A control system for adjusting the running speed of a transfer robot based on load

Abstract

The application relates to the technical field of general automatic control systems, and discloses a control system for adaptively adjusting the running speed of a transfer robot based on a load, which comprises a core controller module and an inertia mirror adaptive law module; the inertia mirror adaptive law module is used for calculating an adaptive gain coefficient according to a basic control effort signal and a speed error, so as to adjust the control parameters of the core controller module; the system is further provided with an error signal discrimination module and an adaptive law arbitration module; when dynamic disturbance or a physical actuator saturation state is monitored, the adaptive gain coefficient of the inertia mirror adaptive law module is inhibited from being updated; the application realizes adaptive adjustment of load changes by using internal information of the control system, and ensures logical self-consistency through a double inhibition mechanism, thereby solving the inherent contradiction between responsiveness and stability in model time variation, and enabling the system to maintain consistent running characteristics under different loads.

Description

Technical Field

[0001] This invention relates to a control system for adaptively adjusting the operating speed of a transfer robot based on load, belonging to the field of general automatic control system technology. Background Technology

[0002] Currently, classical feedback controllers relying on fixed control parameters are widely used due to their clear structure and ease of implementation. For systems where the characteristics of the controlled object are basically constant, such controllers can provide stable and effective regulation. However, in many practical tasks, the system model of the controlled object is not constant but undergoes drastic real-time changes. For example, when a transport robot switches between no-load and full-load states, its system inertia changes by orders of magnitude. This dramatic change in inertia fundamentally alters the dynamic response characteristics of the controlled object itself. This exposes an inherent performance constraint of the aforementioned classical control method relying on fixed parameters: if a set of conservative control parameters is set to cope with the maximum load condition, the system will exhibit slow response and overly conservative regulation under light load or no load, greatly limiting operating efficiency. Conversely, if a set of more aggressive parameters is set to pursue a fast response under light load, the same set of parameters will immediately cause the control system to become unstable when the system load suddenly increases to full load, resulting in severe speed overshoot and oscillation. This is an unacceptable operational risk for precise operating environments that require smooth start and stop.

[0003] To address this issue, existing technologies have proposed various adaptive control schemes. However, these schemes often have inherent logical flaws, making it difficult to ensure system stability and logical consistency under complex operating conditions. For example, Chinese invention patent CN119024678B discloses an adaptive control method and system for a load-driven front fork shock absorber. The core idea of ​​this scheme is to establish a nominal model of the controlled object and identify the deviation between the actual output and the theoretical output of the model as a disturbance for compensation. However, the fundamental flaw of this method is that it fails to distinguish between model parameter changes (such as changes in load inertia) and errors caused by external environmental disturbances from a mechanistic perspective. It categorizes all unmodeled dynamic or time-varying characteristics of the controlled object as external disturbances, resulting in a lack of logical precision in its adaptive adjustment. When the system faces both parameter changes and external disturbances simultaneously, its control logic is prone to confusion. It may misjudge the actual model changes as disturbances and suppress the adjustment, or incorrectly absorb the external disturbances into the parameter adaptation, thereby causing deterioration or even instability in control performance.

[0004] Therefore, the technical problem to be solved by this invention is how to provide a control system that can use the operating information inside its control loop to continuously calculate the inertial changes of the controlled object in real time, and establish a continuous adaptive adjustment mechanism based on the calculation results, so that the system can still autonomously balance the speed of response and the stability of operation when the model changes drastically. Summary of the Invention

[0005] This invention provides a control system for adjusting the operating speed of a transfer robot based on load adaptive adjustment. Its main purpose is to solve the problem that existing control systems lack a mechanism to use internal operating information to calculate the inertial changes of the controlled object, which leads to the inability to take into account both responsiveness and stability when the model is time-varying.

[0006] To achieve the above objectives, the present invention provides a control system for adaptively adjusting the operating speed of a transfer robot based on load, the system comprising:

[0007] A core controller module is configured to generate basic control effort signals to the physical actuators of the transfer robot based on the speed error between the speed setpoint and the speed measurement value, and according to a set of adjustable control parameters.

[0008] An inertial mirror adaptive law module is configured to calculate an adaptive gain coefficient in real time based on the basic control effort signal and velocity error, and use the adaptive gain coefficient to adjust the adjustable control parameters of the core controller module.

[0009] An error signal identification module is configured to: analyze a frequency domain feature of the velocity error in real time; and when the frequency domain feature matches a preset dynamic disturbance signature, suppress the inertial mirror adaptive law module to prevent it from updating the adaptive gain coefficient.

[0010] An adaptive law arbitration module is configured to: monitor whether the physical actuator enters a saturation state; and suppress the operation of the inertial mirror adaptive law module during the period when the physical actuator is in a saturation state, so as to prevent it from updating the adaptive gain coefficient.

[0011] Preferably, the inertial mirror adaptive law module is configured to: monitor the integral value of the basic control effort signal within a first preset time window; monitor the rate of change of the velocity error within a second preset time window; calculate the adaptive gain coefficient based on the ratio between the integral value and the rate of change; and freeze and maintain the last generated adaptive gain coefficient until the velocity error exceeds the steady-state threshold again when the velocity error continues to be below a preset steady-state threshold.

