Dual-arm dexterous hand hybrid control method and system

Abstract

The application discloses a kind of dual-arm dexterous hand hybrid control method and system, method includes: according to the size and position of fingertip force Calculation grasping force center position;With the line of dual-arm center reference point and the line of grasping force center position Construction virtual constraint line segment, with line segment length as geometric offset;Tracking virtual constraint line segment posture change, extract direction vector rotation angular velocity as structure constraint relaxation rate;With geometric offset and relaxation rate as horizontal and vertical coordinates Construction state trajectory, extract trajectory tortuosity, determine that system is in tight coupling or loose coupling state;According to coupling state adjustment position control gain and force control gain, generate hybrid control instruction Output to dual-arm joint controller and dexterous hand finger controller executes.By virtual constraint line segment, the degree of structure coupling between dual-arm and dexterous hand is perceived, the smooth mixing of position and force control is realized, the fixed threshold switching impact is avoided, and the grasping adaptability and stability are improved.

Description

Technical Field

[0001] This invention belongs to the field of robot control technology, and in particular relates to a hybrid control method and system for dual-arm dexterous hands. Background Technology

[0002] A dual-arm dexterous hand system consists of two multi-degree-of-freedom robotic arms and a multi-finger dexterous hand mounted at the end of each arm, capable of performing complex grasping, manipulation, and assembly tasks. In this type of system, the two arms provide support and positioning for the dexterous hand, which performs the grasping and manipulation of objects; there is a close mechanical coupling between the two.

[0003] Existing dual-arm dexterity hand hybrid control methods mainly suffer from the following technical defects: First, the switching between position control and force control relies on a fixed threshold. Existing methods determine the switching control mode by detecting whether the force amplitude exceeds a preset threshold or whether the position error exceeds a preset threshold. This approach has poor adaptability to different operational tasks and objects, and the threshold setting lacks a physical basis.

[0004] Second, there is a lack of awareness of the constraint relationship between the two arms and the dexterous hand. Existing methods treat the control of the two arms and the dexterous hand as two independent or loosely coupled subsystems, without establishing a geometric constraint model between the two arms and the dexterous hand, and therefore cannot perceive the degree of structural coupling between them.

[0005] Third, the control mode switching is abrupt. Existing methods directly switch control modes when switching conditions are detected, causing a sudden change in control output at the moment of switching, resulting in shocks and vibrations in the robotic arm and dexterous hand.

[0006] Therefore, a method is needed that can establish a geometric constraint model between the two arms and the dexterous hand, and guide the switching of hybrid control modes through constraint state perception. Summary of the Invention

[0007] This invention provides a dual-arm dexterity hand hybrid control method and system to solve at least one of the above-mentioned technical problems.

[0008] In a first aspect, the present invention provides a method for hybrid control of dual-arm dexterity hands, comprising: Acquire the end-position signals of both arms in space and the fingertip force signals of each finger of the dexterous hand; The midpoint of the line connecting the ends of the arms is calculated based on the position of the ends of the arms and used as the center reference point of the arms. The center position of the grasping force is calculated based on the magnitude of the fingertip force and the spatial position of the fingertips. A virtual constraint line segment is constructed between the two dexterous arms by connecting the reference point of the two arm centers and the position of the grasping force center. The length of the virtual constraint line segment is used as the geometric offset between the center of the two arms and the center of the grasping force. Track the attitude change of the virtual constraint line segment in space, and extract the rotational angular velocity of the direction vector of the virtual constraint line segment on the time axis as the structural constraint relaxation rate of the dual-arm dexterous hand system; Determine whether the dual-arm dexterous hand system is in a tightly coupled state based on the geometric offset and the structural constraint relaxation rate. If the dual-arm dexterous hand system is in a tightly coupled state, the position control gain and force control gain are adjusted, and a hybrid control command is generated based on the adjusted position control gain and force control gain, which is then output to the dual-arm joint controller and the dexterous hand finger controller for execution.

[0009] In a second aspect, the present invention provides a dual-arm dexterity hand hybrid control system, comprising: The acquisition module is configured to acquire the end position signals of the two arms in space and the fingertip force signals of each finger of the dexterous hand; The calculation module is configured to calculate the midpoint of the line connecting the ends of the two arms based on the end positions of the two arms, and use it as the center reference point of the two arms. It also calculates the center position of the grasping force based on the magnitude of the fingertip force and the spatial position of the fingertip. The construction module is configured to construct a virtual constraint line segment between the two dexterous arms using the line connecting the center reference point of the two arms and the position of the center of the grasping force, and to use the length of the virtual constraint line segment as the geometric offset between the center of the two arms and the center of the grasping force. The extraction module is configured to track the attitude change of the virtual constraint line segment in space and extract the rotational angular velocity of the direction vector of the virtual constraint line segment on the time axis as the structural constraint relaxation rate of the dual-arm dexterous hand system. The judgment module is configured to determine whether the dual-arm dexterous hand system is in a tightly coupled state based on the geometric offset and the structural constraint relaxation rate. The output module is configured to adjust the position control gain and force control gain if the dual-arm dexterous hand system is in a tightly coupled state, and generate a hybrid control command based on the adjusted position control gain and force control gain, and output it to the dual-arm joint controller and dexterous hand finger controller for execution.

[0010] Thirdly, an electronic device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the steps of the dual-arm dexterity hand hybrid control method of any embodiment of the present invention.

[0011] Fourthly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the program is executed by a processor, the processor performs the steps of the dual-arm dexterity hand hybrid control method of any embodiment of the present invention.

