Dual-module control system and method based on multi-sensor fusion and adaptive cooperation

Abstract

The invention relates to the technical field of industrial control, and discloses a dual-module control system and method based on multi-sensor fusion and adaptive collaboration, and the system comprises a sensing processing unit, a first control module, a second control module and a collaborative arbitration module, the logic phase of the first control module is mapped to the second control module in real time through a state shadow register, and the second control module is triggered to compensate the calculation logic before a preset action threshold value according to the total response deviation determined by the action lag time of the execution mechanism and the medium transmission lag time; according to the method, the time domain normalization alignment of the compensation gain and the basic control instruction at the physical action starting point of the controlled object is realized, the logic out-of-step phenomenon generated by physical inertia in industrial control is eliminated, and the action certainty in the production process is guaranteed.

Description

Technical Field

[0001] This invention relates to a dual-module control system and method based on multi-sensor fusion and adaptive collaboration, belonging to the field of industrial control technology. Background Technology

[0002] The current system precisely controls controlled objects such as dispensing heads and reaction vessels based on real-time collection of cleanliness and media flow. Existing dual-module control systems often adopt master-slave redundancy or equal-weight parallel mode, and manage the action sequence of the actuators through a logical arbitration mechanism. Such systems assume that the environmental parameters are globally statically distributed, and the characteristic data acquired by each control module are consistent in timing.

[0003] In actual operating conditions, the pulse-like disturbances generated by the flow of the process medium and the inherent lag in the physical transmission process cause nonlinear phase shifts in the characteristic data acquired by different control modules. Existing logic arbitration mechanisms cannot identify such physical lags, resulting in pseudo-feature differences. This leads to instruction drift or closed-loop oscillations in the actuator at critical process nodes, causing technical problems such as media waste or exceeding cleanliness standards. Increasing sensor accuracy or improving computing power cannot eliminate the disconnect between instruction logic and physical action timing; instead, it increases the complexity of the control loop and the system deployment cost. In addition to hardware bottlenecks in the actuators, the control logic architecture and its calibration mechanism also restrict the determinism of actions. For example, Chinese invention with publication number CN116499933A... The patent discloses a method for testing the cleanliness of a cleaning brush and a method for setting a cleaning process window. It selects cleaning brushes based on particle conditions and sets a cleaning process window. However, the technical approach focuses on static screening of macroscopic parameters and does not address the microscopic time delay matching between the logical jump of control commands and the starting point of physical actions. In high-speed and high-frequency response scenarios, if there is a lack of accurate extraction and dynamic compensation of media transmission delay and mechanical response delay of actuators, even if the process parameters are within the preset window, there will still be a phase deviation between the logical and physical execution processes in the time domain. The decoupling between the logical and physical levels causes severe pressure pulses to be generated at the edge of the execution window. Because the arbitration unit lacks the ability to compensate for physical lag, it triggers logical competition between multiple modules, reducing the continuity of production in non-ideal environments.

[0004] Therefore, how to reconstruct the synchronization relationship between industrial control instructions and physical execution actions, and use logical timing offset to compensate for the physical lag of the actuator, so that the compensation gain is accurately aligned with the basic instructions at the physical level, has become the technical problem to be solved by this invention. Summary of the Invention

[0005] To address the problems mentioned in the background art, the technical solution of the present invention is as follows: A dual-module control system based on multi-sensor fusion and adaptive coordination, comprising: The sensing and processing unit is used to acquire the environmental cleanliness parameters of the controlled object and the physical characteristic parameters of the process medium; The first control module has a built-in cleanliness-execution mapping table, which is used to convert environmental cleanliness parameters into the process execution window of the controlled object by retrieving the cleanliness-execution mapping table, and generate basic control instructions. The second control module is used to calculate the compensation gain of the basic control commands based on physical characteristic parameters; The collaborative arbitration module is connected to both the first and second control modules and has a built-in set of response delay constants. This set includes the first lag time of the controlled object's actuator from receiving the instruction to the start of the physical action. And the second lag time of the process medium being transported in the pipeline to the end of the actuator. The collaborative arbitration module maps the current execution phase of the first control module to the second control module in real time through a status shadow register; the collaborative arbitration module is used to calculate the first lag time. With the second lag time The sum is used to determine the total response deviation. The total response deviation before the current execution phase reaches a preset execution threshold is detected by the status shadow register. At the corresponding timing point, the status shadow register triggers the compensation calculation logic of the second control module; the cooperative arbitration module loads the generated compensation gain into the basic control command at the instant of the instruction jump of the first control module, and outputs the cooperative control vector, so that the compensation gain and the basic control command achieve timing synchronization at the starting point of the actuator's action.

[0006] Preferably, the first control module is used to shorten or extend the duration of the process execution window by retrieving the cleanliness-execution mapping table when the environmental cleanliness parameter fluctuates; the collaborative arbitration module is used to switch the collaborative control vector to the virtual loop execution instruction during the process execution window contraction, driving the process medium from the execution end of the controlled object into the cycle maintenance loop.

