Regulation and control system based on smooth transition of multi-section movement speed

Abstract

The invention relates to the technical field of motion control, and discloses a multi-section motion speed smooth transition-based regulation and control system, which comprises an instruction look-ahead unit for acquiring a state deviation vector of a motion section connection point; the parameter modulation unit constructs a control potential energy model representing energy distribution; the dynamic adjusting unit obtains an inertia coupling matrix of the driving execution unit, calculates a transition correction increment in combination with gradient pointing of a control potential energy model, and converts signal jump into guiding force information; the optimization output unit loads guiding force information into an energy functional to adjust coordinates of a control point of a to-be-machined path, and target spline parameters with first-order derivative continuity are output. Physical domain resonance impact is counteracted in advance by constructing a control potential energy model, phase lag and track precision loss caused by traditional filtering are avoided, and the precision of the to-be-machined path is improved. And smooth and high-precision operation of the multi-axis linkage system is ensured.

Description

Technical Field

[0001] This invention relates to the field of motion control technology, and more specifically to a control system based on smooth transition of multi-segment motion speed. Background Technology

[0002] Currently, the continuous splicing of multiple path segments forms the basis for realizing complex spatial geometry. The complete trajectory is decomposed into several discrete motion command segments, and independent velocity planning is processed within each segment. The steady-state tracking quality and dynamic response characteristics of the control system are related to the continuity of the first or higher-order derivatives of the input command stream. When the motion command generates velocity vector deflection or acceleration step at the junction of adjacent segments, the high-frequency components are prone to exceed the frequency response bandwidth of the servo system. Under the conditions of high-viscosity material extrusion or precision machining, the step of the command derivative at the trajectory junction point causes the transmission chain to generate closed-loop oscillation. This phenomenon causes the tracking error to accumulate continuously with the motion process in the multi-axis coupling state, and the inconsistency of spatial trajectory deformation due to the difference in the dynamic response characteristics of each axis. In order to maintain the processing rhythm, the existing technology usually needs to compromise between shape accuracy and response speed, and defaults to accepting the system impact load generated by the command step.

[0003] To address the aforementioned oscillations, the industry often employs linear methods such as geometric spline smoothing or post-low-pass filtering. However, geometrically fitted paths deviate from the nonlinear dynamic boundary constraints of the actuator, easily leading to drive torque saturation. Post-filtering mechanisms introduce system phase lag, reducing conformity at trajectory corners. Existing technical solutions cannot achieve predictive suppression of command step energy while ensuring real-time performance. For structural vibration-sensitive characteristics, in addition to hardware optimization such as mechanical rigidity, motion control algorithm optimization also faces challenges. For example, Chinese invention patent application with publication number CN120595731A... Please disclose a method for calculating the acceleration and deceleration of each motion axis based on travel and vibration constraints. By establishing an algebraic constraint relationship between the travel and acceleration resultant vectors of each axis, the method uses iterative optimization or lookup table method to find the shortest motion time that meets the vibration constraints. The travel pre-planning mode is a static parameter adaptation when dealing with high-frequency connection of multiple continuous trajectories. It lacks real-time evolution representation of the energy contained in the velocity vector jump at the connection point. In scenarios with multi-axis inertia coupling and variable operating conditions, simple acceleration limiting is difficult to eliminate the resonance caused by command step. It lacks a predictive state mapping mechanism and introduces phase lag when suppressing vibration, resulting in the loss of trajectory corner conformity.

[0004] Therefore, the technical problem to be solved by this invention is how to establish a continuous state mapping mechanism based on physical dynamic constraints, eliminate the excitation energy generated by command step while avoiding phase lag, and improve the closed-loop bandwidth utilization of multi-axis linkage system. Summary of the Invention

[0005] This invention proposes a control system based on smooth transition of multi-segment motion velocity, the system comprising:

[0006] The instruction look-ahead unit is used to acquire the motion instruction flow representing the path to be processed, and to extract the state deviation vector containing velocity vector difference, acceleration vector difference and jerk vector difference at the connection point of adjacent motion segments.

[0007] The parameter modulation unit is used to construct a control potential energy model representing the state energy distribution of each axis in the control state space that maps the multi-axis linkage coupling characteristics, based on the state deviation vector. The control potential energy model defines the spatial curvature distribution based on the inertia level of the controlled object, and is used to transform the path deviation in the geometric domain into the potential well constraint in the energy domain.

[0008] The dynamic adjustment unit is used to obtain the inertia coupling matrix of the drive execution unit, and calculate the transition correction increment characterizing the inter-axis dynamic compensation based on the gradient direction of the control potential energy model and the inertia coupling matrix. It converts the signal mutation caused by the state deviation vector into the guiding force information that evolves along the direction of potential energy decrease in the control state space.

[0009] The optimized output unit is used to load the guiding force information as an external constraint into the energy functional equation to be optimized. Based on the external constraint, the position deflection is adjusted in multiple dimensions according to the coordinates of the control points of the processing path. The output target spline parameters with first derivative continuity and dynamic smoothness are used to update the value of the underlying interpolation register of the drive execution unit, so as to pre-cancele the resonant impact generated by the controlled object in the physical domain within the information domain.