[0012] The dual-arm dexterity hand hybrid control method and system of this application acquires the position signals of the ends of both arms and the fingertip force signals of each finger of the dexterity hand. It calculates the midpoint of the line connecting the ends of both arms as the center reference point of both arms and calculates the position of the grasping force center based on the magnitude of the fingertip force and the spatial position of the fingertips. Then, it constructs a virtual constraint line segment using the line connecting the two, and uses its length as the geometric offset. This unifies the position information of both arms and the force information of the dexterity hand at the spatial geometric level, enabling the system to directly perceive the degree of structural coupling between the arms and the dexterity hand. Furthermore, it constructs a constraint triangle using the ends of the left and right arms and the grasping force center. The absolute value of the difference in the areas of the two sub-triangles is used as the asymmetry, and the difference in the interior angles at the center reference point of both arms deviating from the preset reference angle is used as the structural angular deviation. The product of these two values ​​is used as the geometric offset, thus comprehensively characterizing the degree and direction of geometric constraint distortion. By tracking the rotational angular velocity of the direction vector of the virtual constraint line segment, it calculates the instantaneous rotation angle, rotational angular acceleration, and rotation direction holding time. The ratio of the instantaneous rotation angle to the rotation direction holding time is used as the structural constraint relaxation rate, thus simultaneously reflecting constraint changes. The system assesses the speed and directional stability of the coupling. A state trajectory is constructed using geometric offset as the x-axis and structural constraint relaxation rate as the y-axis. The displacement vector sequence and directional change angle sequence of adjacent state points are calculated. The ratio of cumulative directional change to total displacement is used as the tortuosity and compared with a threshold, thus avoiding the subjectivity of setting separate thresholds and accurately determining tight or loose coupling states. Based on this, a coupling state prediction factor is obtained by measuring the proportion and rate of change of tight coupling within a preset time window. When the factor is greater than the first threshold, the force control gain is increased and the position control gain is decreased; when the factor is less than the second threshold, the position control gain is increased and the force control gain is decreased, thereby adjusting the gain in advance and overcoming hysteresis. Finally, the position control component is obtained by multiplying the end-position error of both arms with the adjusted position control gain, and the force control component is obtained by multiplying the dexterous fingertip force error with the adjusted force control gain. The two are added together to generate a hybrid control command, which is output to the dual-arm joint controller and the dexterous hand finger controller, achieving a smooth blend of position and force control, reducing switching shocks and oscillations, and improving the accuracy of coupling state recognition, grasping stability, and task adaptability. Attached Figure Description

[0013] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0014] Figure 1 A flowchart of a dual-arm dexterity hand hybrid control method provided in an embodiment of the present invention; Figure 2 This is a structural block diagram of a dual-arm dexterity hand hybrid control system provided in an embodiment of the present invention; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, 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, 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.

[0016] Please see Figure 1 The diagram shows a flowchart of a dual-arm dexterity hand hybrid control method according to this application.

[0017] like Figure 1 As shown, the dual-arm dexterity hand hybrid control method specifically includes the following steps: Step S101: Obtain the end position signals of the two arms in space and the fingertip force signals of each finger of the dexterous hand.

[0018] In this step, encoders at each joint of both arms collect joint angle values, and forward kinematics calculations are used to obtain the position coordinates of the left and right arm distal ends in three-dimensional space, serving as the distal end position signals of both arms. Miniature force sensors installed at the fingertips of each finger of the dexterous hand collect the contact force between each finger and the object, obtaining the fingertip force signals of each finger. These fingertip force signals contain information on the magnitude and direction of the force at each fingertip. The distal end position signals and the fingertip force signals are synchronously acquired at the same sampling frequency and aligned on the time axis.

[0019] Step S102: Calculate the midpoint of the line connecting the ends of the two arms based on the end positions of the two arms, and use it as the center reference point of the two arms. Calculate the center position of the grasping force based on the magnitude of the fingertip force and the spatial position of the fingertips.

[0020] In this step, the first end position coordinates of the left arm end in space and the second end position coordinates of the right arm end in space are obtained. The line connecting the first end position coordinates and the second end position coordinates is used as the inter-arm reference baseline. The intersection of the perpendicular plane of the inter-arm reference baseline and the inter-arm reference baseline is used as the center reference point of the two arms. The magnitude of the fingertip force of each finger of the dexterous hand and the spatial coordinates of the fingertip position of each finger are obtained. Using the magnitude of the fingertip force of each finger as the force value, the resultant force components of the fingertip force in three orthogonal directions in space are calculated. The spatial coordinates of each resultant force component when they reach equilibrium are calculated respectively. The average value of the spatial coordinates of each resultant force component when they reach equilibrium is taken as the position of the grasping force center.

[0021] In one specific embodiment, the first spatial coordinates of the left arm's distal end and the second spatial coordinates of the right arm's distal end are first obtained. Both the first and second distal end coordinates are three-dimensional coordinate vectors, containing X-axis, Y-axis, and Z-axis components. The line connecting the first and second distal end coordinates is used as an inter-arm reference baseline, with the starting point at the left arm's distal end and the ending point at the right arm's distal end.

[0022] Using the midpoint of the inter-arm reference baseline as a reference, a plane perpendicular to the inter-arm reference baseline is drawn through the midpoint; this plane is the perpendicular bisector of the inter-arm reference baseline. The intersection of this perpendicular bisector and the inter-arm reference baseline is the midpoint of the inter-arm reference baseline. The intersection of this perpendicular bisector and the inter-arm reference baseline is taken as the center reference point of both arms. The three-dimensional spatial coordinates of the center reference point of both arms are equal to the sum of the coordinates of the first end position and the coordinates of the second end position, divided by two.

[0023] The algorithm acquires the fingertip force magnitude and fingertip spatial coordinates of each finger in a dexterous hand. The fingertip force magnitude is the force value measured by the fingertip force sensor, and the fingertip spatial coordinates are the position coordinates of each fingertip in three-dimensional space. The fingertip spatial coordinates are calculated by the dexterous hand's forward kinematic model based on the angles of each finger joint.

[0024] Using the magnitude of the fingertip force as a weighting coefficient and the spatial coordinates of the fingertip as the weighting object, a weighted average of the fingertip positions is calculated. The weighted average is calculated as follows: multiply the magnitude of the fingertip force of each finger by each component of the corresponding fingertip spatial coordinate to obtain the weighted value of each finger in each component; sum the weighted values ​​of all fingers in the same component and divide by the sum of the magnitudes of the fingertip forces of all fingers to obtain the weighted average coordinates for that component. The weighted average coordinates for the X-axis, Y-axis, and Z-axis components are calculated separately, and the weighted average coordinates of the three components are combined to form a coordinate point in three-dimensional space, which serves as the center position of the grasping force.

[0025] When the fingertip force of a particular finger is less than a preset invalid fingertip force threshold, the contribution of that finger's contact force to the grasping force center is negligible. This particular finger is excluded from the calculation of the grasping force center position; that is, the fingertip force magnitude and fingertip spatial coordinates are not included in the weighted average calculation. The preset invalid fingertip force threshold is determined based on the noise level of the force sensor and the contact force detection sensitivity. Specifically, the outputs of each fingertip force sensor are collected under dexterity hand idle conditions, the noise standard deviation σ is calculated, and the preset invalid fingertip force threshold is determined. , This indicates the full-scale or range of the force sensor, and .

[0026] Step S103: Construct a virtual constraint line segment between the two dexterous arms using the line connecting the center reference point of the two arms and the center position of the grasping force, and use the length of the virtual constraint line segment as the geometric offset between the center of the two arms and the center of the grasping force.