[0012] Preferably, the system further includes: a transient detection module configured to: respond to a step change in the speed setpoint from a steady-state value; and inject a preset detection pulse into the physical actuator as a basic control effort signal before the core controller module responds to the step change; and trigger an inertial mirror adaptive law module to calculate and set an initial adaptive gain coefficient using the control effort of the preset detection pulse and the resulting change in speed error; the core controller module is further configured to use the initial adaptive gain coefficient as an adjustable control parameter to respond to the step change.

[0013] Preferably, the transient detection module is further configured to: immediately instruct the transfer robot to run at a preset constant low speed for a preset period of time after executing a preset detection pulse; and monitor the average base control effort required to maintain the preset constant low speed within the preset period of time to determine a damping mirror parameter; wherein the core controller module is configured to adjust its proportional gain and integral gain using an initial adaptive gain coefficient, and is further configured to independently adjust its derivative gain using the damping mirror parameter.

[0014] Preferably, the system further includes: a setpoint smoothing module configured to: receive a speed setpoint as an original speed setpoint and generate a smoothed speed setpoint; the core controller module is further configured to use the smoothed speed setpoint instead of the original speed setpoint to calculate the speed error; wherein the smoothing degree of the setpoint smoothing module is adjusted by the adaptive gain coefficient generated by the inertial mirror adaptive law module; such that the smoothing degree is enhanced as the adaptive gain coefficient, which represents the inertia of the transfer robot, increases.

[0015] Preferably, the system further includes: a transient spike suppression module configured to: receive the original speed measurement signal stream; apply medium-range filtering logic to the original speed measurement signal stream to generate purified speed measurement values; and provide the purified speed measurement values ​​as speed measurement values ​​to the core controller module for calculating speed errors.

[0016] Preferably, the dynamic disturbance signature is defined as the energy of the velocity error being concentrated in a preset disturbance characteristic frequency band; the error signal identification module uses a digital bandpass filter to monitor whether the velocity error conforms to the dynamic disturbance signature.

[0017] Preferably, saturation state refers to the state in which the actual output value of the physical actuator reaches its output limit, resulting in a discrepancy between the basic control effort signal and the actual output value of the physical actuator; the adaptive law arbitration module monitors whether the physical actuator has entered saturation state by comparing the basic control effort signal with the actual output value of the physical actuator.

[0018] Preferably, the core controller module is a PID controller; the adjustable control parameters include the proportional gain of the PID controller. With integral gain The inertial image adaptive law module is configured to determine the proportional gain based on the adaptive gain coefficients using a pre-defined nonlinear function lookup table. With integral gain The value.

[0019] Preferably, the transient detection module is configured to respond to a step change in the speed setpoint from zero after the transfer robot is stationary and its load changes, thereby completing the calculation and setting of the initial adaptive gain coefficient before the transfer robot begins to perform acceleration.

[0020] Compared with the prior art, the beneficial effects of the present invention are:

[0021] 1. This scheme establishes an information utilization path within the control system. It retrieves the control effort signal and speed error signal generated by the controller itself, and generates an index reflecting the inertial change of the controlled object by analyzing the correlation between the two in the dynamic process. Based on this index, the parameters of the core controller are continuously adjusted through a preset nonlinear relationship. This adjustment mechanism also incorporates the judgment of steady-state conditions. When the system enters stable operation, it maintains the current parameters. This approach enables the control logic to recognize the model changes of the controlled object on its own, and ensures that the response characteristics and stability of the system remain consistent under different loads without relying on external physical measurements.

[0022] 2. By using a transient detection module to initiate a startup, a standard detection pulse is injected into the controlled object before responding to a step speed command. The inertial image adaptive law module is then invoked to analyze the response generated by this pulse, thereby calculating an initial adaptive gain coefficient. This design reuses the original reactive control mechanism as an active pre-calibration at startup. This ensures that the core controller's control parameters match the current true inertia of the controlled object in the first cycle of the startup action, thus avoiding startup sluggishness or shock problems caused by frozen steady-state parameters. Simultaneously, this method… The case also reuses the internal information representing the inertia of the controlled object generated by the inertial mirror adaptive law, using it as a dynamic adjustment parameter and inputting it into a setpoint smoothing module located at the front end of the core controller. When the internal information shows that the inertia of the controlled object increases, the smoothing degree of the smoothing module is automatically enhanced. This structure feeds the control insights from the lower layer to the instruction planning of the upper layer, so that when the system senses that its own load is heavy, it can actively soften the original speed instructions it receives, thereby avoiding the physical impact caused by executing high dynamic instructions under high inertia outside the control loop.

[0023] 3. The control system also has the ability to distinguish between its own operating boundaries and external disturbances. On the one hand, it identifies the error introduced by external dynamic disturbances by analyzing the frequency domain characteristics of the speed error signal, and suppresses the increase of adaptive gain during identification to cut off the positive feedback of the control logic that is misled by the disturbance. On the other hand, it also monitors the saturation state of the physical actuators. When the actuators reach their output limits, it temporarily stops updating the adaptive law. This dual suppression mechanism ensures that the core inertial mirror logic only operates under conditions where the information is real and valid, avoiding the system from generating erroneous adaptive adjustments in nonlinear regions or under strong disturbance environments, thereby maintaining the stability of the control system. Attached Figure Description

[0024] Figure 1 This is a functional block diagram of the adaptive control system and its dual suppression logic of the present invention;