[0010] Preferably, when calculating the transition correction increment, the dynamic adjustment unit adopts the following steps: obtaining the dynamic envelope constraint characterizing the physical output limit of the driving execution unit; determining the compensation weight of each axis according to the distribution of the diagonal elements of the inertia coupling matrix, and allocating the correction requirements generated by the gradient direction to the corresponding linkage axis components according to the compensation weight; performing amplitude limiting processing on the allocated linkage axis components according to the dynamic envelope constraint to generate the transition correction increment, so that the guiding force information can achieve gradual energy dissipation of the state deviation vector without exceeding the physical saturation limit of the driving execution unit.

[0011] Preferably, when constructing the control potential energy model, the parameter modulation unit introduces a damping coefficient that characterizes the transmission chain characteristics of the control system. This coefficient is used to suppress high-frequency command signals by adjusting the spatial curvature distribution of the control potential energy model, thereby reducing the system resonance energy excited by the speed vector jump during the control command generation stage. The damping coefficient is dynamically selected based on the real-time speed fluctuation frequency of the drive execution unit to ensure that the control potential energy model can effectively filter out noise of different frequencies and ensure the steady-state operation accuracy of the multi-axis linkage system.

[0012] Preferably, the dynamic adjustment unit adjusts the iteration step size of gradient analysis in real time according to the magnitude of the state deviation vector. When the magnitude exceeds a preset threshold of 10mm or 5mm / s, the iteration step size of gradient analysis is reduced to improve the mapping accuracy of the guiding force information to the abrupt change of the motion command flow. The dynamic adjustment unit improves the ability to capture nonlinear features at the trajectory corner by shrinking the iteration step size, ensuring that the generated transition correction increment can accurately match the dynamic response characteristics of the controlled object and prevent the accumulation of position tracking errors in high-speed processing environment.

[0013] Preferably, when the optimization output unit adjusts the position deflection, it introduces the guiding force information as an external constraint load into the energy functional equation of the path to be processed. By driving the control point to deflect in the negative gradient direction of the energy functional, the total virtual energy of the control system reaches a minimum value. The optimization output unit iteratively searches for the optimal solution of the energy functional and achieves a smooth transition of the velocity vector by fine-tuning the local geometric topology while ensuring the shape preservation of the trajectory. Finally, it obtains the target spline parameters that satisfy the kinematic constraints.

[0014] Preferably, the system also includes a status monitoring unit, which is used to acquire the feedback pose signal of the drive execution unit in real time, and dynamically compensate the energy distribution gain in the parameter modulation unit through a preset proportional-integral gain function based on the dynamic tracking error between the feedback pose signal and the target spline parameters; the status monitoring unit ensures that the control potential energy model can be adaptively adjusted according to the actual mechanical load changes of the controlled object by constructing a closed-loop adjustment loop for the energy distribution gain, thereby improving the stability of the control system under variable operating conditions.

[0015] Preferably, the state deviation vector extracted by the command look-ahead unit characterizes the geometric topological discontinuity and physical temporal mismatch at the spatial connection point of adjacent motion command segments; among them, the velocity vector difference determines the potential well depth of the control potential energy model, the acceleration vector difference determines the gradient slope of the control potential energy model, and the jerk vector difference determines the rate of change of the guiding force information. Through multi-dimensional state combination, the control system can establish a complete physical impact quantification characterization in a multi-dimensional state space.

[0016] Preferably, the dynamic envelope constraints include the maximum output torque saturation limit of each motor in the drive execution unit, the maximum allowable centripetal acceleration of each transmission shaft in high-speed motion, and the vibration stiffness limit of each mechanical structure; the dynamic adjustment unit ensures that the generated transition correction increment is physically feasible by limiting the linkage shaft components within the convex space defined by the dynamic envelope constraints, thereby avoiding protective shutdown of the drive system or fatigue damage to the mechanical structure due to command overshoot.

[0017] Preferably, the target spline parameters output by the optimized output unit are directly mapped to the periodic sampling point data of the interpolation controller inside the controlled object, enabling the drive execution unit to gradually switch the speed direction and modulus based on the guidance of the control potential energy model when performing multi-segment path connection tasks. After the target spline parameters are updated to the hardware cache, the multi-axis linkage system is driven by high-frequency synchronous pulses to run according to the optimized smooth curve, thereby significantly reducing the mechanical wear and energy consumption of the transmission chain while ensuring processing efficiency.

[0018] The beneficial effects of this invention are:

[0019] 1. In the control of smooth transition of multi-segment motion speed, through the coordinated linkage of command look-ahead unit and guide field construction unit, the instantaneous velocity vector jump of adjacent motion segments is converted into a guide force field that evolves continuously in the state space. This enables the originally discrete trajectory connection points to have the continuity of the first derivative at the command issuance stage, eliminates the excitation energy of the resonant frequency band of the excitation transmission chain, and avoids the phase lag and corner accuracy loss caused by traditional lag filtering methods.