[0007] Preferably, the physical characteristic parameters include the real-time flow rate of the process medium; the second control module is used to calculate the compensation gain based on the deviation between the real-time flow rate and the preset standard flow rate, and to correct the amplitude of the basic control command.

[0008] Preferably, the collaborative arbitration module is also connected to the online monitoring module; the collaborative arbitration module is used to reduce the output weight of the first control module and increase the execution priority of the media circulation logic in the second control module when the environmental cleanliness parameter fed back by the online monitoring module is worse than the preset safety threshold.

[0009] Preferably, the sensing processing unit controls the first control module and the second control module to perform asynchronous sampling at coprime frequencies; the collaborative arbitration module is used to extract the logical residual between the mapped predicted value and the actual feedback value in the state shadow register. Among them, logical residuals The calculation formula is: ,in, The execution phase prediction value stored in the state shadow register. Real-time physical feedback values ​​collected by the sensing and processing unit; collaborative arbitration module, used for logical residuals When the unidirectional trend distribution is satisfied within the preset time window, the output result of the second control module is called to update the state baseline of the first control module.

[0010] Preferably, the collaborative arbitration module is used to simultaneously issue virtual loop commands to the second control module when the process execution window shrinks, and drive the valve of the controlled reflux branch to open, so that the process medium is switched from the execution end of the controlled object to the circulation maintenance loop; the second control module tracks the current execution phase through the status shadow register and switches the output carrier from the actuator to the reflux pump.

[0011] Preferably, the collaborative arbitration module is used to superimpose a micro-perturbation signal onto the status shadow register during the operation of the controlled object, the amplitude of which is lower than the dead zone threshold of the actuator; the second control module is used to collect the feedback characteristics generated by the controlled object in response to the micro-perturbation signal and calculate the physical characteristic parameters of the controlled object's execution end based on the feedback characteristics; the collaborative arbitration module is used to correct the control gain of the first control module based on the physical characteristic parameters of the execution end.

[0012] Preferably, the collaborative arbitration module is used to extract the logical residual sequence of the mapped predicted value and the measured value through the state shadow register, and calculate the kurtosis coefficient of the logical residual sequence within the sliding window; when the kurtosis coefficient exceeds the preset judgment threshold, the collaborative arbitration module determines that there is external electromagnetic interference and keeps the compensation gain at the current state value.

[0013] Preferably, the collaborative arbitration module is used to monitor the update cycle of the status shadow register; the status shadow register is used to perform linear deduction according to a preset phase evolution model to generate a virtual execution phase when the update cycle deviates; the collaborative arbitration module is used to calculate the residual between the virtual execution phase and the measured execution phase of the first control module after the update cycle recovers, and correct the stored value of the status shadow register according to the residual; the status shadow register is used to store the impedance characteristic quantity characterizing the pressure drop data of the process medium at the distribution node; the second control module is used to extract the impedance characteristic quantity and calculate the gain correction value of each execution branch in combination with the current execution phase, so as to balance the flow difference between the execution branches.

[0014] A dual-module control method based on multi-sensor fusion and adaptive coordination includes the following steps: The sensing and processing steps acquire the environmental cleanliness parameters of the controlled object and the physical characteristic parameters of the process medium. The first control step involves retrieving the built-in cleanliness-execution mapping table, converting the environmental cleanliness parameters into the process execution window of the controlled object, and generating basic control instructions. The second control step is to calculate the compensation gain of the basic control command based on the physical characteristic parameters. The collaborative arbitration step maps the current execution phase of the first control step to the second control step in real time through a status shadow register, and calculates the first lag time of the controlled object's actuator from receiving the instruction to the start of the physical action. The second lag time between the process medium and the transmission time of the process medium in the pipeline to the end of the actuator. The sum of these values ​​is used to determine the total response deviation. The total response deviation before the current execution phase reaches a preset execution threshold is detected by the status shadow register. At the corresponding timing point, the state shadow register triggers the compensation calculation logic of the second control step; at the instant of instruction jump in the first control step, the generated compensation gain is loaded into the basic control instruction, and the cooperative control vector is output, so that the compensation gain and the basic control instruction achieve timing synchronization at the starting point of the actuator's action.

[0015] Compared with the prior art, the beneficial effects of the present invention are: 1. In the dual-module system of multi-sensor fusion and adaptive collaboration, the state shadow register locks the logic phase of the first control module in real time as the reference scale for the second control module to perform compensation actions. By introducing a set of response delay constants corresponding to the physical characteristics of the actuator, the timing of the compensation gain calculation is offset by a controlled advance relative to the basic control command. This compensates for the lag caused by the resistance of the physical medium flow channel and the mechanical inertia of the actuator, thereby aligning the physical actions of the controlled object at the end of the execution and eliminating pressure pulses and flow overshoot caused by spatiotemporal asymmetry.

