Self-adaptive variable impedance algorithm for hydraulic four-foot support unloading

Abstract

The invention discloses a self-adaptive variable impedance algorithm for hydraulic four-foot support unloading, which comprises the following specific steps: step 1, single-cycle variable impedance control: introducing a variable impedance control strategy in a single-step cycle, judging that a robot is subjected to large impact when a large force change rate is detected by a force sensor, and judging that the robot is subjected to large impact when the force change rate is larger; at the moment, expected force is rapidly adjusted to achieve force unloading, and therefore impact damage to structural parts is reduced; step 2, dynamic impedance coefficient adjustment: when it is detected that the impact force change rate exceeds a preset threshold value or the speed is increased, the impedance coefficient is dynamically reduced to reduce rigidity and improve flexibility, and rapid force unloading is assisted; and step 3, performing impact detection and real-time response. The hydraulic quadruped robot can dynamically adapt to complex tracks and stress changes, compliant control and whole-course stable operation are achieved, the unloading efficiency and the structural reliability are improved, the compliance and the movement capacity are also considered, and reliable support is provided for efficient walking of the hydraulic quadruped robot in the complex environment.

Description

Technical Field

[0001] The invention belongs to the technical field of robot algorithms, and in particular relates to an adaptive variable impedance algorithm for hydraulic quadruped support unloading. Background Art

[0002] The adaptive variable impedance algorithm for hydraulic quadruped support unloading refers to a control algorithm used in hydraulic systems such as quadruped robots to adjust the support force of the robot's legs to adapt to changes in the external environment and achieve stable support. This algorithm is usually based on the variable impedance control theory and combined with the characteristics of the hydraulic system to achieve dynamic adjustment of the support force. In this algorithm, the support force of the robot's legs can be adaptively adjusted according to the force changes in the external environment to maintain a stable support state. By monitoring the changes in external forces and combining the robot's own state information, the algorithm can adjust the parameters of the hydraulic system in real time to control the size and direction of the leg support force to ensure that the robot remains balanced and stable when moving or standing. This adaptive variable impedance algorithm usually needs to be combined with sensor feedback and real-time control strategies to dynamically adjust the impedance parameters and support force size to adapt to different terrains, external forces and movement requirements. Through this algorithm, the hydraulic quadruped support system can better adapt to complex environments and task requirements and improve the stability and flexibility of the robot.

[0003] Traditional impedance control algorithms are mainly applicable to multi-rigid body systems with fixed bases, which exhibit compliant characteristics when the system is subjected to constant force. However, for a hydraulically driven quadruped robot, its model is a floating base, and the external force it receives changes dynamically with the change of the motion trajectory. In this case, traditional impedance control methods are difficult to meet the requirements of the hydraulic quadruped robot to achieve instant force unloading and maintain compliant characteristics during rapid motion.

[0004] The hydraulic actuator can output a large force. If the force cannot be quickly unloaded during the landing process, it will cause a large impact on the mechanical structure. If the impedance coefficient is set too small, the robot may not be able to provide enough driving force to complete the leg lifting action when the support phase turns into the swing phase. Therefore, in order to achieve the smooth control requirements of the hydraulic quadruped robot when walking quickly, it is necessary to design an adaptive variable impedance algorithm that can combine the robot's motion trajectory and real-time force information. Summary of the invention

[0005] 1. Technical issues to be resolved

[0006] In view of the shortcomings of the prior art, the present invention provides an adaptive variable impedance algorithm for hydraulic quadruped support unloading, which solves the problem that traditional impedance control methods are difficult to meet the requirements of hydraulic quadruped robots to achieve instant unloading and maintain compliant characteristics during fast walking.

[0007] (2) Technical solution

[0008] The technical solution of the present invention is: an adaptive variable impedance algorithm for hydraulic quadruped support force unloading, and the specific steps are as follows:

[0009] Step 1, single-cycle variable impedance control. During a single-step cycle, a variable impedance control strategy is introduced. When a large change rate of force is detected by a force sensor, it is determined that the robot is subjected to a large impact. At this time, the expected force value is quickly adjusted to achieve force unloading, thereby reducing the impact damage to the structural components;

[0010] Step 2, dynamic impedance coefficient adjustment. When it is detected that the large impact force change rate exceeds the preset threshold and the speed increases, the impedance coefficient is dynamically reduced to reduce rigidity and improve compliance to assist in achieving rapid force unloading;

[0011] Step 3, impact detection and real-time response. By real-time monitoring the change rate of the force at the foot end of the robot and the movement speed, when the change rate exceeds the preset threshold or the speed changes, the adjustment of the impedance coefficient and the expected force is immediately triggered to adapt to the external force impact and complete the force unloading operation.