[0020] 2. Combining the dynamic acceleration envelope constraint of the controlled mechanism with the negative gradient analysis mechanism of the virtual potential energy field, when the correction requirement generated by the state deviation vector approaches the physical execution limit, the dynamic adjustment unit adjusts the spatial steepness of the potential energy distribution function in real time to smoothly distribute the high-frequency impact load in the time dimension, thereby achieving gradual energy dissipation for large-span speed jumps without exceeding the saturation limit of the motor output torque.

[0021] 3. Based on the inertia coupling matrix of the multi-axis linkage system, the multi-dimensional vector allocation of the transition correction increment is carried out, so that the axis with smaller inertia and faster dynamic response can share a higher proportion of the trajectory compensation. This asymmetric adjustment mechanism based on physical characteristics ensures the shape preservation of the center of gravity trajectory of the whole system during high-speed motion connection, and solves the spatial oscillation of the closed-loop system induced by the non-consistency of multi-axis response delay. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a functional architecture and instruction execution flow diagram of the multi-axis linkage control system of the present invention; Figure 2 This is a block diagram of the logic architecture of the multi-segment motion smoothing control system guided by the control potential energy of the present invention. Detailed Implementation

[0024] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0025] A control system based on smooth transition of multi-segment motion velocity, the system comprising:

[0026] The instruction look-ahead unit is used to acquire the motion instruction flow representing the path to be processed, and to extract the state deviation vector containing velocity vector difference, acceleration vector difference and jerk vector difference at the connection point of adjacent motion segments.

[0027] The parameter modulation unit is used to construct a control potential energy model representing the state energy distribution of each axis in the control state space that maps the multi-axis linkage coupling characteristics, based on the state deviation vector. The control potential energy model defines the spatial curvature distribution based on the inertia level of the controlled object, and is used to transform the path deviation in the geometric domain into the potential well constraint in the energy domain.

[0028] The dynamic adjustment unit is used to obtain the inertia coupling matrix of the drive execution unit, and calculate the transition correction increment characterizing the inter-axis dynamic compensation based on the gradient direction of the control potential energy model and the inertia coupling matrix. It converts the signal mutation caused by the state deviation vector into the guiding force information that evolves along the direction of potential energy decrease in the control state space.

[0029] The optimized output unit is used to load the guiding force information as an external constraint into the energy functional equation to be optimized. Based on the external constraint, the position deflection is adjusted in multiple dimensions according to the coordinates of the control points of the processing path. The output target spline parameters with first derivative continuity and dynamic smoothness are used to update the value of the underlying interpolation register of the drive execution unit, so as to pre-cancele the resonant impact generated by the controlled object in the physical domain within the information domain.

[0030] Preferably, when calculating the transition correction increment, the dynamic adjustment unit adopts the following steps: obtaining the dynamic envelope constraint characterizing the physical output limit of the driving execution unit; determining the compensation weight of each axis according to the distribution of the diagonal elements of the inertia coupling matrix, and allocating the correction requirements generated by the gradient direction to the corresponding linkage axis components according to the compensation weight; performing amplitude limiting processing on the allocated linkage axis components according to the dynamic envelope constraint to generate the transition correction increment, so that the guiding force information can achieve gradual energy dissipation of the state deviation vector without exceeding the physical saturation limit of the driving execution unit.

[0031] Preferably, when constructing the control potential energy model, the parameter modulation unit introduces a damping coefficient that characterizes the transmission chain characteristics of the control system. This coefficient is used to suppress high-frequency command signals by adjusting the spatial curvature distribution of the control potential energy model, thereby reducing the system resonance energy excited by the speed vector jump during the control command generation stage. The damping coefficient is dynamically selected based on the real-time speed fluctuation frequency of the drive execution unit to ensure that the control potential energy model can effectively filter out noise of different frequencies and ensure the steady-state operation accuracy of the multi-axis linkage system.

[0032] Preferably, the dynamic adjustment unit adjusts the iteration step size of gradient analysis in real time according to the magnitude of the state deviation vector. When the magnitude exceeds a preset threshold of 10mm or 5mm / s, the iteration step size of gradient analysis is reduced to improve the mapping accuracy of the guiding force information to the abrupt change of the motion command flow. The dynamic adjustment unit improves the ability to capture nonlinear features at the trajectory corner by shrinking the iteration step size, ensuring that the generated transition correction increment can accurately match the dynamic response characteristics of the controlled object and prevent the accumulation of position tracking errors in high-speed processing environment.

[0033] Preferably, the parameter modulation unit quantifies the strength of the control potential energy model by using a virtual energy value E defined in the state space by the state deviation vector. The formula for calculating the virtual energy value E is as follows: Where Δ is the state deviation vector, M is the inertia coupling matrix, and T represents the matrix transpose operation; the dynamic adjustment unit obtains the gradient direction of the control potential energy model by differentiating the virtual energy value E with respect to the displacement variable, which serves as the original driving source for the transition correction increment.

