, the center of mass deviates within the tolerance range of the body, and the landing point also changes, leading to frequent modifications of the landing point. Therefore, this embodiment additionally introduces a method of compensation for the desired center of mass position to determine the current landing point position of the leg. Specifically, based on the radius and azimuth of the circular motion of the leg foot on the plane, the landing point of the leg in the world coordinate system is determined. shaft and Projection components along the axes; using the world coordinate system shaft and The desired centroid position along the axis, respectively, for the leg footing point in the world coordinate system. shaft and Compensation is performed on the projection components along the axis to obtain the current foot placement point of the leg in the world coordinate system. shaft and Position along the axis.

[0053] Furthermore, to enable flexible switching between the two methods, this embodiment adds a corresponding... shaft and Shaft identification criteria The desired centroid position obtained based on the centroid planner Thus, we obtain formula (2):

[0054] (2)

[0055] in, The judgment flag is used to determine whether the expected trajectory in the current direction has changed. Specifically, if the expected centroid position at the current moment is equal to the expected centroid position at the previous moment, and the absolute value of the difference between the current centroid position in the world coordinate system and the expected centroid position at the current moment is less than the disturbance threshold, then it indicates that the quadruped robot has not been subjected to an external force disturbance greater than the disturbance threshold. All values ​​are 0; otherwise, it indicates that the quadruped robot is subjected to an external force disturbance greater than the disturbance threshold, and the threshold is set. All are 1.

[0056] by Taking direction as an example: when the lateral velocity is input, the desired... Direction and position change in real time, flag position The landing point enters a follow-up state, dynamically adjusting with the trajectory of the center of mass to achieve lateral movement of the entire machine; when no lateral velocity is input, the desired... The orientation and position remain constant, and the marker position remains constant. In this direction, the aircraft enters a non-follow-up state (landing point compensation is performed to prevent it from changing with the center of mass). At this time, the landing point position is determined by the desired trajectory. Lateral disturbances are offset by the redundancy of the fuselage on the torso side, so there is no need to frequently modify the landing point.

[0057] When subjected to lateral external force disturbance, the landing point will be compensated, achieving a non-following center of mass effect. However, when the disturbance force is too large and the fuselage attitude redundancy cannot completely offset the disturbance, it is still necessary to switch to follow mode to maintain stability. In this embodiment, the judgment flag is used. The assignment of is expressed by the following formula:

[0058] (3)

[0059] (4)

[0060] in The expected center of mass position at the previous moment. The threshold value is set to half the fuselage width. When the center of mass offset exceeds the threshold, the flag is automatically set to 1, and the landing point resumes its follow-up state. The large disturbance is offset by adjusting the landing point position. When the offset returns to the threshold range, the flag is reset to 0, the landing point receives compensation, and the attitude redundancy anti-interference mode is returned.

[0061] Finally, based on the current foot landing point, the foot trajectory is planned using Bézier curves to obtain the continuous foot position within each control cycle. .

[0062] Step 3: For the leg in the support phase, an improved EID estimator is used to estimate the external disturbance, and the estimated external disturbance value is applied to the foot contact force of the leg to compensate for the external disturbance. The compensated foot contact force is obtained by replacing the generalized inverse substitution of the B matrix in the EID estimator with a quadratic programming method.

[0063] Support phase foot force compensation, building upon compensation for the oscillating phase, enhances the quadruped robot's anti-disturbance capability and ensures stable body posture. When subjected to external disturbances (such as ground protrusions or lateral impacts), to maintain the fixed trajectory of the foot landing point, the support phase foot force compensation employs an improved EID estimator to estimate the disturbance magnitude. The disturbance is then applied equivalently to the foot force for compensation. Foot force adjustment counteracts the disturbance, improving anti-interference capability. Details are as follows:

[0064] First, the quadruped robot system is a high-degree-of-freedom system. In order to simplify the control system design, the quadruped robot system needs to be simplified to obtain a three-dimensional single rigid body model, as shown in expression (5):

[0065] (5)

[0066] In the formula, Let this be the position of the center of mass of the quadruped robot in the world coordinate system; The attitude angle of the quadruped robot body relative to the world system; The angular velocity of the quadruped robot as measured by its inertial guidance unit (IMU); for The first derivative, for The second derivative, for The first derivative, for The first derivative, For the mass of the quadruped robot; For foot contact force, These correspond to the indices of the four legs of the quadruped robot; This is the vector of gravitational acceleration; To calculate the angular velocity in the body coordinate system, Euler angle derivative converted to world coordinates The transformation matrix; The inertial tensor in the world coordinate system; This represents the total torque acting on the center of mass of the quadruped machine.