[0016] 2. The first control module dynamically reduces or expands the duration of the process execution window based on the real-time collected environmental cleanliness parameters. During the window contraction period, the collaborative arbitration module switches the output vector to the virtual loop execution command, driving the process medium from the execution end into the circulation maintenance loop. Under the stable state of the physical medium flow characteristics, the control logic flexibly contracts to avoid precipitation, stratification or deterioration of the fluid medium in the pipeline due to stagnation, thereby reducing the system's response time from the disturbance state to the production state.

[0017] 3. The sensing and processing unit controls the dual modules to sample asynchronously at coprime frequencies. The collaborative arbitration module compares the predicted value of the state shadow register with the measured feedback value in real time to generate a residual sequence. Based on the kurtosis coefficient and standard deviation of the residual sequence within the sliding window, external electromagnetic pulse interference is identified. When interference is detected, the current state of the controlled object is maintained through inertial extrapolation logic. The temporal staggered distribution of the dual sampling points is used to identify the sensor's trend zero-point drift, and the intrinsic performance of the system is repaired during long-term operation to suppress the silent degradation of control accuracy. The collaborative arbitration module injects a micro-jitter signal below the dead zone threshold of the actuator in the stable section of the control loop. It uses the feedback characteristics of the actuator to invert the equivalent frictional damping and response bandwidth margin of the actuator. Based on the physical characteristics, the control gain is dynamically corrected to compensate for the physical wear of the actuator. In conjunction with the phase evolution model inertial prediction logic, the communication bus transmission jitter is suppressed, and the continuity of the logic reference of the system is maintained under the conditions of actuator aging or bus congestion. Attached Figure Description

[0018] Figure 1 This is a schematic diagram of the dual-module control logic interaction of the multi-sensor fusion and adaptive collaboration of the present invention. Figure 2 This is a diagram showing the correspondence between the frequency of the micro-perturbation signal and the physical characteristics of the feedback amplitude and phase delay of the present invention. Figure 3 This is a causal diagram illustrating the technical elements that eliminate logic synchronization issues in the dual-module control system of this invention. Detailed Implementation

[0019] The following embodiments are intended for illustrative purposes only and are not intended to limit the scope of protection of the present invention.

[0020] This invention discloses a dual-module control system and method based on multi-sensor fusion and adaptive collaboration, including a sensing and processing unit, a first control module, a second control module, and a collaborative arbitration module. The sensing and processing unit acquires environmental cleanliness parameters of the controlled object and physical characteristic parameters of the process medium. The first control module has a built-in cleanliness-execution mapping table, which is used to convert environmental cleanliness parameters into the process execution window of the controlled object and generate basic control commands. The second control module calculates the compensation gain of the basic control commands based on the physical characteristic parameters. The collaborative arbitration module is connected to both the first and second control modules and has a built-in response delay constant set. The collaborative arbitration module maps the current execution phase of the first control module to the second control module in real time through a state shadow register. The collaborative arbitration module determines the timing deviation of the controlled object at the physical execution level through the response delay constant set, which includes a first lag time. With the second lag time ;in, The time from receiving an instruction to the initiation of a physical action by the actuator of the controlled object. The time it takes for the process medium to travel from the pipeline to the end of the process; the collaborative arbitration module calculates the total response deviation using the following formula. : ,in, The total response deviation; the status shadow register monitors the current execution phase, and the total response deviation before the phase evolves to a preset execution threshold. At the corresponding timing point, the compensation calculation logic of the second control module is triggered; at the instant of the instruction jump of the first control module, the collaborative arbitration module loads the generated compensation gain into the basic control instruction and outputs the collaborative control vector, so that the compensation gain and the basic control instruction achieve timing synchronization at the starting point of the actuator's action.

[0021] Taking the application scenario of dispensing process as an example, the first delay time is preset. The second lag time is 50ms. The total response deviation was calculated to be 120ms. It is 170ms; if the transition phase point of the basic control command is... The status shadow register arrives at the current execution phase. The second control module's gain calculation is triggered at a certain moment; the second control module calculates the compensation gain based on the deviation between the real-time flow and the standard flow, which is used to correct the amplitude of the basic control command; when the first control module... When the output command changes, the collaborative arbitration module combines the calculated compensation gain with the command, ensuring that the logical compensation amount aligns with the physical action start point when the physical medium produces an action at the execution end of the controlled object; the sensing processing unit controls the first and second control modules to perform asynchronous sampling at coprime frequencies; the collaborative arbitration module extracts the logical residual between the mapped predicted value and the actual feedback value in the state shadow register in real time. Logical residuals The calculation formula is: , The execution phase prediction value stored in the state shadow register. Real-time physical feedback values ​​collected by the sensing and processing unit; when the logic residual When a unidirectional trend distribution is satisfied within a preset time window, the collaborative arbitration module calls the output result of the second control module to update the state baseline of the first control module; if the kurtosis coefficient of the logic residual sequence exceeds the preset judgment threshold within the sliding window, the collaborative arbitration module determines that there is external electromagnetic interference and maintains the compensation gain at the current state value; when the environmental cleanliness parameter fluctuates, the first control module reduces or expands the duration of the process execution window by retrieving the cleanliness-execution mapping table; when the environmental cleanliness parameter is worse than the preset safety threshold, the collaborative arbitration module reduces the output weight of the first control module and increases the execution priority of the media circulation logic in the second control module; during the process execution window contraction, the collaborative arbitration module switches the collaborative control vector to the virtual loop execution instruction, driving the process medium from the execution end into the circulation maintenance loop; at the same time, it drives the controlled return branch valve to open, so that the process medium switches from the execution end of the controlled object to the circulation maintenance loop, and the second control module switches the output carrier from the execution mechanism to the return pump.