[0034] Preferably, when the optimization output unit adjusts the position deflection, it introduces the guiding force information as an external constraint load into the energy functional equation of the path to be processed. By driving the control point to deflect in the negative gradient direction of the energy functional, the total virtual energy of the control system reaches a minimum value. The optimization output unit iteratively searches for the optimal solution of the energy functional and achieves a smooth transition of the velocity vector by fine-tuning the local geometric topology while ensuring the shape preservation of the trajectory. Finally, it obtains the target spline parameters that satisfy the kinematic constraints.

[0035] Preferably, the system also includes a status monitoring unit, which is used to acquire the feedback pose signal of the drive execution unit in real time, and dynamically compensate the energy distribution gain in the parameter modulation unit through a preset proportional-integral gain function based on the dynamic tracking error between the feedback pose signal and the target spline parameters; the status monitoring unit ensures that the control potential energy model can be adaptively adjusted according to the actual mechanical load changes of the controlled object by constructing a closed-loop adjustment loop for the energy distribution gain, thereby improving the stability of the control system under variable operating conditions.

[0036] Preferably, the state deviation vector extracted by the command look-ahead unit characterizes the geometric topological discontinuity and physical temporal mismatch at the spatial connection point of adjacent motion command segments; among them, the velocity vector difference determines the potential well depth of the control potential energy model, the acceleration vector difference determines the gradient slope of the control potential energy model, and the jerk vector difference determines the rate of change of the guiding force information. Through multi-dimensional state combination, the control system can establish a complete physical impact quantification characterization in a multi-dimensional state space.

[0037] Preferably, the dynamic envelope constraints include the maximum output torque saturation limit of each motor in the drive execution unit, the maximum allowable centripetal acceleration of each transmission shaft in high-speed motion, and the vibration stiffness limit of each mechanical structure; the dynamic adjustment unit ensures that the generated transition correction increment is physically feasible by limiting the linkage shaft components within the convex space defined by the dynamic envelope constraints, thereby avoiding protective shutdown of the drive system or fatigue damage to the mechanical structure due to command overshoot.

[0038] Preferably, the target spline parameters output by the optimized output unit are directly mapped to the periodic sampling point data of the interpolation controller inside the controlled object, enabling the drive execution unit to gradually switch the speed direction and modulus based on the guidance of the control potential energy model when performing multi-segment path connection tasks. After the target spline parameters are updated to the hardware cache, the multi-axis linkage system is driven by high-frequency synchronous pulses to run according to the optimized smooth curve, thereby significantly reducing the mechanical wear and energy consumption of the transmission chain while ensuring processing efficiency.

[0039] Example 1: In a continuously operating 5-axis linkage high-viscosity fluid extrusion molding system, the drive execution unit of the controlled object needs trajectory stabilization adjustment. Local trajectory planning generates velocity vector deflection at the junction of adjacent motion segments, causing signal abrupt changes that trigger control loop resonance and exceed the frequency response bandwidth of the servo system. The low-pass filtering mechanism introduces phase lag and leads to loss of shape preservation at trajectory corners. The command look-ahead unit acquires the motion command flow characterizing the path to be processed and extracts the state deviation vector containing velocity vector difference, acceleration vector difference, and jerk vector difference at the junction of adjacent motion segments. The parameter modulation unit constructs a control potential energy model characterizing the state energy distribution of each axis within the control state space that maps the multi-axis linkage coupling characteristics, based on the state deviation vector. The control potential energy model defines the spatial curvature distribution based on the inertia magnitude of the controlled object. Simultaneously, the parameter modulation unit defines virtual energy values ​​within the state space through the state deviation vector. The strength of the quantified control potential energy model is determined by the formula for calculating the virtual energy value E. Where Δ is the state deviation vector, M is the inertia coupling matrix, and T is the matrix transpose operator; in this calculation model, in order to eliminate the problem of inconsistent physical dimensions caused by directly multiplying the higher-order derivative components of the state deviation vector with the inertia coupling matrix, the system pre-performs dimensional normalization processing on the state deviation vector. By introducing a diagonal scaling matrix containing the nominal maximum reference value corresponding to each component, the physical deviation containing velocity and acceleration information is transformed into a dimensionless state scalar, thereby providing the extracted virtual energy with a physically self-consistent unified mathematical scale.

