Soil sample preparation adaptive control system and method based on load feedback

Abstract

The invention discloses a soil sample preparation adaptive control system and method based on load feedback, and relates to the technical field of soil sample preparation control. A soil sample preparation adaptive control system based on load feedback comprises a load characteristic module, a working condition identification module, an operation switching module, an instruction generation module, an instruction evaluation module, an instruction updating module and a sample preparation judgment module. According to the method, the multi-time-window load characteristics are input into the dynamic coordinator under the triggering of operation state switching, and the control demand conflict degree of two abnormal working conditions of hard impact and viscous retardation is evaluated according to predefined conflict arbitration logic, so that a control decision is coupled to the abnormal working conditions instead of a single working condition; therefore, the problem of strategy counteracting or risk amplification is avoided.

Description

Technical Field

[0001] This invention relates to the field of soil sample preparation control technology, and in particular to an adaptive control system and method for soil sample preparation based on load feedback. Background Technology

[0002] Soil sample preparation equipment typically uses motor-driven crushing, grinding, or mixing mechanisms to homogenize and represent the collected soil samples, meeting the requirements for sample homogeneity and repeatability in subsequent testing and analysis. In practical applications, soil sources are complex, with significant differences in moisture content and particle size distribution. The load during sample preparation continuously changes with the material state. Therefore, equipment generally relies on monitoring signals such as motor current and speed to reflect crushing resistance and operating conditions, and accordingly controls such as speed, direction, or operating rhythm to achieve stable sample preparation and equipment protection.

[0003] Control measures for abnormal loads often employ fixed threshold alarms, switching preset strategies after identifying a single operating condition, or simple state machine-based switching between several operating states. These methods treat abnormal operating conditions as independent: when instantaneous impacts caused by hard particles occur, they tend to limit the speed and provide protective buffering; when persistent obstruction occurs due to high water content or viscous agglomeration, they tend to increase the driving force or execute a clearing rhythm. However, in real soil samples, hard impacts and viscous obstruction often occur concurrently and influence each other. For example, when cohesive soil encapsulates hard rocks, impact spikes and stalling tendencies alternate. In such cases, single-condition strategies can easily cancel each other out or even induce control oscillations, manifesting as frequent switching, ineffective clearing, and increased energy consumption. In severe cases, this can lead to jamming shutdowns or mechanical overload damage, making it difficult to simultaneously achieve sample preparation efficiency, stability, and equipment safety. Summary of the Invention

[0004] The technical problem this solution aims to solve is: during soil sampling, when two abnormal operating conditions, hard impact and viscous hindrance, are coupled and evolve over time, how to objectively characterize the load characteristics based solely on motor load feedback at different time scales, assess the degree of conflict in control requirements between the two abnormal operating conditions, dynamically adjust the priority weight of control objectives, and collaboratively integrate the basic control commands corresponding to each operating state to generate composite control commands, thereby avoiding oscillations and failures caused by single-strategy switching and achieving adaptive and stable control of the sampling process.

[0005] To achieve the above objectives, the technical solution adopted by the present invention is as follows: An adaptive control method for soil sampling based on load feedback includes: During soil sampling, the current and speed signals of the motor are collected as load feedback signals; the load feedback signals are analyzed within short, medium, and long time windows to obtain multi-time window load characteristics. The working mode of the current soil sampling process is determined based on the load characteristics of multiple time windows. The working mode includes hard impact working mode, viscous retardation working mode and normal working mode. Multiple operating states corresponding to each working condition are preset. Each operating state has entry and exit conditions based on multi-time window load characteristics. The switching between operating states is driven by the real-time determined working condition. When the operation state is switched, the current multi-time window load characteristics are input into the dynamic coordinator. According to the predefined conflict arbitration logic, the degree of conflict of control requirements for the two abnormal operating conditions, hard impact and viscous resistance, is evaluated. Based on the long time window load characteristics, the priority weight of each control objective is dynamically adjusted, and the basic control instructions corresponding to each operation state are weighted and fused to generate composite control instructions. Based on the composite control command, the motor is controlled in coordination. The load feedback signal after control is collected, and the load characteristic changes of multiple time windows before and after control are compared to evaluate the processing effect of the composite control command on the current working condition. If the processing effect does not meet the preset conditions, adjust the control parameters and update the composite control command; Soil sample preparation is considered complete when the load feedback signal remains stable for several consecutive long time windows and no longer meets the identification conditions for hard impact or viscous hindrance conditions.

[0006] As a preferred embodiment of the present invention, the multi-time-window load characteristics include: short-time-window load characteristics: current peak amplitude, current rise time change rate, and impact pulse count calculated based on the current signal within the short-time window; medium-time-window load characteristics: current root mean square value, speed fluctuation rate, and correlation coefficient between the current root mean square value sequence and the speed fluctuation rate sequence calculated based on the current signal and speed signal within the medium-time window; and long-time-window load characteristics: slope index of the current mean change trend, load feedback signal fluctuation amplitude index, and load stability index calculated based on the load feedback signal within the long-time window.