[0067] The state-space model is further constructed from the three-dimensional single rigid body model, as shown in expression (6):

[0068] (6)

[0069] in, for The system state at any given moment (including attitude angle, center of mass position, center of mass velocity, body angular velocity, and gravitational acceleration). System status The derivative, for The timing controller input is the contact force at the tips of the four legs of the quadruped robot. , It is obtained by linearizing the rigid body dynamics equations at the equilibrium point; for The system output at any given time, This is the output matrix, used to obtain information from the system state. The angular velocity and center-of-mass velocity components of the body are extracted as the system output.

[0070] like Figure 3 As shown, based on the system state, controller input, and observation output, the observation residual is defined and the state observer is designed, as shown in expression (7):

[0071] (7)

[0072] in, for System state estimation at time t. For system state estimation The derivative, Define the output and residual for the observer gain. ,in For observational output, it represents the angular velocity of the fuselage. With the velocity of the center of mass .

[0073] Because the forces at the ends of the legs of a quadruped robot need to be constrained by factors such as friction cones, normal forces, and zero torque of the swinging legs during movement, the traditional EID (Enhanced Intelligent Displacement) method cannot be directly applied. The interference is equivalent to the input channel using a matrix (the generalized inverse of the B matrix). In this invention, a quadratic programming (QP) approach is used to solve for the constrained compensation amount. First, the observation residuals are... Amplified by the observer gain, and then multiplied by a transformation matrix. The dimensions of the observed residuals are converted into forces and moments for subsequent QP optimization calculations, ultimately yielding the perturbation force acting on the center of mass. See expression (8):

[0074] (8)

[0075] in, To control the cycle; Represents a 3×3 identity matrix; Represents a 3×3 zero matrix; For observer gain; The transformation matrix is ​​given. According to the center-of-mass dynamics, the plantar force, the resultant force of the center of mass, and the resultant torque satisfy a mapping relationship (i.e., the mapping relationship between the foot compensation force and the disturbance force acting on the center of mass), as shown in expression (9):

[0076] (9)

[0077] Wherein, parameter matrix , For the first The position of each foot tip relative to the center of mass It is a cross product matrix; The foot compensation force is determined by constructing a QP problem based on the mapping relationship, followed by a quadratic programming problem, and finally solving the quadratic programming problem to obtain the foot compensation force. See expression (10):

[0078] (10)

[0079] in, For parameter matrices, for The compensating force at the toes at all times, for The disturbance force that constantly acts on the center of mass. The square of the L2 norm. Indexing the legs of a quadruped robot. This is the set of legs of the quadruped robot that are currently in a swinging phase. This is the set of legs of the quadruped robot in the current supporting phase. This represents the friction cone constraint condition. Represents the friction cone constraint function. Indicates the friction cone constraint threshold. , The upper and lower bounds of the normal force; Indicates in Time of the first The contact force at the foot of one leg. for Time of the first The compensating force at the foot of each leg, Indicates the contact force at the foot. In the normal component, For foot compensation force In the normal direction of the component.

[0080] Due to foot compensatory force The instantaneous changes may be significant, therefore a first-order filter is needed to smooth the result to the final compensation value. The design of the first-order filter is shown in expression (11):

[0081] (11)

[0082] in, is the filter time constant.

[0083] The final command issued by the supporting leg is as follows:

[0084] (12)

[0085] in For Model Predictive Control (MPC) controllers in The contact force at the foot is calculated at all times. Here, is the perturbation value equivalent to the input channel, and is the estimated external perturbation value. for The compensated foot contact force is calculated at all times.