[0012] As a further preferred embodiment of the present invention, it further includes a method for calibrating the expected force of the variable impedance of the hydraulic cylinder. Let the hydraulic quadruped run in a trot gait with the legs swinging in the air without touching the ground, record the upper and lower limits of the force fluctuation received by the knee joint tension and pressure sensor, and set the upper and lower limit values as (x1, x2). If the sensor value becomes larger when the compression cylinder is working, the current force sensor value is y1, and the expected force is e1. If y1 > x2, e1 = x1; otherwise, e1 = x2. If the sensor value becomes larger when the stretching cylinder is working, then if y1 < x2, e1 = x2; otherwise, e1 = x1.

[0013] As a further preferred embodiment of the present invention, it further includes a method for planning the foot end trajectory of the hydraulic quadruped. Swing phase trajectory: Support phase trajectory:

[0014] X = -sv.a + X1, Z = Z0, where X is the foot end trajectory in the X direction, Z is the foot end trajectory in the Z direction, S is half of the step length, A5 = S * 2 / T 2 / 25.13, ω l = π / T, ω h = 2π / T / 2.86, φ = 0.218979522475626, and a is the discrete time.

[0015] As a further preferred embodiment of the present invention, it also includes a method for adaptively adjusting the impedance parameters of the hydraulic cylinder, wherein the impedance coefficient is mainly composed of M mass, B damping and K stiffness, and the foot end trajectory of the electro-hydraulic quadruped is planned to be uniform motion, that is, the acceleration is 0, M is set to 0, the size of B depends on the speed of the quadruped robot, the size of K depends on the current force, the current speed of the hydraulic robot is v, and B(v)=B0e av , B0 mainly depends on the discretized impedance formula (M+BT+KT 2 )Δx(k)=T 2 ΔF(k)+(2M+BT)x(k-1)-MΔx(k-2), T is the discrete time, which depends on the time it takes for the host computer to perform a force control closed-loop operation, Δx(k) is the current displacement compensation value, Δx(k-1) is the displacement compensation value of the previous cycle, Δx(k-2) is the displacement compensation value of the previous cycle, ΔF(k)=F(k)-e1, F(k) is the force value received by the current force sensor, and the total stroke of the hydraulic cylinder is X. Substitute all the above values, and when ΔF(k)=X / 15, the B0 value is obtained;

[0016] K is the elastic coefficient. If F(k-1) is the force sensor value of the previous cycle, F(k-2) is the force sensor value of the cycle before that. The smaller the K value, the larger the Δx(k), that is, the larger the displacement value to be compensated. When the hydraulic quadruped enters the support phase, if it has just landed, a larger displacement compensation is required, then the K value needs to be reduced. If it enters the support phase to the swing phase, a smaller displacement compensation is required to ensure sufficient force to lift, then the K value needs to be increased. So K = a*K0e -(F(k)-F(k-1)) , if Z=z1, Z is the foot end trajectory value in the Z direction, z1 is the minimum value of the foot end trajectory in the Z direction, when Z=z1, the hydraulic quadruped enters the support phase, then a=60 / m, m is the mass of the hydraulic quadruped, otherwise a=1, to ensure that the elastic coefficient of the swing phase is equal to the elastic coefficient of the support phase. The quadruped is subjected to a larger force only when it is in the support phase. K0 also depends on the displacement compensation value Δx(k). When K0 is substituted into the discretized impedance formula, Δx(k) should be ≤X / 15. When K0 is equal to 1, K=K0e -(F(k)-F(k-1)) .