[0025] Figure 2 This is a comparison chart of the speed response of the system of the present invention under different loads and that of a fixed PID controller;

[0026] Figure 3 This is a timing diagram of the adaptive suppression function under dynamic disturbance conditions of the present invention. Detailed Implementation

[0027] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below. It should be noted that the described embodiments are only some embodiments of this invention, not all embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0028] This invention discloses a control system for adjusting the operating speed of a transport robot based on load adaptive adjustment. Within this control framework, it is constructed as a closed-loop adjustment system comprising a core controller module, an inertial image adaptive law module, an error signal identification module, and an adaptive law arbitration module. The system operates by having the inertial image adaptive law module analyze internal information during the control process in real time to generate an adaptive gain coefficient to continuously adjust the control parameters of the core controller module. Simultaneously, the error signal identification module and the adaptive law arbitration module act as monitoring logic, respectively monitoring dynamic disturbances and physical actuator saturation, and applying adjustment measures to the inertial image adaptive law module when these conditions occur. Specific suppression is added to ensure the coordination and stable operation of the entire control system when the controlled object model is time-varying or encounters external disturbances. In a specific implementation, the core controller module is a digital proportional-integral-derivative (PID) controller used in the field, such as an algorithm implementation deployed on a microcontroller (MCU). Within a control cycle, this module first receives the speed setpoint from the upper-level system and obtains the speed measurement value from the speed measurement unit (such as a motor encoder) of the controlled object (i.e., the transfer robot). The PID controller obtains the current speed error by subtracting the two values. Subsequently, the PID controller uses a set of adjustable control parameters, specifically the proportional gain. Integral gain And in some implementations, differential gain is also included. The speed error is subjected to proportional, integral, and differential operations, and the results are superimposed to generate a digital basic control effort signal, such as the duty cycle setting of a pulse width modulation (PWM) signal. This signal is then sent to the physical actuators of the transfer robot, such as motor drivers, to drive the controlled object to adjust its running speed.

[0029] The inertial image adaptive law module's main function in the control system is to provide the core controller module with the ability to cope with real-time changes in the controlled object's model parameters (i.e., inertia). To achieve this function, the module is configured to continuously monitor the core controller module's output (i.e., the basic control effort signal) and input (i.e., velocity error), and calculate the controlled object's inertial state based on the dynamic relationship between the two. In a specific implementation, the module's internal procedure is as follows: First, within a continuous control cycle, the module calculates the integral value of the basic control effort signal over a first preset time window (e.g., 10 control cycles). This integral value characterizes the controller's efforts to eliminate errors. The module performs the following steps: First, it calculates the control effort required to correct the speed error. Second, it synchronously calculates the rate of change of the speed error within a second preset time window (e.g., 10 synchronized control cycles). Third, it calculates an adaptive gain coefficient based on the ratio between the integral value and the rate of change. This ratio, in the control logic, represents the physical inertia of the controlled object; a high ratio corresponds to high inertia, and a low ratio corresponds to low inertia. Fourth, it uses this ratio to look up a preset nonlinear function lookup table to determine a corresponding adaptive gain coefficient. Finally, it uses this adaptive gain coefficient to adjust the control parameters of the core controller module, for example, by adjusting the new... Setting as a base The value is the product of the adaptive gain coefficient; furthermore, to handle steady-state conditions in the system, the module is also configured to monitor the absolute value of the speed error. When the speed error continues to be below a preset steady-state threshold (e.g., below 0.5% of the target speed), the module will freeze and hold the last generated adaptive gain coefficient to avoid gain jitter caused by instability in the denominator (error change rate) when the error is close to zero, until the speed error exceeds the steady-state threshold again before resuming calculation.

[0030] To address the difficulty in distinguishing between time-varying models and external disturbances in control systems, and to prevent the system from incorrectly attributing errors caused by non-inertial factors (such as dynamic load swaying or road bumps) to inertial changes, this system introduces an error signal identification module. This module operates in parallel with the inertial mirror adaptive law module, but its task is to analyze a frequency domain characteristic of the velocity error signal stream in real time. In one specific implementation, this module achieves this function through a digital bandpass filter. The passband of this filter is set to a preset disturbance characteristic frequency band, for example, corresponding to 1Hz to 3Hz, the typical swaying frequency of the liquid load on the transport robot. The module continuously monitors the energy of the error signal passing through this filter. When the energy exceeds a preset threshold, it indicates that the velocity error energy is concentrated in this characteristic frequency band, conforming to a preset dynamic disturbance signature. In this case, the module immediately performs its suppression function, outputting a suppression signal to the inertial mirror adaptive law module. Specifically, this prevents the module from updating, and particularly increasing, the adaptive gain coefficient based on this disturbance-affected error signal, thereby cutting off any potential disturbance-gain positive feedback path in the system.