[0027] In this step, the positions of the left and right arm ends and the center of the grasping force are obtained, and a dual-arm dexterity hand constraint triangle is constructed using the positions of the left and right arm ends and the center of the grasping force as the three vertices. The midline from the center reference point of the two arms to the center of the grasping force in the dual-arm dexterity hand constraint triangle is used as a virtual constraint line segment. The virtual constraint line segment divides the constraint triangle into two sub-triangles. The absolute value of the difference between the areas of the two sub-triangles is used as the asymmetry of the dual-arm dexterity hand constraint triangle. The interior angle of the constraint triangle of the dual-arm dexterity hand at the reference point at the center of the dual arms is obtained, and the difference between the interior angle and the preset reference angle is used as the structural angle deviation of the dual-arm dexterity hand system. Multiply the structural angular deviation by the asymmetry of the dual-arm dexterity hand constraint triangle, and use the product as the geometric offset between the center of the dual arms and the center of the grasping force.

[0028] In one specific embodiment, the positions of the left and right arm ends collected in step S101 and the position of the gripping force center calculated in step S102 are first obtained. The positions of the left and right arm ends and the gripping force center are all coordinate points in three-dimensional space.

[0029] Using the positions of the left and right arm ends and the center of the grasping force as three vertices, a triangle is constructed in three-dimensional space, called the dual-arm dexterity hand constraint triangle. The three vertices of the dual-arm dexterity hand constraint triangle correspond to the two support points of the arms and the center of the dexterity hand's grasping force, respectively. The dual-arm dexterity hand constraint triangle consists of three sides and three interior angles. The three sides are the lines connecting the ends of the left and right arms, the left and right arms to the center of the grasping force, and the right and right arms to the center of the grasping force, respectively.

[0030] Obtain the bi-arm center reference point calculated in step S102. This bi-arm center reference point is located at the midpoint of the line connecting the ends of the left and right arms. Starting from this bi-arm center reference point and ending at the center of the grasping force, draw a line segment from the bi-arm center reference point to the center of the grasping force. This line segment is the midline of the constraint triangle of the bi-arm dexterity hand. Use this midline as the virtual constraint line segment between the bi-arm dexterity hands.

[0031] The virtual constraint line segment divides the dual-arm dexterity hand constraint triangle into two sub-triangles: the first sub-triangle is composed of three vertices: the end of the left arm, the reference point at the center of both arms, and the center of the grasping force; the second sub-triangle is composed of three vertices: the end of the right arm, the reference point at the center of both arms, and the center of the grasping force. The areas of the first and second sub-triangles are calculated respectively, and the absolute value of the difference between the areas of the two sub-triangles is taken as the asymmetry of the dual-arm dexterity hand constraint triangle. The asymmetry characterizes the degree of geometric asymmetry between the two support points of the arms and the center of the grasping force. When the left and right arms are symmetrically distributed relative to the center of the grasping force, the areas of the two sub-triangles are equal, and the asymmetry is zero; when either the left or right arm is offset relative to the center of the grasping force, the areas of the two sub-triangles are not equal, and the asymmetry is positive.

[0032] Obtain the interior angle of the constraint triangle of the dual-arm dexterity hand at the reference point at the center of the dual arms. The interior angle is the angle between the line connecting the end of the left arm to the reference point at the center of the dual arms and the line connecting the end of the right arm to the reference point at the center of the dual arms. Obtain a preset reference angle, which is determined based on the ideal structure when the dual arms are symmetrically supported. Specifically, in a standard symmetrical grasping posture, the ends of the left and right arms and the center of the grasping force form an isosceles triangle with the reference point at the center of the dual arms as its vertex, and the interior angle at this vertex is measured as the reference angle; or, based on the dual-arm workspace and grasping task, the preset reference angle is calibrated experimentally, ranging from 60 degrees to 120 degrees, preferably 90 degrees. When the actual interior angle is equal to this reference angle, the structural angular deviation is zero. Calculate the difference between the interior angle and the preset reference angle, and use this difference as the structural angular deviation of the dual-arm dexterity hand system. The structural opening angle deviation characterizes the degree to which the opening of the line connecting the ends of the two arms deviates from the ideal state relative to the central reference point of the two arms. When the actual opening angle of the ends of the two arms at the central reference point of the two arms is equal to the preset reference angle, the opening angle deviation is zero; when the actual opening angle is greater than or less than the preset reference angle, the opening angle deviation is a positive or negative value.

[0033] The structural angular deviation is multiplied by the asymmetry of the dual-arm dexterity hand constraint triangle, and the product is used as the geometric offset between the center of the two arms and the center of the grasping force. The geometric offset can be positive or negative; its absolute value characterizes the degree of distortion in the geometric constraints of the dual-arm dexterity hand system, and its sign indicates the direction of the distortion. When both the structural angular deviation and the asymmetry are positive or both are negative, the geometric offset is positive, indicating a double offset in the same direction between the geometric constraints of the dual-arm support structure and the center of the grasping force; when their signs are opposite, the geometric offset is negative, indicating that the two constraint distortions cancel each other out in different directions. The larger the absolute value of the geometric offset, the more severe the deviation of the dual-arm dexterity hand system's geometric constraints from the ideal state in space, and the greater the control and adjustment required to compensate for the structural distortion.

[0034] Step S104: Track the attitude change of the virtual constraint line segment in space, and extract the rotational angular velocity of the direction vector of the virtual constraint line segment on the time axis as the structural constraint relaxation rate of the dual-arm dexterous hand system.

[0035] In this step, the first direction vector of the virtual constraint line segment at the current moment, the second direction vector of the virtual constraint line segment at the previous moment, and the third direction vector of the virtual constraint line segment at the previous two moments are obtained; Calculate the first angle between the first direction vector and the second direction vector as the instantaneous rotation angle at the current moment; calculate the second angle between the second direction vector and the third direction vector as the instantaneous rotation angle at the previous moment. The rotational angular acceleration at the current moment is obtained by dividing the difference between the first included angle and the second included angle by the time interval between the current moment and the previous moment. The rotation direction of the direction vector is determined based on the sign change of the rotation angular acceleration. The duration during which the rotation direction does not reverse within a preset number of consecutive moments is taken as the duration during which the rotation direction of the direction vector is maintained. The ratio of the instantaneous rotation angle at the current moment to the duration of the rotation direction is used as the structural constraint relaxation rate of the dual-arm dexterity hand system.

[0036] In one specific embodiment, the dual-arm dexterous hand system refers to a robot operation platform composed of two multi-degree-of-freedom robotic arms and a multi-finger dexterous hand installed at the end of the robotic arms. The two arms are mainly used for support, positioning and movement, while the dexterous hand is mainly used for grasping, manipulating and assembling objects.