[0022] During the operation of the controlled object, the collaborative arbitration module superimposes a micro-perturbation signal onto the status shadow register. The amplitude of this signal is lower than the dead-zone threshold of the actuator. The second control module collects the feedback characteristics generated by the controlled object in response to this micro-perturbation signal and inverts the physical characteristic parameters of the controlled object's actuator based on these characteristics. The injected signal is defined as follows: ,set up And amplitude Below the dead zone threshold The second control module collects and outputs responses. Perform autocorrelation analysis to extract the fundamental amplitude. With phase bias Establish an equivalent damping model for the execution end. Damping ratio is identified in real time using the recursive least squares method. With natural frequency ,like The drift from the reference value exceeds Call the correction operator The basic command amplitude is adjusted to eliminate the interference of mechanical wear on the consistency of injection action. The collaborative arbitration module corrects the control gain of the first control module according to the physical characteristic parameters of the execution end. The collaborative arbitration module monitors the update cycle of the state shadow register. When the update cycle deviates, the state shadow register performs linear deduction according to the preset phase evolution model to generate a virtual execution phase. After the update cycle recovers, the collaborative arbitration module calculates the residual between the virtual execution phase and the measured execution phase of the first control module, and corrects the stored value of the state shadow register according to the residual. The state shadow register stores the impedance characteristic quantity representing the pressure drop data of the process medium at the distribution node. The second control module extracts the impedance characteristic quantity and calculates the gain correction value of each execution branch in combination with the current execution phase to balance the flow difference between execution branches. The dual-module control method based on multi-sensor fusion and adaptive collaboration includes the following steps: the perception processing step obtains the environmental cleanliness parameters of the controlled object and the physical characteristic parameters of the process medium; the first control step retrieves the cleanliness-execution mapping table, converts the environmental cleanliness parameters into process execution windows and generates basic control commands; the second control step calculates the compensation gain based on the physical characteristic parameters; the collaborative arbitration step realizes phase mapping through the state shadow register, according to and Determine the total response deviation And at the corresponding timing point, the compensation calculation is triggered, and the gain is loaded onto the basic control command at the instant of command jump, so that the compensation gain and the basic control command are synchronized in timing at the start of the action.

[0023] Example 1: In a semiconductor precision dispensing application scenario, the controlled object is a micro-jet valve performing high-frequency dispensing. The sensing and processing unit acquires the cleanliness parameters of the environment where the jet valve is located in real time, as well as the physical characteristic parameters of the process medium in the supply pipeline. When the system faces the objective condition of a local aerosol concentration increase due to transient fluctuations in environmental cleanliness, the first control module retrieves the built-in cleanliness-execution mapping table and reduces the duration of the process execution window according to the real-time changes in cleanliness parameters, while simultaneously generating basic control commands. During the physical execution of the micro-jet valve, there is a mechanical response lag from receiving the command to generating the jetting action, and there is a hydrodynamic lag in the transmission of the process medium from the pipeline to the nozzle end. At this time, the collaborative arbitration module calls the built-in response delay constant set to determine the first lag time. The second lag time is 35ms. It is 115ms, and according to The total response deviation was calculated. It takes 150ms; The time from when the actuator receives the basic control command to when the physical action begins. The time it takes for the process medium to travel from the pipeline to the end of the process. The total response deviation is calculated using a status shadow register that maps the current execution phase of the first control module in real time and monitors its evolution. When the status shadow register detects that the current execution phase reaches a timing point 150ms before the transition edge of the basic control command, it triggers the compensation calculation logic of the second control module. The second control module calculates the compensation gain based on the real-time flow deviation of the current process medium, which is used to correct the amplitude of the basic control command. At the instant the basic control command of the first control module transitions, the collaborative arbitration module loads the generated compensation gain into the basic control command. Because the timing of the compensation logic triggering relative to the command transition edge results in a total response deviation... The equivalent controlled advance, the compensation gain, arrives at the end of the micro-injection valve at the physical level at the time point, which coincides with the starting point of the physical action triggered by the basic control command on the time axis.