[0086] In this embodiment, the swing phase is compensated in step 2 to enable the body and feet to move without following the movement during motion. In step 3, the improved EID compensation of the support phase enables better anti-interference capability when the landing point trajectory is fixed. The two work together: the swing phase determines the direction of movement of the quadruped robot, and the support phase performs balance control of the quadruped robot. Therefore, in the quadruped robot system, formula (12) is used to perform adaptive compensation based on the phase:

[0087] (13)

[0088] in for The foot measure of time; To support the phase duty cycle; The gait period; The foot contact force calculated by the MPC controller corresponds to ; These are estimates of external disturbances calculated using the improved EID estimator; The position of the foot after compensation. According to formula (12), when it is in the support phase, the foot amount is the compensated foot ground return force, and when it is in the swing phase, the foot amount is the compensated foot trajectory.

[0089] This embodiment uses the process of a quadruped robot experiencing lateral impacts during movement as an example for experimentation, as detailed below:

[0090] 1. This embodiment uses a quadruped robot and the control parameters are as follows: quadruped robot mass Moment of inertia of a quadruped robot The control period is 2ms, the gait period is 450ms, and the ground contact leg duty cycle is [not specified]. Filter time constant .

[0091] 2. In the embodiment, a lateral step thrust of 15N and 20N is used for lateral thrust, with each lateral thrust lasting for 0.2s, to simulate the situation where the quadruped robot is subjected to sudden disturbance in the working scenario.

[0092] 3. Because this invention makes the quadruped robot's feet and center of mass non-reactive, the primary evaluation target is the quadrilateral formed by the robot's feet. The trajectory of the projection center of the plane is determined by Figure 3 It can be seen that when subjected to lateral disturbance, the center of its quadrilateral projection... There was no significant deviation in direction. Secondly, after fixing the foot trajectory, a comparison was made between adding improved EID compensation and not adding EID compensation. Figure 4As shown, using the improved EID compensation, the centroid offset is smaller under the same thrust, proving that it has stronger resistance to lateral disturbances.

[0093] Example 2:

[0094] This embodiment provides a computer-readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it implements the steps of the quadruped robot gait adaptive control method based on active anti-interference described in Embodiment 1.

[0095] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0096] Example 3:

[0097] This embodiment provides a computer device, including a processor and a memory storing a plurality of computer instructions. When the computer instructions are executed by the processor, they implement the steps of the quadruped robot gait adaptive control method based on active anti-interference described in Embodiment 1.

[0098] The memory and processor are electrically connected directly or indirectly to enable data transmission or interaction. For example, these components can be electrically connected to each other via one or more communication buses or signal lines. The memory stores a computer program that can run on the processor, which implements the method of the present invention by running the computer program stored in the memory.

[0099] The memory may be, but is not limited to, Random Access Memory (RAM), Read Only Memory (ROM), Programmable Read-Only Memory (PROM), Erasable Programmable Read-Only Memory (EPROM), Electrically Erasable Programmable Read-Only Memory (EEPROM), etc. The memory stores the program, and the processor executes the program upon receiving an execution instruction.

[0100] The processor may be an integrated circuit chip with data processing capabilities. The aforementioned processor can be a general-purpose processor, including a Central Processing Unit (CPU), a Network Processor (NP), etc. It can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this invention. The general-purpose processor can be a microprocessor or any conventional processor.

[0101] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0102] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the appended claims.