[0037] First, obtain the first direction vector of the virtual constraint line segment at the current moment, the second direction vector of the virtual constraint line segment at the previous moment, and the third direction vector of the virtual constraint line segment at the two moments before. The direction vector is a unit vector pointing from the center reference point of the two arms to the center of the grasping force, representing the orientation of the virtual constraint line segment in space. The first direction vector is the direction of the virtual constraint line segment at the current moment from the center reference point of the two arms to the center of the grasping force; the second direction vector is the direction of the virtual constraint line segment at the previous moment; and the third direction vector is the direction of the virtual constraint line segment at the two moments before.

[0038] Calculate the angle between the first direction vector and the second direction vector, and use this angle as the instantaneous rotation angle at the current moment. The angle is calculated as follows: calculate the dot product of the first direction vector and the second direction vector, divide by the product of the magnitudes of the first and second direction vectors, and take the inverse cosine of the result to obtain the radian value of the angle. Both the first and second direction vectors are unit vectors with a magnitude of 1. The instantaneous rotation angle represents the angular displacement of the virtual constraint line segment in space from the previous moment to the current moment.

[0039] Calculate the angle between the second direction vector and the third direction vector, and use this angle as the instantaneous rotation angle at the previous moment. The method for calculating the instantaneous rotation angle at the previous moment is the same as the method for calculating the instantaneous rotation angle at the current moment, and it represents the angular displacement of the virtual constraint line segment in space from the previous two moments to the previous moment.

[0040] The instantaneous rotation angle at the current moment is subtracted from the instantaneous rotation angle at the previous moment. The difference obtained by subtraction is divided by the time interval between the current moment and the previous moment. The quotient is taken as the rotational angular acceleration at the current moment. The rotational angular acceleration represents the rate of change of the rotational speed of the virtual constraint line segment. When the rotational angular acceleration is positive, it indicates that the rotation of the virtual constraint line segment is accelerating; when the rotational angular acceleration is negative, it indicates that the rotation of the virtual constraint line segment is decelerating; when the rotational angular acceleration is zero, it indicates that the rotational speed of the virtual constraint line segment remains constant.

[0041] The direction of rotation of the direction vector is determined based on the sign change of the rotational angular acceleration. If the sign of the rotational angular acceleration at the current moment is opposite to that at the previous moment, the direction of rotation of the direction vector is determined to have reversed. If the sign of the rotational angular acceleration at the current moment is the same as that at the previous moment, the direction of rotation of the direction vector is determined not to have reversed. The rotational angular acceleration sign is recorded at each consecutive moment, and if no reversal occurs within a preset number of consecutive moments, the duration of this consecutive period without reversal is taken as the duration of the direction vector's rotation direction. The starting point of the duration of the rotation direction maintenance is the moment of the last reversal, and the ending point is the current moment. If no reversal occurs within the time interval from the previous two moments to the current moment, the duration of the rotation direction maintenance is equal to the time difference between the current moment and the moment of the last reversal.

[0042] The instantaneous rotation angle at the current moment is divided by the duration of the rotation direction being maintained, and the quotient is taken as the structural constraint relaxation rate of the dual-arm dexterous hand system. When the instantaneous rotation angle is large and the duration of the rotation direction being maintained is long, the value of the structural constraint relaxation rate is moderate, indicating that the system rotates stably in the same direction for a relatively long time. When the instantaneous rotation angle is small but the duration of the rotation direction being maintained is extremely short, the value of the structural constraint relaxation rate is large, indicating that the system's direction changes frequently but with small amplitude each time. When the instantaneous rotation angle is large and the duration of the rotation direction being maintained is extremely short, the value of the structural constraint relaxation rate is the largest, indicating that the system's direction undergoes a drastic and abrupt rotation. The structural constraint relaxation rate is related to both the amplitude and directional stability of the virtual constraint line segment's attitude change in space. The larger the instantaneous rotation angle and the shorter the duration of the rotation direction being maintained, the larger the structural constraint relaxation rate, indicating a higher degree of structural constraint relaxation between the dual arms and the dexterous hand.

[0043] Step S105: Determine whether the dual-arm dexterous hand system is in a tightly coupled state based on the geometric offset and the structural constraint relaxation rate.

[0044] In this step, the geometric offset is used as the horizontal axis and the structural constraint relaxation rate is used as the vertical axis. The geometric offset value and the structural constraint relaxation rate value at the current moment are combined into a state point in a two-dimensional plane. Obtain the state points corresponding to multiple moments within a preset time window before the current moment, and connect the multiple state points in chronological order to obtain the state trajectory of the state points in the two-dimensional plane. Calculate the displacement vector between two adjacent state points in the state trajectory to obtain a displacement vector sequence, and calculate the direction change angle between two adjacent displacement vectors in the displacement vector sequence to obtain a direction change angle sequence; The sum of the absolute values ​​of the directional change angles in the directional change angle sequence is taken as the cumulative directional change of the state trajectory, and the sum of the magnitudes of the displacement vectors in the displacement vector sequence is taken as the total displacement of the state trajectory. The ratio of the cumulative directional change to the total displacement is used as the tortuosity of the state trajectory, and it is determined whether the tortuosity of the state trajectory is less than a preset tortuosity threshold. When the degree of tortuosity is less than a preset tortuosity threshold, the dual-arm dexterity hand system is determined to be in a tightly coupled state. When the tortuosity is not less than the preset tortuosity threshold, the dual-arm dexterity hand system is determined to be in a loosely coupled state.

[0045] In one specific embodiment, the current value of the geometric offset calculated in step S103 and the current value of the structural constraint relaxation rate calculated in step S104 are first obtained. The geometric offset is a scalar representing the degree of geometric distortion between the center of the two arms and the center of the gripping force; the structural constraint relaxation rate is a scalar representing the rate of change of the virtual constraint line segment's attitude in space.

[0046] Using the geometric offset as the x-axis and the structural constraint relaxation rate as the y-axis, the geometric offset value and the structural constraint relaxation rate value at the current moment are combined to form a coordinate point in a two-dimensional plane, called a state point. The x-axis of the two-dimensional plane represents the geometric offset, and the y-axis represents the structural constraint relaxation rate. Multiple moments within a preset time window preceding the current moment are traversed to obtain the state point corresponding to each moment. The length of the preset time window is determined based on the dynamic response time of the dual-arm dexterous hand system, typically taken as 3 to 5 times the minimum response time of the system.

[0047] The multiple state points are connected sequentially in chronological order. Starting with the earliest state point within the preset time window, a straight line segment connects it to the state point at the next instant, then to the state point at the instant after that, and so on, until the current state point is reached. This results in the state trajectory of the state points in the two-dimensional plane. The state trajectory is a broken line in the two-dimensional plane.