[0024] During continuous system operation, if the sensing and processing unit detects that the environmental cleanliness parameter is worse than the preset safety threshold, the collaborative arbitration module lowers the output weight of the first control module and increases the execution priority of the media circulation logic in the second control module. To prevent the process medium from stagnating or precipitating in the pipeline during the contraction of the process execution window, the collaborative arbitration module switches the output collaborative control vector to the virtual loop execution command. At this time, the controlled return branch valve opens, driving the process medium into the circulation maintenance loop. The second control module continues to track the logic execution phase of the first control module through the state shadow register and switches the output carrier from the micro-jet valve to the return pump. In this way, the process medium maintains its flow characteristics during the protective logic contraction, resolving the contradiction between protecting the controlled object and maintaining the stability of the medium. Simultaneously, the sensing and processing unit controls the first and second control modules to perform asynchronous sampling at coprime frequencies, with the sampling frequency of the first control module set to 1003Hz and the sampling frequency of the second control module set to 997Hz. The collaborative arbitration module extracts the logical residual between the mapped predicted value and the actual physical feedback value in the state shadow register in real time. Logical residuals The calculation formula is: ,in, For logical residuals, The execution phase prediction value stored in the state shadow register is the real-time physical feedback value acquired by the sensing processing unit; when this logic residual... If the sensor source shows a unidirectional increasing trend and the deviation exceeds 15% within a preset 500ms time window, the system determines that the sensor source has a trend of zero-point drift. The collaborative arbitration module calls the output of the second control module to reshape the state reference of the first control module. This mechanism utilizes the time-domain interleaving characteristics of dual-module sampling to achieve the calibration of the sensing deviation without introducing an external reference.

[0025] When the controlled object enters the steady-state operation phase, the collaborative arbitration module injects a micro-amplitude jitter signal with a preset frequency of 60Hz and an amplitude lower than the dead zone threshold of the injection valve into the state shadow register. The second control module collects the feedback characteristics of the injection valve generated by this micro-amplitude jitter signal and inverts the physical characteristic parameters of the injection valve's actuator end, including the equivalent friction damping and response bandwidth margin. The collaborative arbitration module corrects the gain of the basic control command of the first control module based on the inverted physical characteristic parameters of the actuator end to compensate for the response hysteresis caused by mechanical wear. For pulse interference in complex electromagnetic environments, the collaborative arbitration module extracts the logical residual sequence between the mapped predicted value and the measured value through the state shadow register. The system calculates the kurtosis coefficient of the logical residual sequence within the sliding window. If it exceeds the preset threshold, the system determines that there is transient interference and freezes the compensation gain update of the second control module. The first control module performs linear deduction based on the phase evolution model to generate a virtual execution phase. This deduction logic maintains the continuity of control commands during bus transmission jitter or data frame loss. After the interference disappears, the stored value of the state shadow register is smoothly corrected by a third-order S-curve. By embedding physical response characteristics into the architecture design of logical compensation timing, the system redefines the original simple instruction alignment as physical action alignment that includes execution end characteristics, so that the controlled object achieves the predetermined process determinism at the execution end.

[0026] Example 2: In the dynamic response verification test of the precision fluid control system, the prototype of this invention adopts a dual-module control system and method based on multi-sensor fusion and adaptive coordination, while the control group adopts a dual-module redundant control system without a state shadow register and response delay constant set. The test platform is built on a distributed industrial control architecture, and the physical experimental environment is equipped with a mass flow meter with a range of 0 mL / min to 50 mL / min and a resolution of 0.01 mL / min, and a pressure transmitter with a sampling frequency of 1000 Hz. The core parameters in the test include the sampling period and the first lag time. and the second lag time Sampling period The sampling frequency is set to 1ms to balance the real-time performance of control commands with the computational load on the processor. This is for systems with a bus communication rate of 1Mbps and a controlled injection valve operating frequency of 200Hz. Increasing the sampling frequency reduces the sampling period. A time delay of 1ms satisfies the Nyquist sampling theorem and ensures real-time scheduling margin; first delay time With the second lag time The value is determined according to the physical measurement procedure. The physical displacement process from the emission of the electromagnetic trigger signal to the droplet leaving the nozzle is captured by a high-speed camera with a shooting frequency of 2000fps. The time from receiving the basic control command to the start of the physical action of the actuator is measured. The time is 42.5 ms, which is the time it takes for the process medium to travel from the pipeline to the end of the actuator. It is 108.2ms; according to the formula The total response deviation was calculated. The duration was 150.7ms; to simulate an industrial electromagnetic environment, Gaussian white noise with a signal-to-noise ratio of 22dB and power frequency harmonic interference with a frequency of 50Hz were actively superimposed in the test signal source.

[0027] The experimental process constructs a gradient verification system by changing the environmental cleanliness parameters of the controlled object and the disturbance intensity of the process medium flow rate; the sensing and processing unit controls the first control module and the second control module to perform asynchronous sampling at coprime frequencies of 1003Hz and 997Hz; when the environmental cleanliness parameter fluctuates around the preset safety threshold of 85, the system initiates the collaborative arbitration logic; Table 1 is the system response performance gradient verification data table, used to compare the timing deviation between the sample group and the control group at the starting point of the physical action; the logic residuals in the table From the formula The calculation shows that, among which The execution phase prediction value stored in the state shadow register. Real-time physical feedback values ​​collected by the sensing and processing unit; the time-series deviation quantification measures the distance on the time axis between the actual physical action point of the actuator and the logical jump point of the basic control command.