[0007] As a preferred embodiment of the present invention, the determination of the operating mode includes: constructing a hard impact strength index based on the current peak amplitude, the rate of change of the current rising edge, and the impact pulse count; constructing a viscous resistance strength index based on the root mean square value of the current, the rotational speed fluctuation rate, and the correlation coefficient; forming a two-dimensional feature space by combining the hard impact strength index and the viscous resistance strength index; obtaining the decision boundary in the two-dimensional feature space by pre-calibrating typical soil samples; mapping the currently calculated hard impact strength index and viscous resistance strength index to the two-dimensional feature space; and outputting the operating mode as hard impact condition, viscous resistance condition, or normal condition according to the decision boundary.

[0008] As a preferred technical solution of the present invention, the switching between the operating states includes: setting entry conditions and exit conditions for each operating state; the entry condition is triggered when the operating mode remains consistent within a consecutive preset number of time windows; the exit condition is triggered when the operating mode does not meet the entry condition of the operating state within a consecutive preset number of time windows; after entering any operating state, a minimum dwell time is set, and switching to other operating states is prohibited within the minimum dwell time.

[0009] As a preferred embodiment of the present invention, the dynamic coordinator includes: a feature input module, used to receive the operating state switching trigger signal and the current multi-time window load characteristics, and to obtain the load evolution trend information characterized by the long-time window load characteristics; a conflict assessment module, used to form impact control demand indicators and damping control demand indicators based on the multi-time window load characteristics according to a predefined conflict arbitration logic, and to output the degree of control demand conflict for two abnormal operating conditions, hard impact and viscous damping; a priority scheduling module, used to dynamically calculate and update the priority weight of each control target according to the degree of control demand conflict and the load evolution trend information; and an instruction fusion module, used to perform weighted fusion of the basic control instructions corresponding to each operating state according to the priority weight, generate a composite control instruction, and output it to the motor control terminal.

[0010] As a preferred embodiment of the present invention, the control of the degree of demand conflict includes: generating an impact demand index based on short-time window load characteristics, and generating a stall demand index based on medium-time window load characteristics and long-time window load characteristics; mapping the impact demand index and stall demand index to a conflict degree level according to a predefined conflict arbitration logic, wherein the conflict degree level includes a first level, a second level, and a third level; establishing a correspondence between the conflict degree level and the priority weight update rules, wherein when the conflict degree level is the first level, the second level, or the third level, the corresponding first group, the second group, or the third group of preset weight update rules are called respectively to update the priority weights of the current limiting control target, the speed maintaining control target, and the torque smoothing control target.

[0011] As a preferred embodiment of the present invention, the generation of the composite control command includes: taking the degree of demand conflict and the load evolution trend characterized by the load characteristics of a long-term window as inputs, setting corresponding priority weights for the current limiting control target, the speed maintenance control target, and the torque smoothing control target respectively, and updating the priority weights according to the degree of demand conflict and the load evolution trend; applying the priority weights to the basic control commands corresponding to each operating state, and weighting and fusing the basic control commands to obtain the composite control command; setting weight boundaries and weight update rate limits for the priority weights, so that the priority weights change within a preset range and the change in adjacent update cycles does not exceed a preset upper limit.

[0012] As a preferred embodiment of the present invention, the evaluation of the processing effect includes: calculating the corresponding multi-time-window load characteristics before and after executing the composite control command; evaluating the degree of mitigation of the hard impact condition based on the impact pulse count and current peak amplitude of the short-time-window load characteristics; evaluating the degree of mitigation of the viscous stagnation condition based on the root mean square value of the current and the speed fluctuation rate of the medium-time-window load characteristics; evaluating the load evolution trend and the degree of stability improvement based on the slope index of the current mean change trend, the load feedback signal fluctuation amplitude index, and the load stability index of the long-time-window load characteristics; and generating an evaluation result of the processing effect of the composite control command on the current condition based on the degree of mitigation of the hard impact condition, the degree of mitigation of the viscous stagnation condition, and the degree of stability improvement.

[0013] As a preferred embodiment of the present invention, the updating of the composite control command includes: when the evaluation result indicates that the composite control command does not meet the preset conditions, recalculating the current multi-time window load characteristics based on the controlled load feedback signal, and updating the demand conflict degree or updating the priority weight based on the multi-time window load characteristics; re-weighting and fusing the basic control commands corresponding to each operating state according to the updated priority weight, updating and outputting the composite control command.