[0048] Calculate the displacement vector between two adjacent state points in the state trajectory. For the displacement vector between the i-th state point and the (i+1)-th state point, its direction is determined by the position of the (i+1)-th state point relative to the i-th state point, and its magnitude is the Euclidean distance between the two state points. Arrange all the displacement vectors between adjacent state points in chronological order to obtain a displacement vector sequence.

[0049] Calculate the direction change angle between two adjacent displacement vectors in the displacement vector sequence. For the direction change angle between the i-th displacement vector and the (i+1)-th displacement vector, using the direction of the i-th displacement vector as the reference direction, calculate the rotation angle of the (i+1)-th displacement vector relative to the i-th displacement vector, and use this rotation angle as the direction change angle. Arrange the direction change angles between all adjacent displacement vectors in chronological order to obtain a direction change angle sequence.

[0050] The sum of the absolute values ​​of the directional change angles in the sequence of directional change angles is taken as the cumulative directional change of the state trajectory. The cumulative directional change represents the total magnitude of the directional change of the state trajectory during its movement; the greater the cumulative directional change, the greater the curvature of the state trajectory.

[0051] The sum of the magnitudes of the displacement vectors in the displacement vector sequence is taken as the total displacement of the state trajectory. The total displacement represents the total path length traversed by the state trajectory from the starting state point to the ending state point. The larger the total displacement, the greater the total distance the state trajectory moves in the two-dimensional plane.

[0052] The cumulative directional change is divided by the total displacement, and the quotient is taken as the tortuosity of the state trajectory. When the total displacement is zero, the tortuosity is set to zero. The tortuosity is a real number greater than or equal to zero. The smaller the tortuosity, the closer the state trajectory is to a straight line, that is, the more stable the relationship between the geometric offset and the structural constraint relaxation rate; the larger the tortuosity, the more curved the state trajectory is, that is, the more unstable the relationship between the two characteristic quantities.

[0053] The tortuosity is compared with a preset tortuosity threshold. The preset tortuosity threshold is determined based on the statistical distribution differences of the tortuosity of the state trajectory in tightly coupled and loosely coupled states of the dual-arm dexterous hand system. Specifically, the calibration method is as follows: multiple sets of tightly coupled and loosely coupled samples are collected respectively, and the tortuosity of each set of state trajectories is calculated according to the above steps to obtain the mean tortuosity under tightly coupled state. mean tortuosity under loose coupling state Take the preset tortuosity threshold. Alternatively, an ROC curve can be used to determine the threshold with the goal of maximizing classification accuracy. Tightly coupled samples can be obtained through rigid connection between the arms and the dexterous hand or through standard symmetrical grasping, while loosely coupled samples can be obtained through free movement or asymmetric perturbation.

[0054] When the tortuosity is less than the preset tortuosity threshold, the dual-arm dexterous hand system is determined to be in a tightly coupled state. In the tightly coupled state, the geometric constraints between the arms and the dexterous hand are stable, the posture changes of the virtual constraint line segments are gradual, and the state trajectory moves in an approximately straight line in the two-dimensional plane with minimal directional changes.

[0055] When the tortuosity is greater than or equal to the preset tortuosity threshold, the dual-arm dexterous hand system is determined to be in a loosely coupled state. In the loosely coupled state, the geometric constraints between the arms and the dexterous hand are unstable, the posture of the virtual constraint line segments changes drastically, and the state trajectory frequently changes direction in the two-dimensional plane with large directional changes.

[0056] Step S106: If the dual-arm dexterous hand system is in a tightly coupled state, the position control gain and force control gain are adjusted, and a hybrid control command is generated based on the adjusted position control gain and force control gain, and output to the dual-arm joint controller and dexterous hand finger controller for execution.

[0057] In this step, the tight coupling determination results of each time within a preset time window before the current time are obtained, and the number of determinations of tight coupling state within the preset time window is divided by the total number of determinations within the preset time window to obtain the tight coupling ratio. Calculate the change in the tight coupling ratio at the current moment relative to the tight coupling ratio at the previous moment, and obtain the ratio change rate. The product of the tight coupling ratio and the ratio change rate is used as the coupling situation prediction factor. When the coupling situation prediction factor is greater than the preset first factor threshold, the force control gain is increased by a first step length and the position control gain is decreased by a second step length. The first step length is the adjustment step length of the force control gain and the second step length is the adjustment step length of the position control gain. The first step length is greater than the second step length. When the coupling situation prediction factor is less than the preset second factor threshold, the position control gain is increased by the second step size, and the force control gain is decreased by the first step size. When the coupling situation prediction factor is greater than or equal to the second factor threshold and less than or equal to the first factor threshold, the position control gain and the force control gain remain unchanged. The deviation between the current position and the desired position of the two arms' ends is obtained as the position error of the two arms' ends; the deviation between the current fingertip force and the desired fingertip force of each finger of the dexterous hand is obtained as the dexterous fingertip force error. The position control component is obtained by multiplying the position error of the two arms end by the adjusted position control gain, and the force control component is obtained by multiplying the force error of the dexterous fingertips by the adjusted force control gain. The position control component and the force control component are added together, and the result of the addition is output as the hybrid control command.

[0058] In one specific embodiment, the tight coupling determination results for each moment within a preset time window prior to the current moment are first obtained. The tight coupling determination results are the determination results output in step S105, and each moment's determination result indicates either a tight coupling state or a loose coupling state. The length of the preset time window is determined based on the dynamic response time of the dual-arm dexterous hand system; specifically, the minimum system response time is set to... The sampling period is Then the window length , Take 3 to 5, and the number of sampling points within the window Take a value between 5 and 30. For example... , hour, The timeframe is set to 0.3 to 0.5 seconds. This range covers at least one dynamic response cycle while avoiding decision lag caused by an excessively long window.

[0059] The total number of times the tightly coupled state occurs within the preset time window is counted. This total number of occurrences is then divided by the total number of decisions made within the preset time window. The quotient is taken as the tight coupling percentage. The tight coupling percentage is a value between 0 and 1, representing the proportion of time the system is in a tightly coupled state within the preset time window.

[0060] Obtain the tight coupling percentage corresponding to the previous time step, subtract the tight coupling percentage of the previous time step from the tight coupling percentage of the current time step, and use the difference obtained from the subtraction operation as the percentage change rate. A positive percentage change rate indicates that the system is transitioning to a tight coupling state, a negative percentage indicates that the system is transitioning to a loose coupling state, and a zero percentage indicates that the coupling state of the system remains stable. The tight coupling percentage corresponding to the previous time step is the tight coupling percentage calculated with the previous time step as the end point of the window.