[0031] Similarly, to address the mismatch between the linear control logic and the nonlinear boundary (i.e., saturation) of the physical actuator in the control system, this system also includes an adaptive law arbitration module. This module monitors whether the physical actuator has entered its physical limit, i.e., saturation. In a specific implementation, saturation is defined as the physical actuator's (e.g., motor driver) actual output value, such as actual output current or PWM duty cycle, reaching its output limit, causing a discrepancy between the basic control effort signal issued by the core controller module and the actual output value of the physical actuator. The adaptive law arbitration module compares the values ​​of these two signals in real time, and once this discrepancy is detected, it determines that the system has entered saturation. During the entire period when the physical actuator is in saturation, the arbitration module performs its suppression function, temporarily suspending the operation of the inertial mirror adaptive law module and preventing it from updating the adaptive gain coefficients. This procedure ensures that the adaptive logic of the inertial mirror is only allowed to operate when the control system is operating in its linear region, i.e., when its internal information (effort and error) is valid, thereby avoiding the gain integral accumulation problem caused by actuator saturation and improving the system's stability after the extreme conditions are removed.

[0032] To address the potential start-up blind zone issue that may occur during steady-state freeze in the inertial mirror adaptive law module—specifically, when the load on the controlled object (transfer robot) changes from idle to full load while the controller maintains its old gain—this system further includes a start-up transient detection module. This module's operation involves transforming the original reactive adaptive mechanism into an active pre-parameter calibration upon receiving a start command. Specifically, when the module detects a step change in the speed setpoint from its steady-state value (e.g., zero), it temporarily prevents the core controller module from immediately responding to this step. Before responding, it injects a preset, standardized probe pulse (e.g., a PWM signal with a 20% duty cycle and a duration of 50 milliseconds) into the controlled object via a physical actuator as the basic control effort signal. Simultaneously, this module triggers the inertial mirror adaptive law module, causing it to analyze the control effort of the probe pulse and the resulting instantaneous speed error change, and, based on the aforementioned ratio calculation logic, instantaneously... An initial adaptive gain coefficient is calculated and set, which represents the current true inertia. Subsequently, the core controller module is allowed to use this newly calibrated initial adaptive gain coefficient as its adjustable control parameter to execute the step start command. Furthermore, in a more in-depth implementation, the start transient detection module is also configured to execute a procedure for separating model parameters. That is, after executing the aforementioned detection pulse to identify inertia, it immediately instructs the transport robot to run at a preset constant low speed, such as 0.1 m / s, for a preset period of time, such as 200 milliseconds. During this constant low-speed operation, the system is in a steady state (acceleration approximately zero). The module then monitors and calculates the average base control effort required to maintain this speed. According to control principles, this steady-state effort value is independent of inertia and mainly used to overcome system friction and damping; therefore, this value is determined as a damping mirror parameter. Finally, the core controller module is configured to adjust its proportional gain using the first value (initial adaptive gain coefficient). With integral gain (Primarily corresponding to inertia), and independently uses a second value (damped mirror parameter) to adjust its differential gain. (Mainly corresponds to damping).

[0033] This invention also utilizes internal system control information for adjusting upper-level commands. To this end, the system further includes a setpoint smoothing module. This module, located in the control link before the core controller module, receives the planned speed setpoint from the upper layer as an initial speed setpoint and outputs a smoothed speed setpoint to the core controller module. The core controller module then uses this smoothed value to calculate the speed error. This module is typically implemented as a digital low-pass filter, and the smoothing degree of this filter, for example, its time constant... It is not fixed, but dynamically adjusted by the adaptive gain coefficient generated by the inertial mirror adaptive law module, i.e., the inertial representation; its adjustment logic is determined that the smoothness increases as the inertia of the transfer robot, represented by the adaptive gain coefficient, increases, for example, the time constant. The adaptive gain coefficient increases with increasing values, and vice versa. This feedforward-feedback cooperative configuration allows the system to smooth any sharp speed step command into a gentle acceleration curve when it senses heavy load inertia, thus avoiding the physical shock risk of executing high-dynamic commands under high inertia at the command source. Finally, to ensure the numerical stability of the ratio algorithm of the inertial mirror adaptive law module, especially to avoid its susceptibility to transient noise spikes in the measurement signal, this system also includes a transient spike suppression module. This module is located at the input of the information stream and is configured to receive the raw speed measurement signal stream from the speed measurement unit (such as the encoder). The module uses a median filter logic instead of an average or low-pass filter because median filtering can effectively filter out single, abnormally large noise spikes without introducing significant phase delay. Its implementation involves the module maintaining a list of N (e.g., The latest raw speed measurement values ​​are displayed in a sliding window. In each control cycle, the five data points in the window are sorted, and only the median of the sorted values ​​is output as the purified speed measurement value. This purified speed measurement value is then provided to the core controller module for calculating the speed error.