[0040] Based on the principle of energy dissipation in structural dynamics, the parameter modulation unit analyzes and drives the real-time speed feedback signal of the execution unit through fast Fourier transform, extracting the main frequency with the largest amplitude fluctuation in the spectrum as the real-time speed fluctuation frequency. The parameter modulation unit is based on the damping attenuation formula. Real-time calculation of dimensionless damping coefficient ,in, To pre-determine the inherent resonant frequency limit of the transmission chain, the condition monitoring unit continuously collects the position deviation sequence between the actual mechanical trajectory and the target spline parameters, and inputs it into the discrete proportional-integral controller to calculate the dimensionless energy distribution gain parameter. The system will damping coefficient With energy distribution gain parameter Multiplication is used to scale the diagonal elements of the inertia coupling matrix M in real time, thereby targeting and weakening the potential energy of the system in a specific high-frequency band. This scaling operation is not intended to change the objective physical mass of the actuator, but rather to perform mapping calculations in the virtual admittance control algorithm running inside the underlying interpolation controller. By modifying the virtual inertia parameters used in the feedforward control loop, the computational impedance of the system in the information domain to high-frequency command mutations is effectively increased. The dynamic adjustment unit obtains the inertia coupling matrix M that drives the actuator, and, in conjunction with the gradient direction of the control potential energy model and the inertia coupling matrix, calculates the transition correction increment that characterizes the inter-axis dynamic compensation, converting the signal mutations caused by the state deviation vector into guiding force information that evolves along the direction of potential energy reduction in the control state space.

[0041] In this adjustment loop, the diagonal element distribution of the inertia coupling matrix determines the compensation weight of each axis, so that the correction demand generated by the gradient pointing is allocated to the corresponding linkage axis component according to the compensation weight, providing a physical boundary for the generation of the transition correction increment. The gradient pointing drives the linkage axis component to converge towards the direction of the minimum of the total virtual energy of the system. The coupling feedback of inertia allocation and potential field gradient changes the adjustment boundary of hysteresis filtering in the control algorithm dimension. The system obtains the dynamic envelope constraint representing the physical output limit of the drive execution unit and performs amplitude limiting processing on the allocated linkage axis component. Within the control architecture, it solves the constraints of instruction tracking and servo feedback bandwidth limitation. Based on the objective law that the electromagnetic torque of the servo motor decreases with increasing speed, it reads the two-dimensional numerical table of static torque and speed of each motor from the internal memory. In each position interpolation cycle, it uses the current actual speed of the spindle to interpolate and query the numerical table to obtain the transient maximum output torque limit of each axis. The combination of the output torque limit and the preset maximum allowable centripetal acceleration parameter constructs a dynamic polyhedron boundary enclosed by linear inequalities in the multi-dimensional control state space. When the correction component assigned to a specific linkage axis exceeds the boundary of the dynamic polyhedron, the system orthogonally projects the correction component along the boundary normal direction onto the nearest section of the polyhedron to achieve physical limiting. The optimized output unit loads the guiding force information as an external constraint term into the energy functional equation. Based on the external constraint term, the position deflection is adjusted according to the coordinates of the control point of the path to be processed, and the target spline parameter with first derivative continuity and dynamic smoothness is output. This target spline parameter is mapped to the periodic sampling point data of the interpolation controller inside the controlled object, and the value of the underlying interpolation register of the drive execution unit is updated. Based on this, the system cancels the resonance impact generated by the controlled object in the physical domain in the information domain, and achieves energy dissipation of the state deviation vector without exceeding the physical saturation limit of the drive execution unit.

[0042] Example 2: In high-dynamic multi-axis linkage trajectory processing, the servo frequency response bottleneck at the connection point of adjacent motion segments causes system resonance. This test was conducted to verify the effectiveness of the trajectory stabilization adjustment mechanism based on the potential energy model in dissipating resonant energy. The test platform adopted a five-axis linkage CNC test bench, equipped with a laser interferometer with a position resolution of 0.1μm to collect physical trajectory data. Gaussian white noise with a signal-to-noise ratio of 20dB was fed into the front end of the speed loop of the servo driver, and 50Hz power frequency interference harmonics were superimposed to construct a test benchmark that included electromagnetic disturbances in the industrial environment. The sampling period of the control loop was set to balance the real-time performance of data acquisition and the computational load of the interpolation controller. When the spectral bandwidth of the acceleration vector difference in the trajectory motion command stream increases, in order to prevent high-frequency signal aliasing distortion under the Nyquist sampling theorem, the sampling period was determined to tend towards the lower limit supported by the hardware clock. Based on this decision logic, the sampling period of the control loop in the high-dynamic extrusion scenario of this test was determined to be 125μs, providing a definite time benchmark for feature vector extraction.

[0043] The experiment was divided into a comparative sample group based on a low-pass filtering algorithm and an experimental sample group using the control potential energy model of the present invention. A spatial spline processing path containing a 90° right-angle turn was input to the command look-ahead unit of both systems, and the state deviation vector Δ at the turn point was extracted. The parameter modulation unit of the present invention calculated the virtual energy value E based on the acquired inertia coupling matrix M and the state deviation vector Δ, and analyzed the gradient direction of the control potential energy model in the state space. The dynamic adjustment unit output a transition correction increment to the interpolation register based on this gradient direction. Experimental data showed that at 50Hz… Under interference conditions, the drive unit of the comparison sample excited a high-frequency oscillation with an amplitude of 15.4 μm after the turning point. Due to the phase lag of the filter component, the actual trajectory profile deviated from the theoretical path by 8.2 μm. The sample of this invention guided the convergence of the direction of the minimum virtual energy of the system of each linkage axis through the transition correction increment. The measured oscillation amplitude at the turning point converged to 1.8 μm, and the maximum profile deviation was maintained at 2.1 μm. The above data show that the constructed virtual energy feedback loop can suppress the resonance impact excited by power frequency disturbance and speed change without introducing time domain lag.