[0028] Table 1: System Response Performance Gradient Verification Data Table

[0029] According to the experimental data shown in Table 1, within the operating range of cleanliness level higher than 85, the timing deviation of the control group increased from 148.5 ms to 155.6 ms as the flow disturbance amplitude increased. The sample group of this invention uses a state shadow register to map the current execution phase of the first control module to the second control module in real time, and the total response deviation before the execution phase reaches the preset execution threshold is... The corresponding timing point triggers the compensation calculation logic; the timing deviation of the sample group is maintained within 3ms; when the environmental cleanliness parameter is worse than the safety threshold of 85, the first control module executes the process execution window shrinkage procedure; as the cleanliness level drops below 50, the logic residual... The value rose to 1.520 rad, indicating that the sensor source was experiencing a trend of zero-point drift or was affected by strong interference. Due to the physical dead zone limitation of the micro-jet valve, the compensation correction effect tended to stabilize under the over-limit conditions of cleanliness level 40 and flow disturbance of 65.0%, and the timing deviation of the sample group increased to 28.6 ms. The test group calculated the physical characteristic parameters of the jet valve's actuator end based on the feedback characteristics by superimposing a micro-perturbation signal with a frequency of 60 Hz and an amplitude of 0.15 V onto the state shadow register. The equivalent friction damping increase value was 18.5%. The collaborative arbitration module corrected the control gain of the first control module based on the physical characteristic parameters of the actuator end, so that the system maintained the coordination of action under mechanical wear conditions.

[0030] Example 3: This example combines Figures 1 to 3 This section describes the dual-module control system and method based on multi-sensor fusion and adaptive collaboration, such as... Figure 1 As shown, this system consists of a sensing and processing unit, a first control module, a second control module, and a collaborative arbitration module, forming a collaborative control logic architecture. The sensing and processing unit is responsible for collecting environmental cleanliness parameters and transmitting them to the first control module, while simultaneously acquiring physical characteristic parameters and transmitting them to the second control module. The first control module retrieves the mapping table based on the environmental parameters and generates basic control commands, with its current execution phase being mapped and output in real time. The second control module calculates the compensation gain based on the physical characteristics. The collaborative arbitration module, as the core node, has a built-in set of response delay constants containing the first and second lag times. It uses a state shadow register to receive the current execution phase and triggers the compensation calculation logic. Finally, the compensation gain is loaded into the basic control command at the collaborative control vector output to achieve timing synchronization of the action start point.

[0031] like Figure 2 As shown, the horizontal axis represents the frequency of the micro-perturbation in Hz, the left vertical axis represents the feedback amplitude in V, and the right vertical axis represents the phase delay in ms. The bars filled with horizontal stripes represent the feedback amplitude, and the bars filled with vertical stripes represent the phase delay. The bar chart clearly shows the nonlinear distribution of the feedback amplitude and phase delay as the frequency of the micro-perturbation changes. This data relationship is used to invert the equivalent frictional damping and response bandwidth margin of the actuator; Figure 3 As shown in the diagram, this analysis details the causal relationship between the technical elements of each module of the system and the core objective of ensuring action determinism and eliminating logical inconsistencies. The perception and processing unit branch includes environmental cleanliness parameters, media physical characteristics, and asynchronous sampling elements of coprime frequencies. The first control module branch covers the cleanliness execution mapping table, process execution window, and basic control commands. The second control module branch involves compensation gain calculation, micro-disturbance feedback characteristics, and media loop logic. The collaborative arbitration module branch integrates the response delay constant set and the total response deviation. The status shadow register and logic residual analysis R, along with other branches, work together to achieve the technical success of eliminating logic inconsistency.

[0032] Example 4: In a field deployment scenario for a motion control system of a precision machining center, the control bus communication cycle is set to 2ms. When the system encounters a sudden strong magnetic field interference, causing millisecond-level transmission jitter or loss of synchronization messages on the Ethernet control automation technology bus, the collaborative arbitration module monitors the update cycle deviation of the status shadow register in real time. If the synchronization signal is missing for three consecutive communication cycles, the status shadow register initiates autonomous evolution logic, performing linear deduction based on the logic phase change rate locked in the previous stable cycle to generate a virtual execution phase. The second control module extracts the virtual execution phase as a phase anchor point to maintain continuous output of compensation gain. After the bus synchronization signal is restored, the collaborative arbitration module calculates the deviation between the virtual execution phase and the measured logic phase sent by the first control module. If the deviation is within a preset safety threshold, the system initiates a phase soft alignment procedure, using a third-order S-curve algorithm to smoothly correct the stored value of the status shadow register in the next five execution cycles, avoiding instantaneous jumps in control commands and ensuring the continuity of the actuator's motion trajectory. The sensing and processing unit then... Gradient adjustment of cleanliness parameters The first control module sends out a sequence of pulse commands, and the pressure transmitter feeds back the peak injection pressure. Calculate the pressure offset rate Adjusting the width of the process execution window based on a binary search algorithm ,like Trigger window collapse, step value as Iterate until return Tolerance range, record the target window value and write it to For the address space of the index, complete the discretization mapping matrix encapsulation to construct a deterministic mapping between controlled environmental fluctuations and execution time.