[0061] The tight coupling percentage is multiplied by the rate of change of that percentage, and the product is used as the coupling situation prediction factor. When both the tight coupling percentage and the rate of change of that percentage are positive, the coupling situation prediction factor is positive and relatively large, indicating that the system is currently in a tight coupling state and the degree of tight coupling is deepening, and the system will evolve towards a stronger tight coupling state. When both the tight coupling percentage and the rate of change of that percentage are negative, the coupling situation prediction factor is negative, indicating that the system is currently in a tight coupling state but the degree of tight coupling is weakening, and the system will evolve towards a loose coupling state. When the tight coupling percentage is low, regardless of whether the rate of change of that percentage is positive or negative, the absolute value of the coupling situation prediction factor is small, indicating that the system is currently in a loose coupling state and the change in the coupling state is not significant.

[0062] The coupling state prediction factor is compared with a preset first factor threshold and a preset second factor threshold. The first factor threshold is a positive number, and the second factor threshold is a negative number, wherein the first factor threshold is greater than the second factor threshold. The first factor threshold and the second factor threshold are determined according to the sensitivity requirements of the coupling state change of the dual-arm dexterity hand system. Typically, 20% to 30% of the maximum tight coupling ratio is taken as the first factor threshold, and 20% to 30% of the minimum tight coupling ratio is taken as the second factor threshold.

[0063] When the coupling situation prediction factor is greater than the first factor threshold, it is determined that the system is currently in a tightly coupled state and the degree of tight coupling is deepening. Therefore, it is necessary to strengthen the dominance of force control and weaken the dominance of position control. The current force control gain is increased by a first step, and the current position control gain is decreased by a second step. The first step is the adjustment step size of the force control gain, and the second step is the adjustment step size of the position control gain. The first step is greater than the second step to ensure that the adjustment range of the force control gain is greater than that of the position control gain, enabling the system to respond quickly to the deepening trend of tight coupling.

[0064] When the coupling situation prediction factor is less than the second factor threshold, it is determined that the system is currently in a loosely coupled state and the degree of loose coupling is deepening. Therefore, it is necessary to strengthen the dominance of position control and weaken the dominance of force control. The current position control gain is increased by the second step size, and the current force control gain is decreased by the first step size. Let the initial force control gain be... The initial position control gain is Then the first step is long Second step length .

[0065] When the coupling state prediction factor is greater than or equal to the second factor threshold and less than or equal to the first factor threshold, the coupling state of the system is determined to be in a stable range, and there is no need to adjust the control gain. The current position control gain and force control gain remain unchanged.

[0066] After completing gain adjustment or determining to keep the gain constant, the deviation between the current position and the desired position of the two arm ends in space is obtained as the arm end position error. The deviation between the current fingertip force and the desired fingertip force of each finger of the dexterous hand is obtained as the dexterous fingertip force error. The desired position is provided by the upper-level task planner, and the desired fingertip force is determined according to the operation task.

[0067] The position error at the ends of both arms is multiplied by the adjusted position control gain, and the product of the multiplication is used as the position control component. The force error at the dexterous fingertips is multiplied by the adjusted force control gain, and the product of the multiplication is used as the force control component. The position control component and the force control component are added together, and the sum of the addition is used as the hybrid control command.

[0068] The hybrid control commands are output to the left arm joint controller, the right arm joint controller, and the finger controllers of each finger of the dexterous hand, so that the arms and the dexterous hand can perform movements according to the hybrid control commands.

[0069] In summary, the method of this application acquires the end-effector position signals of both arms and the force signals of the dexterous fingertips; calculates the midpoint of the line connecting the end-effectors of both arms as the center reference point of both arms, and calculates the center position of the grasping force based on the magnitude and position of the fingertip force; constructs a virtual constraint line segment using the line connecting the center reference point of both arms and the center position of the grasping force, and uses the line segment length as the geometric offset; tracks the attitude change of the virtual constraint line segment, and extracts the rotational angular velocity of the direction vector as the structural constraint relaxation rate; constructs a state trajectory using the geometric offset and relaxation rate as the horizontal and vertical coordinates, extracts the trajectory tortuosity, and compares it with a threshold to determine whether the system is in a tightly coupled or loosely coupled state; adjusts the position control gain and force control gain according to the coupling state, generates a hybrid control command output to the dual-arm joint controller and the dexterous hand finger controller for execution; and senses the degree of structural coupling between the dual arms and the dexterous hand through the virtual constraint line segment, achieving a smooth hybrid of position and force control.

[0070] Please see Figure 2 The diagram shows a structural block diagram of a dual-arm dexterity hybrid control system according to this application.

[0071] like Figure 2 As shown, the dual-arm dexterity hybrid control system 200 includes an acquisition module 210, a calculation module 220, a construction module 230, an extraction module 240, a judgment module 250, and an output module 260.

[0072] The acquisition module 210 is configured to acquire the end-position signals of the two arms in space and the fingertip force signals of each finger of the dexterous hand; the calculation module 220 is configured to calculate the midpoint of the line connecting the ends of the two arms based on the end-positions of the arms, using it as the center reference point of the two arms, and to calculate the center position of the grasping force based on the magnitude of the fingertip force and the spatial position of the fingertips; the construction module 230 is configured to construct a virtual constraint line segment between the two arms and the dexterous hand using the line connecting the center reference point of the two arms and the center position of the grasping force, and to use the length of the virtual constraint line segment as the geometric offset between the center of the two arms and the center of the grasping force; the extraction module 240 is configured to... The virtual constraint line segment's attitude change in space is tracked, and the rotational angular velocity of the direction vector of the virtual constraint line segment on the time axis is extracted as the structural constraint relaxation rate of the dual-arm dexterous hand system. The judgment module 250 is configured to determine whether the dual-arm dexterous hand system is in a tightly coupled state based on the geometric offset and the structural constraint relaxation rate. The output module 260 is configured to adjust the position control gain and force control gain if the dual-arm dexterous hand system is in a tightly coupled state, and generate a hybrid control command based on the adjusted position control gain and force control gain, and output it to the dual-arm joint controller and dexterous hand finger controller for execution.

[0073] It should be understood that Figure 2 The modules and references described in the document Figure 1The steps described in the text correspond to those in the method described above. Therefore, the operations, features, and corresponding technical effects described above also apply to the method described in the text. Figure 2 The various modules in the document will not be described in detail here.