[0017] (III) Beneficial effects

[0018] The present invention provides an adaptive variable impedance algorithm for hydraulic quadruped support unloading. It has the following beneficial effects:

[0019] The adaptive variable impedance algorithm proposed in the present invention can dynamically combine the motion trajectory and real-time force information of the hydraulic quadruped robot, adapt to complex working conditions, and effectively make up for the limitations of traditional impedance control methods in dynamic environments. Improve the efficiency of unloading and reduce impact. By detecting the rate of change of force in real time and adjusting the expected force value, the robot can quickly unload force when it lands, significantly reducing the impact of the hydraulic actuator output on the mechanical structure, extending the life of the equipment, and improving the operation stability. To ensure the support and leg-lifting performance, a dynamic impedance coefficient adjustment strategy is adopted to maintain sufficient rigidity in the support phase to provide the driving force for leg lifting, and quickly reduce the impedance coefficient during impact to achieve smooth unloading, taking into account the flexibility and movement ability of the robot, and ensuring the stable realization of fast walking performance. Compatible with dynamic trajectory adjustment, the algorithm fully considers the floating base characteristics of the hydraulic quadruped robot and the characteristics of force changing with the motion trajectory. It can continuously adapt to the force requirements of different trajectories in a dynamic environment and ensure efficient motion control capabilities. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a shape diagram of a trot gait when moving;

[0021] Figure 2 It is the X direction trajectory and Z direction trajectory diagram;

[0022] Figure 3 B(v)=B0e av Graphs;

[0023] Figure 4 When K0 is equal to 1, K = K0e -(F(k)-F(k-1)) image;

[0024] Figure 5 This is the actual force curve of a single leg of a common impedance hydraulic quadruped;

[0025] Figure 6 This is a curve diagram of the actual force on a single leg after adding adaptive variable impedance. DETAILED DESCRIPTION

[0026] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0027] See also Figure 1-6, an embodiment of the present invention provides a technical solution: an adaptive variable impedance algorithm for hydraulic quadruped support force unloading, including single-cycle variable impedance control. During a single-step cycle, a variable impedance control strategy is introduced. When a large change rate of force is detected by a force sensor, it is determined that the robot is subjected to a large impact. At this time, the expected force value is quickly adjusted to achieve force unloading, thereby reducing the impact damage to the structural components.

[0028] Dynamic impedance coefficient adjustment. Considering the influence of the setting of the impedance coefficient on the compliance and motion ability of the robot, a moderate impedance coefficient is maintained under normal support conditions to ensure that the robot can provide sufficient leg-lifting force during the support phase; when a large impact (the change rate of force exceeds a preset threshold) is detected, the impedance coefficient is dynamically reduced to reduce rigidity and improve compliance, assisting in achieving rapid force unloading.

[0029] Impact detection and real-time response. By real-time monitoring the change rate of the force at the foot end of the robot and the motion speed, when the change rate exceeds the preset threshold or the speed changes, the adjustment of the impedance coefficient and the expected force is immediately triggered to adapt to the external force impact and complete the force unloading operation, thereby ensuring the running stability and walking efficiency of the robot in complex terrains.

[0030] Hydraulic cylinder variable impedance expected force calibration method

[0031] Let the hydraulic quadruped run in a trot gait with the legs swinging in the air without touching the ground, and record the upper and lower limits of the force fluctuations received by the knee joint tension and compression sensors. Let the upper and lower limit values be (x1, x2). If the sensor value increases when the cylinder is compressed, the current force sensor value is y1, and the expected force is e1. If y1 > x2, e1 = x1; otherwise, e1 = x2. If the sensor value increases when the cylinder is stretched, then if y1 < x2, e1 = x2; otherwise, e1 = x1.

[0032] Hydraulic quadruped foot end trajectory planning method

[0033] The core of impedance control is to establish an expected force-position relationship. If virtual inertia is introduced, the dynamic behavior of the system will become more "slow" or "lagging", which is not conducive to rapid response. Therefore, it is best for the hydraulic quadruped to move at a constant speed during movement, so that the inertia of the object will not have too much impact. The gait during movement is the trot gait, and the shape is like a steamed bun, as Figure 1 shown.

[0034] Swing phase trajectory: Support phase trajectory:

[0035] X = -sv.a + X1, Z = Z0. X is the foot end trajectory in the X direction, Z is the foot end trajectory in the Z direction, S is half of the step length, A5 = S * 2 / T 2 / 25.13, ω l=π / T,ω h =2π / T / 2.86, φ=0.218979522475626, a is the discrete time. The X-direction trajectory and the Z-direction trajectory are as follows Figure 2 Shown:.