[0033] Example 5: In the initial debugging scenario of a precision industrial control system, the cleanliness-execution mapping table data is acquired in a controlled atmospheric environment chamber by setting the environmental cleanliness level in 5-step increments using an aerosol generator. The range of variation covers a quantitative interval from 100 to 40. For each set level, the first control module generates 1000 continuous pulse signals, and the pressure transmitter monitors the pressure fluctuation amplitude at the moment of injection. When the peak pressure fluctuation exceeds 10% of the standard working pressure, the first control module reduces the process execution window in 0.5ms increments based on the pressure feedback signal until the pressure fluctuation returns to the steady-state deviation range of 3%. The first control module stores the process execution window data for each level in the cleanliness-execution mapping table field of its internal memory, establishing a mapping relationship between environmental cleanliness parameters and process execution windows.

[0034] When the system faces a new physical deployment environment and needs to initialize the response delay constant set, the collaborative arbitration module sends a step jump command to the first control module, and the physical signals are simultaneously collected by the sensors installed at the end of the actuator; the collaborative arbitration module calculates the difference between the timestamp of the logical command issuance time and the physical feedback time to obtain the first lag time. The second lag time of the process medium in the pipeline is calculated by monitoring the pressure step rise time caused by the arrival of the process medium by a pressure probe installed at the end of the execution, combined with the start time of the physical action. The collaborative arbitration module is based on the formula. Calculate the total response deviation ,in The total response deviation, The first lag time, This is the second lag time; the calculation result is written into the storage space of the collaborative arbitration module to configure the status shadow register and establish a timing reference between the logical phase and the physical action.

[0035] Example 6: The sensing and processing unit operates at 0.1 within a controlled environment chamber. The environmental cleanliness parameter is adjusted by gradient (g / m³). For each set quantization level, the first control module records the process execution window width value corresponding to the controlled object when the flow deviation is within the 2% tolerance range. The cleanliness-execution mapping table is defined as a discretized data matrix stored in the internal memory of the first control module. The row vectors represent the quantized environmental cleanliness level, and the column vectors store the corresponding process execution window duration values. The data filling procedure executes the instruction loop 500 times for each quantization level and calculates the average time width value corresponding to that level. The calculated fixed-point value is stored in the address space corresponding to the cleanliness-execution mapping table, thus establishing a mapping relationship between the environmental cleanliness parameter and the process execution window width.

[0036] The collaborative arbitration module measures the basic control command flip rate of the first control module within a 100ms sampling window after system startup to establish a slope benchmark for the phase evolution model. When a deviation in the update cycle of the state shadow register is detected, the state shadow register performs deduction according to the preset phase evolution model to generate a virtual execution phase. This procedure calculates the logical residual between the virtual execution phase and the measured execution phase of the first control module. To determine the correction factor for the stored value, and within the first synchronous clock cycle after the update cycle resumes, to adjust the logic residual... The compensation amount is loaded into the storage bits of the status shadow register to establish a timing reference under bus jitter conditions, where the logic residual... The calculation formula is as follows: ,in, For logical residuals, The execution phase prediction value stored in the state shadow register. Real-time physical feedback values ​​collected by the sensing and processing unit.

[0037] 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.

[0038] 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 dual-module control system based on multi-sensor fusion and adaptive coordination, characterized in that, include: The sensing and processing unit is used to acquire the environmental cleanliness parameters of the controlled object and the physical characteristic parameters of the process medium; The first control module has a built-in cleanliness-execution mapping table, which is used to convert environmental cleanliness parameters into the process execution window of the controlled object by retrieving the cleanliness-execution mapping table, and generate basic control instructions. The second control module is used to calculate the compensation gain of the basic control commands based on the physical characteristic parameters. The collaborative arbitration module is connected to both the first and second control modules and has a built-in set of response delay constants. This set includes the first lag time of the controlled object's actuator from receiving the instruction to the start of the physical action. And the second lag time of the process medium being transported in the pipeline to the end of the actuator. ; The collaborative arbitration module maps the current execution phase of the first control module to the second control module in real time through the status shadow register; The collaborative arbitration module is used to calculate the first lag time. With the second lag time The sum is used to determine the total response deviation. The total response deviation before the current execution phase reaches a preset execution threshold, as detected by the status shadow register. At the corresponding timing point, the status shadow register triggers the compensation calculation logic of the second control module; the cooperative arbitration module loads the generated compensation gain into the basic control command at the instant of the instruction jump of the first control module, and outputs the cooperative control vector, so that the compensation gain and the basic control command achieve timing synchronization at the starting point of the actuator's action.

2. The dual-module control system based on multi-sensor fusion and adaptive coordination according to claim 1, characterized in that, The first control module is used to shorten or expand the duration of the process execution window by retrieving the cleanliness-execution mapping table when the environmental cleanliness parameters fluctuate. The collaborative arbitration module is used to switch the collaborative control vector to the virtual loop execution instruction during the process execution window contraction, driving the process medium from the execution end of the controlled object into the cycle maintenance loop.