Claims

1. A gait adaptive control method for a quadruped robot based on active anti-interference, characterized in that, Includes the following steps: Collect the position of the quadruped robot's center of mass in the world coordinate system and the state of its legs at the current moment, and use the center of mass planner to obtain the expected position of the center of mass at the current moment; For legs in the swing phase, based on the current position of the centroid in the world coordinate system, the current expected position of the centroid, and the previous expected position of the centroid, it is determined whether the quadruped robot is subjected to an external force disturbance greater than the disturbance threshold. If the quadruped robot is subjected to an external force disturbance greater than the disturbance threshold, the current footing position of the legs is determined by the center of mass following method; otherwise, the current footing position of the legs is determined by the expected center of mass position compensation method. For legs in the support phase, an improved equivalent input perturbation estimator is used to estimate external perturbations, and the estimated external perturbation values ​​are applied to the foot contact force of the leg to compensate for the external perturbation. The compensated foot contact force is obtained by replacing the generalized inverse substitution of the B matrix in the equivalent input perturbation estimator with a quadratic programming method. The method of estimating external disturbances using an improved equivalent input disturbance estimator includes: Construct a three-dimensional single rigid body model of the quadruped robot and obtain its state space model. Define the observation residuals and design the state observer based on the system state, controller inputs, and observation outputs; The observation residual is amplified by the state observer gain, and the dimensions of the amplified observation residual are converted into forces and torques by the transformation matrix to obtain the perturbation force acting on the center of mass. Based on the dynamics of the center of mass, a mapping relationship is established between the foot compensation force and the disturbance force acting on the center of mass. Based on the mapping relationship, a quadratic programming problem is constructed, and the foot compensation force is obtained by solving the quadratic programming problem. The foot-end compensation force is passed through a filter, and the external disturbance estimate is obtained based on the filter output; wherein, the construction of the quadratic programming problem is as follows: ； In the formula, For parameter matrices, for The compensating force at the toes at all times, for The disturbance force that constantly acts on the center of mass. Indexing the legs of a quadruped robot. This is the set of legs of the quadruped robot that are currently in a swinging phase. This is the set of legs of the quadruped robot in the current supporting phase. This represents the friction cone constraint condition. Represents the friction cone constraint function. Indicates the friction cone constraint threshold. , The upper and lower bounds of the normal force; Indicates in Time of the first The contact force at the foot of one leg. for Time of the first The compensating force at the foot of each leg, Indicates the contact force at the foot. In the normal component, For foot compensation force In the normal direction of the component.

2. The quadruped robot gait adaptive control method based on active anti-interference as described in claim 1, characterized in that, The step of determining whether the quadruped robot is subjected to an external force disturbance greater than the disturbance threshold based on the current position of the centroid in the world coordinate system, the expected position of the centroid at the current moment, and the expected position of the centroid at the previous moment includes: If the expected centroid position at the current moment is equal to the expected centroid position at the previous moment, and the absolute value of the difference between the current centroid position in the world coordinate system and the expected centroid position at the current moment is less than the disturbance threshold, then the quadruped robot is not subjected to an external force disturbance greater than the disturbance threshold; otherwise, it means that the quadruped robot is subjected to an external force disturbance greater than the disturbance threshold.

3. The gait adaptive control method for a quadruped robot based on active anti-interference as described in claim 1, characterized in that, The method of determining the current foot placement position of the leg using a center-of-mass follower approach includes: Based on the current position of the center of mass in the world coordinate system, the velocity of the center of mass, the expected velocity of the center of mass, and the phase of the leg swing at the current moment, the position of the leg landing point at the current moment is calculated using a heuristic landing point planning method.

4. The gait adaptive control method for a quadruped robot based on active anti-interference as described in claim 1, characterized in that, The method of determining the current foot placement position of the leg using the desired centroid position compensation method includes: Based on the radius and azimuth of the circular motion of the leg and foot on the plane, determine the position of the leg's footing point in the world coordinate system. shaft and Projection components along the axial direction; Using the world coordinate system shaft and The desired centroid position along the axis, respectively, for the leg footing point in the world coordinate system. shaft and Compensation is performed on the projection components along the axis to obtain the current foot placement point of the leg in the world coordinate system. shaft and Position along the axis.

5. The gait adaptive control method for a quadruped robot based on active anti-interference as described in claim 1, characterized in that, Based on the current position of the foot landing point, the foot trajectory is planned using Bézier curves to obtain the continuous foot position within each control cycle.

6. The gait adaptive control method for a quadruped robot based on active anti-interference as described in claim 1, characterized in that, The transformation matrix contains two rows and two columns of elements. The elements in the first row and first column are the ratio of the inertial tensor to the control period in the world coordinate system of the quadruped robot. The elements in the first row and second column and the second row and first column are both zero matrices. The elements in the second row and second column are the ratio of the product of the mass of the quadruped robot and the identity matrix to the control period.

Images