[0036] Adaptive Adjustment Method of Hydraulic Cylinder Impedance Parameters

[0037] The impedance coefficient is mainly composed of M, mass, B, damping and K, and the stiffness. Since the foot trajectory of the electro-hydraulic quadruped is planned to be uniform motion, that is, the acceleration is 0, M is generally set to 0, the size of B depends on the speed of the quadruped robot, and the size of K depends on the current force.

[0038] Assume the current speed of the hydraulic robot is v, B(v) = B0e av , B0 mainly depends on the discretized impedance formula (M+BT+KT 2 )Δx(k)=T 2 ΔF(k)+(2M+BT)Δx(k-1)-MΔx(k-2), T is the discrete time, which depends on the time it takes for the host computer to perform a force control closed loop operation, Δx(k) is the current displacement compensation value, Δx(k-1) is the displacement compensation value of the previous cycle, Δx(k-2) is the displacement compensation value of the previous cycle, ΔF(k)=F(k)-e1, F(k) is the force value received by the current force sensor. Let the total stroke of the hydraulic cylinder be X. Substitute all the above values ​​and obtain the B0 value when ΔF(k)=X / 15. a∈(0,1). B(v)=B0e av Curve like Figure 3 Shown:.

[0039] K is the elastic coefficient, F(k-1) is the force sensor value of the previous cycle, and F(k-2) is the force sensor value of the cycle before that. The smaller the K value, the larger the Δx(k), that is, the larger the displacement value to be compensated. When the hydraulic quadruped enters the support phase, if it has just landed, a larger displacement compensation is required, then the K value needs to be reduced. If it enters the support phase to the swing phase, a smaller displacement compensation is required to ensure that there is enough force to lift, then the K value needs to be increased. So K = a*K0e -(F(k)-F(k-1)) If Z = z1, Z is the foot end trajectory value in the Z direction, z1 is the minimum foot end trajectory value in the Z direction, when Z = z1, the hydraulic quadruped enters the support phase, then a = 60 / m, m is the mass of the hydraulic quadruped, otherwise a = 1. This is to ensure that the elastic coefficient of the swing phase is greater than the elastic coefficient of the support phase, because the quadruped is subjected to greater force when it is in the support phase.

[0040] K0 also depends on the displacement compensation value Δx(k). When K0 is substituted into the discretized impedance formula, x(k) should be ≤X / 15. When K0 is equal to 1, K=K0e -(F(k)-F(k-1)) Image as Figure 4Shown:.

[0041] The magnitudes of B0 and K0 mainly depend on the magnitudes of the cylinder displacement sensor and the force sensor. When B0 and K0 are temporarily determined, it is necessary to ensure that Δx(k) and X are of the same magnitude, and Δx(k) should be as small as possible as small as X / 15. If the compensation displacement is too large, it will affect the trot gait.

[0042] The single leg structure of a hydraulic quadruped robot usually includes three degrees of freedom, namely the side swing joint, hip joint and knee joint. The side swing joint is mainly used to achieve the steering function, so there is no need to compensate the displacement of the side swing hydraulic cylinder during normal walking. In contrast, during the landing phase, the knee joint hydraulic cylinder is the main load bearer, so displacement compensation usually only needs to be designed and optimized for the knee joint hydraulic cylinder.

[0043] Comparison of actual effects:

[0044] Figure 5 is the actual force curve of a single leg of a common impedance hydraulic quadruped. Figure 6 The actual force situation of a single leg after adding adaptive variable impedance. It is obvious that the curve is smoother and the peak force is larger, which proves that the algorithm can have a better buffering effect for hydraulic quadrupeds with heavier weight.

[0045] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and the present invention can be implemented in other specific forms without departing from the spirit or basic features of the present invention. Therefore, no matter from which point of view, the embodiments should be regarded as exemplary and non-restrictive. The scope of the present invention is defined by the attached claims rather than the above description, and it is intended that all changes falling within the meaning and scope of the equivalent elements of the claims are included in the present invention. Any figure mark in the claims should not be regarded as limiting the claims involved.

[0046] In addition, it should be understood that although the present specification is described according to implementation modes, not every implementation mode contains only one independent technical solution. This description of the specification is only for the sake of clarity. Those skilled in the art should regard the specification as a whole. The technical solutions in each embodiment may also be appropriately combined to form other implementation modes that can be understood by those skilled in the art.