3. The dual-module control system based on multi-sensor fusion and adaptive coordination according to claim 1, characterized in that, Physical characteristic parameters include the real-time flow rate of the process medium; The second control module is used to calculate the compensation gain based on the deviation between the real-time flow rate and the preset standard flow rate, and is used to correct the amplitude of the basic control command.

4. A dual-module control system based on multi-sensor fusion and adaptive coordination according to claim 1, characterized in that, The collaborative arbitration module is also connected to the online monitoring module. When the environmental cleanliness parameter fed back by the online monitoring module is worse than the preset safety threshold, the collaborative arbitration module reduces the output weight of the first control module and increases the execution priority of the media circulation logic in the second control module.

5. A dual-module control system based on multi-sensor fusion and adaptive coordination according to claim 1, characterized in that, The sensing and processing unit controls the first control module and the second control module to perform asynchronous sampling at coprime frequencies; The collaborative arbitration module is used to extract the logical residual between the mapped predicted value and the actual feedback value in the state shadow register. Among them, logical residuals The calculation formula is: ,in, The execution phase prediction value stored in the state shadow register. Real-time physical feedback values ​​collected by the sensing and processing unit; collaborative arbitration module, used for logical residuals When the unidirectional trend distribution is satisfied within the preset time window, the output result of the second control module is called to update the state baseline of the first control module.

6. A dual-module control system based on multi-sensor fusion and adaptive coordination according to claim 4, characterized in that, The collaborative arbitration module is used to synchronously issue virtual loop commands to the second control module when the process execution window shrinks, and drive the valve of the controlled reflux branch to open, so that the process medium is switched from the execution end of the controlled object to the circulation maintenance loop; the second control module tracks the current execution phase through the status shadow register and switches the output carrier from the actuator to the reflux pump.

7. A dual-module control system based on multi-sensor fusion and adaptive coordination according to claim 1, characterized in that, The collaborative arbitration module is used to superimpose micro-perturbation signals onto the status shadow register during the operation of the controlled object. The amplitude of the micro-perturbation signals is lower than the dead zone threshold of the actuator. The second control module is used to collect the feedback characteristics of the controlled object in response to micro-disturbance signals, and calculate the physical characteristic parameters of the controlled object's execution end based on the feedback characteristics; the collaborative arbitration module is used to correct the control gain of the first control module based on the physical characteristic parameters of the execution end.

8. A dual-module control system based on multi-sensor fusion and adaptive coordination according to claim 1, characterized in that, The collaborative arbitration module is used to extract the logical residual sequence between the mapped predicted value and the measured value through the state shadow register, and calculate the kurtosis coefficient of the logical residual sequence within the sliding window. When the kurtosis coefficient exceeds the preset judgment threshold, the collaborative arbitration module determines that there is external electromagnetic interference and keeps the compensation gain at the current state value.

9. A dual-module control system based on multi-sensor fusion and adaptive coordination according to claim 1, characterized in that, The collaborative arbitration module is used to monitor the update cycle of the status shadow register; the status shadow register is used to generate a virtual execution phase by performing linear deduction according to the preset phase evolution model when the update cycle deviates. The collaborative arbitration module is used to calculate the residual between the virtual execution phase and the measured execution phase of the first control module after the update cycle is restored, and to correct the stored value of the state shadow register according to the residual. The state shadow register is used to store the impedance characteristic quantity that characterizes the pressure drop data of the process medium at the distribution node. The second control module is used to extract impedance characteristics and calculate the gain correction value of each execution branch in combination with the current execution phase, in order to balance the flow differences between execution branches.

10. A dual-module control method based on multi-sensor fusion and adaptive coordination, used to implement the dual-module control system based on multi-sensor fusion and adaptive coordination as described in claim 1, characterized in that, Includes the following steps: The sensing and processing steps acquire the environmental cleanliness parameters of the controlled object and the physical characteristic parameters of the process medium. The first control step involves retrieving the built-in cleanliness-execution mapping table, converting the environmental cleanliness parameters into the process execution window of the controlled object, and generating basic control instructions. The second control step is to calculate the compensation gain of the basic control command based on the physical characteristic parameters. The collaborative arbitration step maps the current execution phase of the first control step to the second control step in real time through a status shadow register, and calculates the first lag time of the controlled object's actuator from receiving the instruction to the start of the physical action. The second lag time between the process medium and the transmission time of the process medium in the pipeline to the end of the actuator. The sum of these values ​​is used to determine the total response deviation. ; The total response deviation before the current execution phase reaches a preset execution threshold is detected by the status shadow register. At the corresponding timing point, the state shadow register triggers the compensation calculation logic of the second control step; at the instant of instruction jump in the first control step, the generated compensation gain is loaded into the basic control instruction, and the cooperative control vector is output, so that the compensation gain and the basic control instruction achieve timing synchronization at the starting point of the actuator's action.

Images