[0044] To define the boundary of the transition correction increment, a gradient test benchmark was set with the state deviation vector magnitude increasing from 1 mm / s to 30 mm / s to examine the nonlinear dynamic response limit of the system. Observation results show that when the state deviation vector magnitude is in the range of 1 mm / s to 15 mm / s, the trajectory tracking error of the sample group of this invention increases linearly with the magnitude and remains within 3.5 μm. When the magnitude exceeds 15 mm / s, the error growth rate changes abruptly. When the magnitude reaches 25 mm / s, the tracking error climbs to 12.7 μm, and a performance degradation inflection point appears. This phenomenon maps to the physical saturation mechanism of the torque current inside the drive execution unit, indicating that when the virtual energy generated by the local command jump exceeds the electromagnetic power limit of the motor, the potential energy compensation in the information domain cannot offset the energy overflow in the physical domain. The above gradient data establishes the physical working window of the system through smooth adjustment by the potential energy model, confirming that the transition correction increment set according to the inertia coupling matrix limits the multi-axis transient error within the available response bandwidth of the electromechanical system, and resolves the technical contradiction between high-frequency command tracking and mechanical resonance in multi-axis linkage control.

[0045] Example 3: In a high-dynamic fluid multi-axis extrusion molding process, the controlled object experiences trajectory direction reversal during the processing of an acute-angle trajectory. This causes the magnitude of the state deviation vector extracted at the connection point of adjacent motion segments to exceed the linear tracking limit. The fixed gradient analytical step size triggers the search process to diverge at the bottom of the control potential energy model, instructing the look-ahead unit to extract the state deviation vector. The dynamic adjustment unit calculates the current modulus of the state deviation vector in both displacement and velocity dimensions. It then compares this current modulus with the spatial threshold of 10mm and the velocity threshold of 5mm / s, representing the physical following limit of the device. When the current modulus exceeds either the spatial or velocity threshold, the dynamic adjustment unit activates iterative step-size contraction control. The parameter modulation unit, based on the inertia coupling matrix M of the drive execution unit and the state deviation vector Δ, calculates the current modulus according to the formula... The system calculates the virtual energy value E in the current control state space, where Δ is the state deviation vector, M is the inertia coupling matrix, and T is the matrix transpose operator. The dynamic adjustment unit differentiates the virtual energy value E with respect to the displacement variable to obtain the gradient direction of the control potential energy model. When calculating this partial derivative, the system pre-extracts the geometric coordinate system mapping function of the path to be processed in the CNC machining code, establishes the Jacobian matrix between the state deviation vector at the discrete time sampling point and the spatial displacement variable at the execution end, and then uses the chain rule to transform the potential energy gradient for higher-order kinematic deviations into the gradient vector of the displacement of the control point in the interpolation trajectory space. The dynamic adjustment unit multiplies the initial iteration step size by the decay factor to generate the current iteration step size. The decay factor is set as the reciprocal of the ratio of the current modulus to the corresponding over-limit threshold. This dimensionless decay factor causes the deviation vector that deviates from the physical limit to shrink the corresponding search step size.

[0046] The optimization output unit updates the coordinates of the multi-axis control points along the negative direction of the gradient with the current iteration step size. Based on the updated coordinates, it re-extracts the state deviation vector to calculate the virtual energy value for the next round. When the difference between the virtual energy values ​​of two consecutive iterations is less than the energy threshold of 0.05J, which characterizes the steady-state fluctuation of the system, the optimization output unit terminates the iteration. The coordinate update amount accumulated in a single iteration is determined as the transition correction increment. The transition correction increment converts the signal mutation caused by the state deviation vector into guiding force information that evolves along the direction of potential energy decrease in the control state space. The optimization output unit introduces the guiding force information as an external constraint load into the energy functional equation of the path to be processed and outputs the target spline parameters. The system converts the local trajectory change kinetic energy into command displacement compensation through an adaptive decay iterative solution procedure, so that the generated correction data is adapted to the dynamic response characteristics of the controlled object and the position tracking error is limited to the physical following tolerance range of the servo mechanism.