[0034] Example 1: In a high-density warehousing and logistics scenario, a transfer robot is scheduled to execute a series of high-frequency start-stop and dynamic load tasks. This scenario places coupled requirements on the stability and responsiveness of the control system. Initially, the transfer robot is in an unloaded and stationary state, and the adaptive gain coefficient maintained by its core controller module corresponds to the unloaded model. At this time, a heavy-loaded pallet is placed on the stationary robot, causing a change in its controlled object model (inertia). Immediately afterwards, the system receives a speed step command starting from zero, i.e., a non-steady-state speed setpoint. Under this specific working condition, transient detection is initiated. The detection module is triggered first. This module temporarily prevents the core controller module from responding to step changes and, before responding, injects a preset probe pulse into the physical actuators of the transfer robot as a basic control effort signal. The control effort of this probe pulse, and the slight speed error change caused by the high inertial load, are analyzed in real time by the inertial image adaptive law module. This analysis is based on the ratio of control cost to error correction effect, and an initial adaptive gain coefficient corresponding to the current load state is calculated and set instantaneously. Only then is the core controller module allowed to start and uses this newly updated initial adaptive gain coefficient to adjust its adjustable control parameters (i.e., and As a result, the initial acceleration of the transport robot under heavy load was executed smoothly, avoiding the sluggish start-up or control instability that might have been caused by using outdated (unloaded) gain parameters. Subsequently, the heavily loaded transport robot traveled on uneven ground, causing dynamic disturbances caused by external vibrations to be mixed into the controlled object model. This disturbance manifested as a high-frequency speed error signal in the control loop. At this time, the inertial mirror adaptive law module, during operation, tended to interpret this high-frequency error as inertial mismatch and tended to increase the adaptive gain coefficient. Simultaneously, the error signal identification module worked in parallel. Through its internal digital bandpass filter, it analyzed the frequency domain characteristics of the speed error in real time and detected that the energy of the error signal was concentrated in the frequency band that conformed to the preset dynamic disturbance signature. The error signal identification module immediately output a suppression signal to prevent the inertial mirror adaptive law module from further increasing the adaptive gain coefficient. This logical suppression cut off the path of the control logic being misled by external disturbances, keeping the parameters of the core controller module stable.

[0035] During the mission, the heavily loaded transport robot was instructed to climb a steep slope fully loaded. This caused the core controller module to calculate a high base control effort signal to eliminate speed errors. This signal exceeded the maximum output capacity of its physical actuator, causing the actuator to enter a saturation state. Under this condition, the inertial mirror adaptive law module observed a large control effort command value and a small error correction effect (because the actuator had reached its limit). Its internal calculation logic tended to adjust the adaptive gain coefficient to a high value. At this time, the adaptive law arbitration module was activated. This module detected the saturation state by comparing the command value of the base control effort signal with the actual output value of the physical actuator. The arbitration module immediately executed the suppression function, temporarily suspending the inertial mirror self-control. The adaptive law module freezes its adaptive gain coefficient at the last effective value before saturation. This arbitration mechanism helps prevent control parameters from being contaminated by erroneous information under physical limits. When the transfer robot climbs to the top of the slope and the saturation state is released, the core controller module can seamlessly resume normal control with an uncontaminated gain coefficient, avoiding instability caused by gain accumulation. The pre-calibration of the transient detection module, combined with the logic suppression of the error signal identification module and the adaptive law arbitration module during operation, ensures that the control parameters of the core controller module remain in a state adapted to the controlled object and operating conditions under the three conditions of sudden model change, external dynamic disturbance, and physical actuator saturation.

[0036] Example 2: To objectively verify the control performance of the control system of the present invention in the field when dealing with time-varying and external disturbances of the controlled object model, i.e., its responsiveness and robustness, a hardware-in-the-loop test platform for the control system was built. This test platform uses an industrial controller with a 1ms operating cycle as the hardware carrier for the core controller module and other logic modules. The controlled object, i.e., the transfer robot, is simulated through a variable inertia flywheel system driven by a servo motor. The physical inertia of this system can be controlled by switching between two sets of flywheels via an electromagnetic clutch, achieving low inertia... (Simulated no-load) and high inertia Switching between (simulated full load) and other modes, where Set as Five times; the physical actuator is a servo motor driver with a defined maximum output current limit to simulate saturation; speed measurements are provided by a 10,000-line incremental encoder on the motor shaft; additionally, a separate torque motor is coaxially connected to the servo motor to apply a 2Hz, amplitude-controllable sinusoidal torque to simulate a preset dynamic disturbance signature; to evaluate the synergistic effect of the various control logic modules in this invention, three test groups were set up: control group A, which is a standard PID controller with fixed parameters, which... Parameters based on high inertia The system is tuned under operating conditions, a conservative tuning strategy used in engineering to avoid full-load oscillations; Control group B is a partially functional system, containing only the core controller module and the inertial image adaptive law module, used to verify the basic adaptive function, but lacking the identification and arbitration of disturbances and saturation; and the present invention sample group adopts the complete control system as described in the specific implementation, including the core controller module, the inertial image adaptive law module, the error signal identification module, and the adaptive law arbitration module; all test groups perform a standard test sequence, which sequentially includes: under low inertia Velocity step response test (0 to 100 rad / s) at high inertia Velocity step response test (0 to 100 rad / s) at high inertia Applying a 2Hz dynamic disturbance during steady-state operation at 100 rad / s, and under high inertia... The system executes a high acceleration command sufficient to trigger the physical actuator to saturate. During the test, the key performance indicators of the control system, including the rise time (10%-90%) of the speed step response and the speed overshoot, as well as the steady-state speed oscillation amplitude under disturbance, were recorded in real time, as shown in Table 1.