[0074] In other embodiments, the present invention also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, causes the processor to perform the dual-arm dexterity hand hybrid control method in any of the above method embodiments. In one embodiment, the computer-readable storage medium of the present invention stores computer-executable instructions, which are configured as follows: Acquire the end-position signals of both arms in space and the fingertip force signals of each finger of the dexterous hand; The midpoint of the line connecting the ends of the arms is calculated based on the position of the ends of the arms and used as the center reference point of the arms. The center position of the grasping force is calculated based on the magnitude of the fingertip force and the spatial position of the fingertips. A virtual constraint line segment is constructed between the two dexterous arms by connecting the reference point of the two arm centers and the position of the grasping force center. The length of the virtual constraint line segment is used as the geometric offset between the center of the two arms and the center of the grasping force. Track the attitude change of the virtual constraint line segment in space, and extract the rotational angular velocity of the direction vector of the virtual constraint line segment on the time axis as the structural constraint relaxation rate of the dual-arm dexterous hand system; Determine whether the dual-arm dexterous hand system is in a tightly coupled state based on the geometric offset and the structural constraint relaxation rate. If the dual-arm dexterous hand system is in a tightly coupled state, the position control gain and force control gain are adjusted, and a hybrid control command is generated based on the adjusted position control gain and force control gain, which is then output to the dual-arm joint controller and the dexterous hand finger controller for execution.

[0075] Computer-readable storage media may include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application program required for at least one function; the data storage area may store data created based on the use of the dual-arm dexterity hybrid control system, etc. Furthermore, the computer-readable storage medium may include high-speed random access memory, and may also include memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, the computer-readable storage medium may optionally include memory remotely disposed relative to a processor, which can be connected to the dual-arm dexterity hybrid control system via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0076] Figure 3This is a schematic diagram of the structure of the electronic device provided in the embodiment of the present invention, such as... Figure 3 As shown, the device includes a processor 310 and a memory 320. The electronic device may also include an input device 330 and an output device 340. The processor 310, memory 320, input device 330, and output device 340 can be connected via a bus or other means. Figure 3 Taking a bus connection as an example, the memory 320 is the computer-readable storage medium described above. The processor 310 executes various server functions and data processing by running non-volatile software programs, instructions, and modules stored in the memory 320, thereby implementing the dual-arm dexterity hybrid control method described in the above embodiment. The input device 330 can receive input digital or character information and generate key signal inputs related to user settings and function control of the dual-arm dexterity hybrid control system. The output device 340 may include a display screen or other display device.

[0077] The aforementioned electronic device can execute the method provided in the embodiments of the present invention, and has the corresponding functional modules and beneficial effects for executing the method. Technical details not described in detail in this embodiment can be found in the method provided in the embodiments of the present invention.

[0078] In one implementation, the above-described electronic device is applied in a dual-arm dexterity hybrid control system for a client application, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to: Acquire the end-position signals of both arms in space and the fingertip force signals of each finger of the dexterous hand; The midpoint of the line connecting the ends of the arms is calculated based on the position of the ends of the arms and used as the center reference point of the arms. The center position of the grasping force is calculated based on the magnitude of the fingertip force and the spatial position of the fingertips. A virtual constraint line segment is constructed between the two dexterous arms by connecting the reference point of the two arm centers and the position of the grasping force center. The length of the virtual constraint line segment is used as the geometric offset between the center of the two arms and the center of the grasping force. Track the attitude change of the virtual constraint line segment in space, and extract the rotational angular velocity of the direction vector of the virtual constraint line segment on the time axis as the structural constraint relaxation rate of the dual-arm dexterous hand system; Determine whether the dual-arm dexterous hand system is in a tightly coupled state based on the geometric offset and the structural constraint relaxation rate. If the dual-arm dexterous hand system is in a tightly coupled state, the position control gain and force control gain are adjusted, and a hybrid control command is generated based on the adjusted position control gain and force control gain, which is then output to the dual-arm joint controller and the dexterous hand finger controller for execution.

[0079] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of various embodiments or some parts of embodiments.

[0080] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for hybrid control of two-arm dexterous hands, characterized in that, include: Acquire the end-position signals of both arms in space and the fingertip force signals of each finger of the dexterous hand; The midpoint of the line connecting the ends of the arms is calculated based on the position of the ends of the arms and used as the center reference point of the arms. The center position of the grasping force is calculated based on the magnitude of the fingertip force and the spatial position of the fingertips. A virtual constraint line segment is constructed between the two dexterous arms by connecting the reference point of the two arm centers and the position of the grasping force center. The length of the virtual constraint line segment is used as the geometric offset between the center of the two arms and the center of the grasping force. Track the attitude change of the virtual constraint line segment in space, and extract the rotational angular velocity of the direction vector of the virtual constraint line segment on the time axis as the structural constraint relaxation rate of the dual-arm dexterous hand system; Determine whether the dual-arm dexterous hand system is in a tightly coupled state based on the geometric offset and the structural constraint relaxation rate. If the dual-arm dexterous hand system is in a tightly coupled state, the position control gain and force control gain are adjusted, and a hybrid control command is generated based on the adjusted position control gain and force control gain, which is then output to the dual-arm joint controller and the dexterous hand finger controller for execution.

2. The dual-arm dexterity hand hybrid control method according to claim 1, characterized in that, The calculation of the midpoint of the line connecting the ends of the two arms, based on the position of the ends of the arms, serves as the center reference point for the two arms. The calculation of the center position of the grasping force based on the magnitude of the fingertip force and the spatial position of each fingertip includes: Obtain the first end position coordinates of the left arm end in space and the second end position coordinates of the right arm end in space. Use the line connecting the first end position coordinates and the second end position coordinates as the inter-arm reference baseline. Use the intersection of the perpendicular plane of the inter-arm reference baseline and the inter-arm reference baseline as the center reference point of the two arms. The magnitude of the fingertip force of each finger of the dexterous hand and the spatial coordinates of the fingertip position of each finger are obtained. Using the magnitude of the fingertip force of each finger as the force value, the resultant force components of the fingertip force in three orthogonal directions in space are calculated. The spatial coordinates of each resultant force component when they reach equilibrium are calculated respectively. The average value of the spatial coordinates of each resultant force component when they reach equilibrium is taken as the position of the grasping force center.