[0047] Example 4: When the system faces the deployment of a new line for a heterogeneous multi-axis extrusion equipment, the parameter modulation unit initiates an unloaded self-test procedure for the underlying dynamic characteristics before planning the trajectory. The test module sequentially injects a sweeping sinusoidal excitation current with a frequency linearly increasing from 1Hz to 500Hz into the front end of the speed loop of the servo driver. The signal acquisition module synchronously reads the angular acceleration response and torque fluctuation data fed back by the encoders of each axis. The system analyzes the rotational inertia of the independent axis and the cross-coupling inertia between the linkage axes based on the transfer function relationship between the input current and the output angular acceleration. The system writes the obtained rotational inertia value into the diagonal position of the inertia coupling matrix M, and writes the cross-coupling inertia value into the off-diagonal position of the inertia coupling matrix M. The parameter modulation unit solidifies the inertia coupling matrix M containing the dynamic attributes of the multi-axis into the non-volatile memory, establishing the basic curvature distribution of the control potential energy model in the state space.

[0048] After solidifying the inertia coupling matrix M, the dynamic adjustment unit initiates the boundary determination program for the multi-axis linkage physical limit. The control loop sends a step-increment test step command to the drive execution unit. The laser interferometer captures the actual trajectory contour of the controlled object's end in the physical space in real time. The dynamic adjustment unit calculates the follow-up deviation between the command coordinates and the actual trajectory contour, and records the current transient motion parameters when the follow-up deviation reaches the preset follow-up tolerance limit. The system locks the velocity vector difference magnitude corresponding to the transient motion parameter as the velocity threshold that triggers the iterative step size contraction logic, and locks the corresponding displacement vector difference magnitude as the spatial threshold. The relevant threshold parameters are directly used as the underlying comparison benchmark of the trajectory controller. The system synchronizes the dynamic parameters and control state space data, and enters the fully automatic standby state.

[0049] Example 5: When the system faces the condition of analyzing the coordinates of discrete spline control points using a multi-axis linkage interpolation controller, the optimization output unit extracts the guiding force information from the gradient output of the dynamic adjustment unit based on the control potential energy model, and constructs an energy functional equation in the built-in digital signal processor. The energy functional equation is composed of the linear superposition of an internal tension potential energy term characterizing the geometric bending characteristics of the spline curve and an external work term characterizing dynamic correction and compensation. The optimization output unit calculates the total energy algebraic value of the local trajectory interval based on this energy functional equation. The calculation formula is: ,in, This represents the algebraic value of the total energy. This refers to the virtual stiffness coefficient set by the system based on the frequency response characteristics of the controlled object's foundation machinery. The geometric deformation parameter is characterized by the second-order sum of squared differences in the position coordinates of adjacent multi-axis control points. This refers to the equivalent force components of the guiding force information decomposed on the corresponding control axis. This optimized output unit represents the local positional deflection of the current control point coordinates relative to the nominal displacement, so that the total energy is algebraically represented. With minimization as the objective, a system of nonlinear algebraic equations is established along the discretized sampling interval of the path to be processed.

[0050] The optimized output unit synchronously acquires the real-time spindle feed rate of the current sampling period, compares this real-time spindle feed rate with the rated dynamic extreme value configured at the underlying level, and dynamically adjusts the virtual stiffness coefficient. The weight of the virtual stiffness coefficient monotonically decreases as the real-time spindle feed rate approaches the physical saturation limit. To expand the search domain for solving local position deflection, the optimized output unit uses the conjugate gradient algorithm to iteratively solve the nonlinear algebraic equations. When the difference between the coordinate update vectors generated in two consecutive iterations is less than 0.05 μm, it is determined that the solution has reached the energy convergence state. The coordinate convergence value at this time is extracted to generate the corrected target spline parameters. The target spline parameters contain the position command sequence of each linkage axis synchronized with the time base of the underlying servo drive. Based on this, the system converts discrete signal jumps into smooth displacement compensation commands in the information control domain, and smooths out the physical oscillations caused by the sudden change of velocity vector in the spatial geometric dimension.

[0051] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A control system based on smooth transition of multi-segment motion speed, characterized in that, The system includes: The instruction look-ahead unit is used to acquire the motion instruction flow representing the path to be processed, and to extract the state deviation vector containing velocity vector difference, acceleration vector difference and jerk vector difference at the connection point of adjacent motion segments. The parameter modulation unit is used to construct a control potential energy model representing the state energy distribution of each axis in the control state space that maps the multi-axis linkage coupling characteristics, based on the state deviation vector. The control potential energy model defines the spatial curvature distribution based on the inertia level of the controlled object, and is used to transform the path deviation in the geometric domain into the potential well constraint in the energy domain. The dynamic adjustment unit is used to obtain the inertia coupling matrix of the drive execution unit, and calculate the transition correction increment characterizing the inter-axis dynamic compensation based on the gradient direction of the control potential energy model and the inertia coupling matrix. It converts the signal mutation caused by the state deviation vector into the guiding force information that evolves along the direction of potential energy decrease in the control state space. The optimized output unit is used to load the guiding force information as an external constraint into the energy functional equation to be optimized. Based on the external constraint, the position deflection is adjusted in multiple dimensions according to the coordinates of the control points of the processing path. The output target spline parameters with first derivative continuity and dynamic smoothness are used to update the value of the underlying interpolation register of the drive execution unit, so as to pre-cancele the resonant impact generated by the controlled object in the physical domain within the information domain.