[0037] Table 1: Comparison Test Data of Control System Performance

[0038]

[0039] Experimental data show that in condition (1) low inertia step, the rise time of control group A (fixed PID) is 1.35s; the inertial mirror adaptive law modules of control group B and the present invention both identified low inertia and adjusted the control gain accordingly, with rise times of 0.41s and 0.42s respectively; in condition (2) high inertia step, all three groups showed stability, with the adaptive gain coefficients of control group B and the present invention both converging to values ​​matching the high inertia condition; the results of conditions (3) and (4) show the performance differences under different control logics; when a 2Hz disturbance is applied in condition (3), the inertial mirror adaptive law module of control group B interprets the velocity error introduced by the dynamic disturbance as model inaccuracy and continuously increases its adaptive gain coefficient, resulting in a control gain ( The ) is amplified, which in turn amplifies the 2Hz disturbance, forming positive feedback, and the final speed oscillation amplitude reaches 6.2rad / s; while in the sample group of the present invention, the error signal identification module identifies the error as conforming to the preset dynamic disturbance signature through frequency domain analysis, and suppresses the update of the inertial mirror adaptive law module, so that the control gain is kept at the value determined by the working condition (2), and the steady-state oscillation is suppressed to 0.4rad / s.

[0040] In the saturation start-up of condition (4), the inertial mirror adaptive law module of control group B observed a large basic control effort signal (command value) and a small speed error change (actual execution limited) during the physical actuator saturation period, which led to it calculating a high adaptive gain coefficient. When the system left the saturation state, this gain caused a speed overshoot of 41.5%. In the sample group of the present invention, after the adaptive law arbitration module detected the saturation state, it suppressed the operation of the inertial mirror adaptive law module, freezing its gain at the value before entering saturation. When the system left the saturation state, the controller used this gain to restore control, and the overshoot was 3.1%. The experimental data show that the error signal identification module and the adaptive law arbitration module in the sample group of the present invention, as the cooperative unit of control logic, enable the inertial mirror adaptive law module to avoid the risk of adverse adjustment due to changes in information flow under dynamic disturbance and saturation state. This allows the entire control system to maintain the adaptability to changes in the inertial model of the controlled object while also maintaining the robustness of control under complex conditions.

[0041] Example 3: This example combines Figures 1 to 3 A description of a control system for adjusting the operating speed of a transfer robot based on load adaptive adjustment is provided, such as... Figure 1 As shown, in this system, a core controller module is based on the speed error between the speed setpoint and the speed measurement value, and according to a set of adjustable control parameters such as... , The system generates a basic control effort signal to a physical actuator, such as a motor driver. The physical actuator drives the controlled object of the transfer robot to run. The robot's load or inertia changes in real time, and its running speed is fed back as a speed measurement value. At the same time, an inertial mirror adaptive law module receives the basic control effort signal and speed error in real time and calculates the adaptive gain coefficient accordingly to adjust the control parameters of the core controller module. The system also includes a dual suppression mechanism. The first is an adaptive law arbitration module, which monitors whether the physical actuator has entered a saturation state and sends a suppression signal when saturation occurs. The second is an error signal identification module, which analyzes the frequency domain characteristics of the speed error to identify the dynamic disturbance signature and sends a suppression signal when a disturbance occurs. Both suppression signals are used to prevent the inertial mirror adaptive law module from updating its gain coefficient.

[0042] like Figure 2 As shown, the horizontal axis represents time (s), and the vertical axis represents velocity (rad / s). The target velocity is set to 100 rad / s. The figure indicates no-load operation. - The curve (solid line) of the system of the present invention shows the system of the present invention under no-load conditions. The response under operating conditions has the fastest rate of increase, indicated by full load. - The curve (dashed line) of the system of the present invention shows the system of the present invention under full load. The response speed under operating conditions is slower than that under no-load conditions, but it can still reach the set value smoothly and quickly. For comparison, the full-load condition is indicated. - The curve (dotted line) for a fixed PID controller shows the performance of a PID controller with fixed parameters under full load. The response under operating conditions not only has a slow rise time, but also exhibits significant overshoot and oscillation when reaching the setpoint. For example... Figure 3 As shown, the process begins with the controlled object transmitting a speed measurement value containing disturbances to the core controller module. After calculation, the core controller module sends the speed error signal stream containing disturbance characteristics to the error signal identification module. After detecting the dynamic disturbance signature through frequency domain analysis such as bandpass filtering, the error signal identification module immediately sends a suppression signal to the inertial mirror adaptive law module. Upon receiving the suppression signal, the inertial mirror adaptive law module performs the action of pausing the gain coefficient update and instructs the core controller module to maintain the current control parameters. The core controller module then outputs a stable control signal to the physical actuator, thereby cutting off the interference path of external disturbances to the adaptive logic.