3. The method for hybrid control of a dual-arm dexterous hand according to claim 1, characterized in that, The construction of a virtual constraint line segment between the two dexterous arms using the line connecting the center reference point of the two arms and the center of the grasping force, and the use of the length of the virtual constraint line segment as the geometric offset between the center of the two arms and the center of the grasping force, includes: Obtain the position of the left arm end, the position of the right arm end, and the position of the gripping force center. Construct a dual-arm dexterity hand constraint triangle with the left arm end, the right arm end, and the gripping force center as the three vertices. The midline from the center reference point of the two arms to the center of the grasping force in the dual-arm dexterity hand constraint triangle is used as a virtual constraint line segment. The virtual constraint line segment divides the constraint triangle into two sub-triangles. The absolute value of the difference between the areas of the two sub-triangles is used as the asymmetry of the dual-arm dexterity hand constraint triangle. The interior angle of the constraint triangle of the dual-arm dexterity hand at the reference point at the center of the dual arms is obtained, and the difference between the interior angle and the preset reference angle is used as the structural angle deviation of the dual-arm dexterity hand system. Multiply the structural angular deviation by the asymmetry of the dual-arm dexterity hand constraint triangle, and use the product as the geometric offset between the center of the dual arms and the center of the grasping force.

4. The dual-arm dexterity hand hybrid control method according to claim 1, characterized in that, The step of tracking the attitude change of the virtual constraint line segment in space and extracting the rotational angular velocity of the direction vector of the virtual constraint line segment on the time axis as the structural constraint relaxation rate of the dual-arm dexterity hand system includes: Obtain the first direction vector of the virtual constraint line segment at the current moment, the second direction vector of the virtual constraint line segment at the previous moment, and the third direction vector of the virtual constraint line segment at the two moments before the previous moment; Calculate the first angle between the first direction vector and the second direction vector as the instantaneous rotation angle at the current moment; calculate the second angle between the second direction vector and the third direction vector as the instantaneous rotation angle at the previous moment. The rotational angular acceleration at the current moment is obtained by dividing the difference between the first included angle and the second included angle by the time interval between the current moment and the previous moment. The rotation direction of the direction vector is determined based on the sign change of the rotation angular acceleration. The duration during which the rotation direction does not reverse within a preset number of consecutive moments is taken as the duration during which the rotation direction of the direction vector is maintained. The ratio of the instantaneous rotation angle at the current moment to the duration of the rotation direction is used as the structural constraint relaxation rate of the dual-arm dexterity hand system.

5. The dual-arm dexterity hand hybrid control method according to claim 1, characterized in that, The step of determining whether the dual-arm dexterous hand system is in a tightly coupled state based on the geometric offset and the structural constraint relaxation rate includes: Using the geometric offset as the horizontal axis and the structural constraint relaxation rate as the vertical axis, the current geometric offset value and the structural constraint relaxation rate value are combined into a state point in a two-dimensional plane. Obtain the state points corresponding to multiple moments within a preset time window before the current moment, and connect the multiple state points in chronological order to obtain the state trajectory of the state points in the two-dimensional plane. Calculate the displacement vector between two adjacent state points in the state trajectory to obtain a displacement vector sequence, and calculate the direction change angle between two adjacent displacement vectors in the displacement vector sequence to obtain a direction change angle sequence; The sum of the absolute values ​​of the directional change angles in the directional change angle sequence is taken as the cumulative directional change of the state trajectory, and the sum of the magnitudes of the displacement vectors in the displacement vector sequence is taken as the total displacement of the state trajectory. The ratio of the cumulative directional change to the total displacement is used as the tortuosity of the state trajectory, and it is determined whether the tortuosity of the state trajectory is less than a preset tortuosity threshold. When the tortuosity is less than a preset tortuosity threshold, the dual-arm dexterity hand system is determined to be in a tightly coupled state. When the degree of tortuosity is not less than the preset degree of tortuosity threshold, the dual-arm dexterity hand system is determined to be in a loosely coupled state.

6. The dual-arm dexterity hand hybrid control method according to claim 1, characterized in that, The adjustment of position control gain and force control gain, and the generation of hybrid control commands based on the adjusted position control gain and force control gain, which are then output to the dual-arm joint controller and the dexterous hand finger controller for execution, include: Obtain the tight coupling determination results for each moment within a preset time window prior to the current moment, and divide the number of tight coupling determinations within the preset time window by the total number of determinations within the preset time window to obtain the tight coupling percentage. Calculate the change in the tight coupling ratio at the current moment relative to the tight coupling ratio at the previous moment, and obtain the ratio change rate. The product of the tight coupling ratio and the ratio change rate is used as the coupling situation prediction factor. When the coupling situation prediction factor is greater than the preset first factor threshold, the force control gain is increased by a first step length and the position control gain is decreased by a second step length. The first step length is the adjustment step length of the force control gain and the second step length is the adjustment step length of the position control gain. The first step length is greater than the second step length. When the coupling situation prediction factor is less than the preset second factor threshold, the position control gain is increased by the second step size, and the force control gain is decreased by the first step size. When the coupling situation prediction factor is greater than or equal to the second factor threshold and less than or equal to the first factor threshold, the position control gain and the force control gain remain unchanged. The deviation between the current position and the desired position of the two arms' ends is obtained as the position error of the two arms' ends; the deviation between the current fingertip force and the desired fingertip force of each finger of the dexterous hand is obtained as the dexterous fingertip force error. The position control component is obtained by multiplying the position error of the two arms end by the adjusted position control gain, and the force control component is obtained by multiplying the force error of the dexterous fingertips by the adjusted force control gain. The position control component and the force control component are added together, and the result of the addition is output as the hybrid control command.

7. A dual-arm dexterity hand hybrid control system, characterized in that, include: The acquisition module is configured to acquire the end position signals of the two arms in space and the fingertip force signals of each finger of the dexterous hand; The calculation module is configured to calculate the midpoint of the line connecting the ends of the two arms based on the end positions of the two arms, and use it as the center reference point of the two arms. It also calculates the center position of the grasping force based on the magnitude of the fingertip force and the spatial position of the fingertip. The construction module is configured to construct a virtual constraint line segment between the two dexterous arms using the line connecting the center reference point of the two arms and the position of the center of the grasping force, and to use the length of the virtual constraint line segment as the geometric offset between the center of the two arms and the center of the grasping force. The extraction module is configured to track the attitude change of the virtual constraint line segment in space and extract the rotational angular velocity of the direction vector of the virtual constraint line segment on the time axis as the structural constraint relaxation rate of the dual-arm dexterous hand system. The judgment module is configured to determine whether the dual-arm dexterous hand system is in a tightly coupled state based on the geometric offset and the structural constraint relaxation rate. The output module is configured to adjust the position control gain and force control gain if the dual-arm dexterous hand system is in a tightly coupled state, and generate a hybrid control command based on the adjusted position control gain and force control gain, and output it to the dual-arm joint controller and dexterous hand finger controller for execution.

8. An electronic device, characterized in that, include: At least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by a processor, it implements the method described in any one of claims 1 to 6.

Images