2. The control system based on smooth transition of multi-segment motion speed according to claim 1, characterized in that, When calculating the transition correction increment, the dynamic adjustment unit adopts the following steps: obtaining the dynamic envelope constraint that characterizes the physical output limit of the driving execution unit; determining the compensation weight of each axis according to the distribution of the diagonal elements of the inertia coupling matrix, and allocating the correction requirements generated by the gradient direction to the corresponding linkage axis components according to the compensation weight; performing amplitude limiting processing on the allocated linkage axis components according to the dynamic envelope constraint to generate the transition correction increment, so that the guiding force information can achieve gradual energy dissipation of the state deviation vector without exceeding the physical saturation limit of the driving execution unit.

3. The control system based on smooth transition of multi-segment motion speed according to claim 1, characterized in that, When constructing the control potential energy model, the parameter modulation unit introduces a damping coefficient that characterizes the transmission chain of the control system. This coefficient is used to suppress high-frequency command signals by adjusting the spatial curvature distribution of the control potential energy model, thereby reducing the system resonance energy excited by the speed vector jump during the control command generation stage. The damping coefficient is dynamically selected based on the real-time speed fluctuation frequency of the drive execution unit to ensure that the control potential energy model can effectively filter out noise of different frequencies and ensure the steady-state operation accuracy of the multi-axis linkage system.

4. The control system based on smooth transition of multi-segment motion speed according to claim 1, characterized in that, The dynamic adjustment unit adjusts the iteration step size of gradient analysis in real time according to the magnitude of the state deviation vector. When the magnitude exceeds a preset threshold of 10mm or 5mm / s, the iteration step size of gradient analysis is reduced to improve the mapping accuracy of the guiding force information to the abrupt change of the motion command flow. The dynamic adjustment unit improves the ability to capture nonlinear features at trajectory corners by shrinking the iteration step size, ensuring that the generated transition correction increment can accurately match the dynamic response characteristics of the controlled object and prevent the accumulation of position tracking errors in high-speed machining environments.

5. The control system based on smooth transition of multi-segment motion speed according to claim 1, characterized in that, When the optimized output unit adjusts the position deflection, it introduces the guiding force information as an external constraint load into the energy functional equation of the path to be processed. By driving the control point to deflect in the negative gradient direction of the energy functional, the total virtual energy of the control system reaches a minimum value. The optimized output unit iteratively searches for the optimal solution of the energy functional and achieves a smooth transition of the velocity vector by fine-tuning the local geometric topology while ensuring the shape preservation of the trajectory. Finally, it obtains the target spline parameters that satisfy the kinematic constraints.

6. The control system based on smooth transition of multi-segment motion speed according to claim 1, characterized in that, The system also includes a status monitoring unit, which is used to acquire the feedback pose signal of the drive execution unit in real time, and dynamically compensate the energy distribution gain in the parameter modulation unit through a preset proportional-integral gain function based on the dynamic tracking error between the feedback pose signal and the target spline parameters. The status monitoring unit ensures that the control potential energy model can be adaptively adjusted according to the actual mechanical load changes of the controlled object by constructing a closed-loop adjustment loop for the energy distribution gain, thereby improving the stability of the control system under variable operating conditions.

7. The control system based on smooth transition of multi-segment motion speed according to claim 1, characterized in that, The state deviation vector extracted by the command look-ahead unit characterizes the geometric topological discontinuity and physical temporal mismatch at the spatial connection point of adjacent motion command segments. Among them, the velocity vector difference determines the potential well depth of the control potential energy model, the acceleration vector difference determines the gradient slope of the control potential energy model, and the jerk vector difference determines the rate of change of the guiding force information. Through multi-dimensional state combination, the control system can establish a complete physical impact quantification characterization in a multi-dimensional state space.

8. The control system based on smooth transition of multi-segment motion speed according to claim 2, characterized in that, The dynamic envelope constraints include the maximum output torque saturation limit of each motor in the drive execution unit, the maximum allowable centripetal acceleration of each transmission shaft under high-speed motion, and the vibration stiffness limit of each mechanical structure. The dynamic adjustment unit ensures that the generated transition correction increment is physically feasible by limiting the linkage shaft components within the convex space defined by the dynamic envelope constraints, thus avoiding protective shutdown of the drive system or fatigue damage to the mechanical structure due to command overshoot.

9. The control system based on smooth transition of multi-segment motion speed according to claim 1, characterized in that, The target spline parameters output by the optimized output unit are directly mapped to the periodic sampling point data of the interpolation controller inside the controlled object. This enables the drive execution unit to gradually switch the speed direction and magnitude based on the guidance of the control potential energy model when performing multi-segment path connection tasks. After the target spline parameters are updated to the hardware cache, the multi-axis linkage system is driven by high-frequency synchronous pulses to run according to the optimized smooth curve, thereby reducing mechanical wear and energy consumption of the transmission chain while ensuring processing efficiency.

Images