[0043] Example 4: To determine the reproducible settings of key parameters in the control system of this invention, specifically involving the preset nonlinear function relationship lookup table in the inertial mirror adaptive law module and the preset disturbance characteristic frequency band and energy threshold in the error signal identification module, a systematic offline calibration procedure is executed. This procedure is based on the analysis of the response data of the control system under controlled excitation. The calibration environment is a test bench equipped with a data acquisition system, capable of applying known loads and controllable external disturbance torques. The transfer robot is fixed on the test bench, and its core controller module and related logic modules run in the controller hardware. The controller's operating cycle is set to 1ms, and the speed measurement value is provided by the motor encoder with a resolution of 10,000 lines / revolution. The steps for calibrating the preset nonlinear function relationship lookup table are as follows: First, a series of known loads are applied to the transfer robot in sequence, covering its working range, including no load (marked as...). ), 25% of rated load ( ), 50% of rated load ( ), 75% of rated load ( ) and 100% rated load ( ); in each load Under these conditions, a set of standardized speed control tasks are executed: accelerating from a standstill to 50 rad / s, maintaining a constant speed for 2 seconds, and then decelerating to a standstill, repeating this process 5 times; during each task execution, the instantaneous value of the basic control effort signal is recorded at a period of 1 ms. Instantaneous value of speed error For each acceleration and deceleration phase, the basic control effort signal is calculated within the time window of that phase according to the procedures defined in the specific implementation method. Integral value within And speed error within the same time window rate of change within Then, the ratio of this task can be calculated. For the same load The average of the ratios calculated from the next 5 repeated tasks is used to obtain the workload. Corresponding representative ratio Meanwhile, during this offline calibration phase, for each load... By using a domain-specific system identification and controller tuning method (using the Ziegler–Nichols method with step response), a set of parameters that enable a specific load are independently determined. The PID controller (core controller module) achieves the control performance target of a rise time of less than 0.6s and an overshoot of less than 5%. and Value; then, select no load. The proportional gain below As a baseline, calculate each load The corresponding adaptive gain coefficient Finally, the loads will be... Corresponding representative ratio Its corresponding adaptive gain coefficient pair These discrete data points form the lookup table. For ratios not included in the table, their corresponding values ​​can be determined using linear interpolation. The values ​​thus form a pre-defined lookup table of nonlinear function relationships, which is then stored in the controller's non-volatile memory.

[0044] The parameter specifications for the calibration error signal discrimination module are as follows: Select 50% of the rated load. The conditions were set to ensure the transfer robot operated stably at 50 rad / s. Using the perturbation torque motor on the test bench, sinusoidal perturbation torques of different frequencies were sequentially applied to the system, ranging from 0.5 Hz to 5.0 Hz, in 0.5 Hz increments, with each frequency lasting 10 seconds. During the application of each frequency perturbation, the speed error signal was recorded. The time series of the error signal was analyzed. For each recorded error signal segment, its power spectral density was calculated using Fast Fourier Transform. The power spectrum was analyzed to identify the frequency range in which the error signal energy was higher than the baseline level when no disturbance was applied after the disturbance was applied. This frequency range was determined as the preset disturbance characteristic frequency band, which was identified in this calibration as 1.5Hz to 2.5Hz. Subsequently, the baseline error energy within this selected frequency band (1.5Hz-2.5Hz) was recorded under steady-state operation of the system without any external disturbance. Ultimately, the preset threshold Set as a multiple of the baseline energy The selection of this multiplier is used to balance the sensitivity of disturbance detection with the avoidance of false triggering. These determined frequency band parameters and energy thresholds are also stored in the controller. By executing the above systematic calibration procedure, the core nonlinear image relationship of the inertial image adaptive law module and the judgment basis of the error signal identification module are determined, so that when the control system of the present invention is deployed, its adaptive and disturbance rejection functions have a reproducible basis.

[0045] Example 5: To determine the deployment parameters of the control system of the present invention on a specific transport robot, and to handle parameter drift that may occur due to individual differences in hardware or long-term operation, a standardized on-site pre-deployment calibration procedure is executed. This procedure is performed before the transport robot is first put into operation or after replacing key components. The initial state is that the transport robot is in an unloaded and stationary state, located on a flat and level test ground. Its control system has been loaded with the offline calibration parameters as in Example 3, including a preset nonlinear function relationship lookup table. This calibration procedure is used to verify and fine-tune the response characteristics of the inertial image adaptive law module and determine the dynamic adjustment relationship of the setpoint smoothing module. In the first step of the calibration procedure, the system automatically executes the inertial detection sequence. First, the controller injects a preset detection pulse with parameters consistent with those of the offline calibration in Example 3 into the physical actuator and records the resulting speed error change. The inertial image adaptive law module is called, and according to the ratio calculation logic of the specific implementation method, it calculates the representative ratio under the current unloaded state. The controller will The ratio of idle data recorded during offline calibration The two are compared; if the deviation is within the tolerance range set according to the system accuracy requirements, the lookup table is valid at that operating point. Next, a known standard test load, such as 50% of the rated load, is placed on the robot. The process of repeated probe pulse injection and ratio calculation is used to obtain the ratio under the field test load. and compared with the offline calibration results Compare; if If the deviation is within the tolerance range, the offline calibrated lookup table is confirmed to be suitable for the current hardware; if any ratio exceeds the tolerance, the system can calculate a correction factor, which is used to adjust the adaptive gain coefficient output from the lookup table at runtime. Fine-tuning can be performed, for example, by calculating a scaling factor based on the deviation ratio and applying it to... value.

[0046] The second step in the calibration procedure is to determine the smoothness level in the setpoint smoothing module, i.e., the time constant. , and adaptive gain coefficient or ratio The functional relationship between them; this process utilizes the inertial identification results that have been verified or fine-tuned in the first step. First, in the no-load state (corresponding to or The controller receives a step command for the original speed setpoint from 0 to a preset test speed, such as 80% of the maximum speed. The time constant of the setpoint smoothing module is used for this step command. From the minimum value set based on the system's response capability Start; the controller monitors the maximum acceleration during this acceleration process. ;like Below the no-load acceleration limit predetermined according to robot operation safety regulations Then record the pairing. or Subsequently, a standard test load was placed on the robot. (correspond or Repeated velocity step test, this time from Start by gradually increasing the size until a minimum value is found. The value that makes the maximum acceleration under this load... Not exceeding the predetermined load acceleration safety limit Record pairings or Using at least these two calibration data points, a model is established through methods such as linear fitting, piecewise linear interpolation, or table lookup interpolation. Follow or Changing functional relationship or The functional relationship is then embedded in the setpoint smoothing module, which is used to dynamically adjust the smoothness of the speed command based on the inertial information output in real time by the inertial mirror adaptive law module. By executing this standardized pre-calibration procedure, the control system of the present invention establishes a reproducible mapping relationship between its adaptive logic and the specific hardware platform, while the parameters of its feedforward smoothing control are also set according to the operational safety constraints.