Claims

1. An adaptive variable impedance algorithm for hydraulic quadruped support unloading, characterized in that: The specific steps include the following: Step 1, single-cycle variable impedance control. During a single-step cycle, a variable impedance control strategy is introduced. When a large change rate of force is detected by the force sensor, it is determined that the robot is subjected to a large impact. At this time, the expected force value is quickly adjusted to achieve force unloading, thereby reducing the impact damage to the structural components; Step 2, dynamic impedance coefficient adjustment. When the change rate of the impact force exceeds the preset threshold and the speed increases, the coefficient is dynamically reduced to reduce rigidity and improve compliance to assist in achieving rapid force unloading; Step 3, impact detection and real-time response. By real-time monitoring the change rate of the force at the foot end of the robot and the movement speed, when the change rate exceeds the preset threshold or the speed changes, the adjustment of the impedance coefficient and the expected force is immediately triggered to adapt to the external force impact and complete the force unloading operation.

2. The adaptive variable impedance algorithm for hydraulic quadruped support unloading according to claim 1, characterized in that: It also includes a calibration method for the expected force of the variable impedance of the hydraulic cylinder. Let the hydraulic quadruped perform a trot gait with an empty swing, record the upper and lower limits of the force fluctuation received by the knee joint pull-pressure sensor, and set the upper and lower limit values as (x1, x2). If the sensor value increases when the compression cylinder is working, the current force sensor value is y1, and the expected force is e1. If y1 > x2, e1 = x1; otherwise, e1 = x2. If the sensor value increases when the stretching cylinder is working, then if y1 < x2, e1 = x2; otherwise, e1 = x1.

3. The adaptive variable impedance algorithm for hydraulic quadruped support unloading according to claim 1, characterized in that: Also included are the hydraulic quadruped foot end trajectory planning method and the swing phase trajectory: Support phase trajectory: X=-sv.a+X1,Z=Z0,X is the foot end trajectory in X direction, Z is the foot end trajectory in Z direction, S is half the step length, A5=S*2 / T 2 / 25.13, ω l =π / T,ω h =2π / T / 2.86, φ=0.218979522475626, a is the discrete time.

4. The adaptive variable impedance algorithm for hydraulic quadruped support unloading according to claim 1, characterized in that: It also includes a method for adaptively adjusting the impedance parameters of the hydraulic cylinder. The impedance coefficient is mainly composed of M mass, B damping and K stiffness. The foot end trajectory of the electro-hydraulic quadruped is planned to be uniform motion, that is, the acceleration is 0, M is set to 0, the size of B depends on the speed of the quadruped robot, the size of K depends on the current force, the current speed of the hydraulic robot is v, B(v) = B0e av , B0 mainly depends on the discretized impedance formula (M+BT+KT 2 )Δx(k)=T 2 ΔF(k)+(2M+BT)Δx(k-1)-MΔx(k-2), T is the discrete time, which depends on the time it takes for the host computer to perform a force control closed-loop operation. Δx(k) is the current displacement compensation value, Δx(k-1) is the displacement compensation value of the previous cycle, Δx(k-2) is the displacement compensation value of the previous cycle, F(k)=F(k)-e1, F(k) is the force value received by the current force sensor, and the total stroke of the hydraulic cylinder is X. Substitute all the above values, and when ΔF(k)=X / 15, the B0 value is obtained; K is the elastic coefficient, F(k-1) is the force sensor value of the previous cycle, and F(k-2) is the force sensor value of the previous cycle. The smaller the K value is, the larger the Δx(k) value is, that is, the larger the compensated displacement value is. When the hydraulic quadruped enters the support phase, if it has just landed, a larger displacement compensation is required, then the K value needs to be reduced. If it enters the support phase to the swing phase, a smaller displacement compensation is required to ensure sufficient force to lift, then the K value needs to be increased, so K = a*K0e -(F(k)-F(k-1)) , if Z=z1, Z is the foot end trajectory value in the Z direction, z1 is the minimum value of the foot end trajectory in the Z direction, when Z=z1, the hydraulic quadruped enters the support phase, then a=60 / m, m is the mass of the hydraulic quadruped, otherwise a=1, to ensure that the elastic coefficient of the swing phase is equal to the elastic coefficient of the support phase. The quadruped is subjected to a larger force only when it is in the support phase. K0 also depends on the displacement compensation value Δx(k). When K0 is substituted into the discretized impedance formula, Δx(k) should be ≤X / 15. When K0 is equal to 1, K=K0e -(F(k)-F(k-1)) .

Images