[0047] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0048] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A control system for adaptively adjusting the operating speed of a transfer robot based on load, characterized in that, The system includes: A core controller module is configured to generate basic control effort signals to the physical actuators of the transfer robot based on the speed error between the speed setpoint and the speed measurement value, and according to a set of adjustable control parameters. An inertial mirror adaptive law module is configured to calculate an adaptive gain coefficient in real time based on the basic control effort signal and velocity error, and use the adaptive gain coefficient to adjust the adjustable control parameters of the core controller module. An error signal identification module is configured to: analyze a frequency domain feature of the velocity error in real time; and when the frequency domain feature matches a preset dynamic disturbance signature, suppress the inertial image adaptive law module to prevent it from updating the adaptive gain coefficient. The dynamic disturbance signature is defined as the energy of the velocity error being concentrated in a preset disturbance feature frequency band. The error signal identification module uses a digital bandpass filter to monitor whether the velocity error matches the dynamic disturbance signature. An adaptive law arbitration module is configured to: monitor whether the physical actuator enters a saturation state; and suppress the operation of the inertial mirror adaptive law module during the period when the physical actuator is in a saturation state, so as to prevent it from updating the adaptive gain coefficient; The inertial mirror adaptive law module is configured to: monitor the integral value of the basic control effort signal within a first preset time window; monitor the rate of change of the velocity error within a second preset time window; calculate the adaptive gain coefficient based on the ratio between the integral value and the rate of change; and freeze and hold the last generated adaptive gain coefficient until the velocity error exceeds the steady-state threshold again when the velocity error continues to be below a preset steady-state threshold.

2. The control system for adaptively adjusting the operating speed of a transfer robot based on load, as described in claim 1, is characterized in that... The system further includes: a transient detection module configured to: respond to a step change in the speed setpoint from the steady-state value; and inject a preset detection pulse into the physical actuator as a basic control effort signal before the core controller module responds to the step change; and trigger an inertial mirror adaptive law module to calculate and set an initial adaptive gain coefficient using the control effort of the preset detection pulse and the resulting change in speed error; the core controller module is further configured to use the initial adaptive gain coefficient as an adjustable control parameter.

3. The control system for adaptively adjusting the operating speed of a transfer robot based on load, as described in claim 2, is characterized in that... The transient detection module is further configured to: immediately instruct the transfer robot to run at a preset constant low speed for a preset period of time after executing a preset detection pulse; and monitor the average base control effort required to maintain the preset constant low speed during the preset period of time to determine a damping mirror parameter. The core controller module is configured to adjust its proportional gain and integral gain using an initial adaptive gain coefficient, and is further configured to independently adjust its differential gain using a damping mirror parameter.

4. The control system for adaptively adjusting the operating speed of a transfer robot based on load, as described in claim 1, is characterized in that... The system further includes: a setpoint smoothing module configured to receive a speed setpoint as an original speed setpoint and generate a smoothed speed setpoint; the core controller module is further configured to use the smoothed speed setpoint instead of the original speed setpoint to calculate the speed error; wherein the smoothing degree of the setpoint smoothing module is adjusted by the adaptive gain coefficient generated by the inertial mirror adaptive law module, so that the smoothing degree is enhanced when the adaptive gain coefficient, which represents the inertia of the transfer robot, increases.

5. A control system for adaptively adjusting the operating speed of a transfer robot based on load, as described in claim 1, is characterized in that... The system further includes a transient spike suppression module configured to: receive the raw speed measurement signal stream; apply medium-range filtering logic to the raw speed measurement signal stream to generate purified speed measurement values; and provide the purified speed measurement values ​​as speed measurement values ​​to the core controller module.

6. The control system for adaptively adjusting the operating speed of a transfer robot based on load, as described in claim 1, is characterized in that... Saturation refers to a state in which the actual output value of a physical actuator reaches its output limit, resulting in a discrepancy between the basic control effort signal and the actual output value of the physical actuator. The adaptive law arbitration module monitors whether the physical actuator has entered saturation by comparing the basic control effort signal with the actual output value of the physical actuator.

7. A control system for adaptively adjusting the operating speed of a transfer robot based on load, as described in claim 1, is characterized in that... The core controller module is a PID controller; the adjustable control parameters include the proportional gain K of the PID controller. p With integral gain K i The inertial mirror adaptive law module is configured to determine the proportional gain K based on the adaptive gain coefficients through a preset nonlinear function lookup table. p With integral gain K i The value.

8. A control system for adaptively adjusting the operating speed of a transfer robot based on load, as described in claim 2, is characterized in that... The transient detection module is configured to respond to a step change in the speed setpoint from zero after the transfer robot is stationary and its load changes, thereby completing the calculation and setting of the initial adaptive gain coefficient before the transfer robot begins to perform acceleration.

Images