[0014] The soil sampling adaptive control system based on load feedback includes: Load characteristic module: During the soil sampling process, the current signal and speed signal of the motor are collected as load feedback signals; the load feedback signals are analyzed to obtain multi-time window load characteristics; Operating condition identification module: Determines the operating condition mode of the current soil sampling process based on multi-time window load characteristics; Operation switching module: Presets multiple operating states corresponding to each operating mode, and drives the switching between operating states according to the real-time determined operating mode; Instruction generation module: When triggered by a change in running state, the current multi-time-window load characteristics are input into the dynamic coordinator to generate composite control instructions; Command evaluation module: Based on the composite control command, the motor is controlled in coordination, the load feedback signal after control is collected, the load characteristic changes of multiple time windows before and after control are compared, and the processing effect of the composite control command on the current working condition is evaluated. Instruction update module: If the processing effect does not meet the preset conditions, adjust the control parameters and update the composite control instruction; Sampling determination module: Soil sampling is determined to be complete when the load feedback signal is stable for several consecutive long time windows and no longer meets the identification conditions for hard impact or viscous hindrance conditions.

[0015] The present invention has the following advantages: This invention collects motor current and speed signals as load feedback signals during soil sampling and performs joint analysis within short, medium, and long time windows to obtain multi-time window load characteristics. This enables simultaneous characterization of instantaneous impacts, continuous stagnation, and long-term evolution trends, improving the comprehensiveness and robustness of the perception of changes in sample preparation load.

[0016] This invention inputs multi-time-window load characteristics into a dynamic coordinator when the operating state is switched, and evaluates the degree of conflict between control requirements for two abnormal operating conditions, hard impact and viscous hindrance, based on a predefined conflict arbitration logic. This makes control decisions oriented towards the coupling of abnormal operating conditions rather than a single operating condition, thereby avoiding the problem of strategies canceling each other out or amplifying risks.

[0017] This invention dynamically adjusts the priority weights of each control objective based on the load characteristics of a long time window, and generates composite control instructions by weighted fusion of the basic control instructions corresponding to each operating state. This achieves dynamic coordination of control objectives such as current limiting, speed maintenance, and torque smoothing, and balances equipment protection and blockage clearing efficiency in scenarios with concurrent impact and stagnation.

[0018] This invention evaluates the processing effect by coordinating motor control based on composite control commands and collecting load feedback signals after control to compare the changes in load characteristics before and after control across multiple time windows. This forms an effect-oriented closed-loop verification mechanism, improving the adaptability and effectiveness of composite control commands to the current operating conditions. Attached Figure Description

[0019] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are only schematic diagrams of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort. Figure 1 This is a schematic diagram of the soil sampling adaptive control system based on load feedback used in an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0021] Example 1, an adaptive control method for soil sampling based on load feedback, includes the following steps: In this embodiment, the current signal is the measured current value of the motor drive circuit, and the speed signal is the measured speed value of the motor shaft. The load feedback signal is continuously collected throughout the soil sampling process and forms time series data with a fixed sampling period. To ensure the consistency of feature calculation, the collected current signal and speed signal are first time-aligned so that the current sampling value and speed sampling value at the same sampling time constitute the same load feedback sampling point; then, data preprocessing is performed, which includes at least removing obvious outliers, smoothing and filtering high-frequency measurement noise, and marking the removal of lost sampling segments.

[0022] The short, medium, and long time windows are three types of analysis windows defined on the same sampling sequence. The short time window is used to characterize the instantaneous abrupt changes in the load feedback signal, the medium time window is used to characterize the continuous fluctuations and stagnation characteristics of the load feedback signal, and the long time window is used to characterize the evolution trend and stability characteristics of the load feedback signal. All three time windows are updated in a sliding manner: when a new sampling point arrives, the window scrolls forward and updates the corresponding statistics, ensuring that the multi-time-window load characteristics are continuously output during soil sampling and remain synchronized with real-time load changes. To ensure data comparability during the sampling process of the same batch of soil samples, the window length and update step size of the short, medium, and long time windows remain unchanged within that batch.

[0023] The multi-time-window load characteristics include: short-time-window load characteristics: current peak amplitude, current rise time change rate, and impact pulse count calculated based on the current signal within the short-time window; medium-time-window load characteristics: current root mean square value, speed fluctuation rate, and correlation coefficient between the current root mean square value sequence and the speed fluctuation rate sequence calculated based on the current signal and speed signal within the medium-time window; and long-time-window load characteristics: slope index of the current mean change trend, load feedback signal fluctuation amplitude index, and load stability index calculated based on the load feedback signal within the long-time window.

[0024] In this embodiment, the source and meaning of the above features are further defined as follows.

[0025] Short-time window load characteristics (used to characterize instantaneous impact): The current peak amplitude is the deviation of the instantaneous extreme value of the current signal within the short-time window from the baseline value, where the baseline value is statistically obtained from the steady-state segment of the current signal within the short-time window. The current peak amplitude is used to characterize the transient load impact intensity caused by hard particles, quartz sand, or small gravel entering the sample preparation chamber in the soil sample. The current rise edge change rate is the rate of change of the current signal during the peak rise phase within the short-time window, used to characterize the steepness of the load rise when the impact occurs. To ensure consistent comparability of this change rate, the start and end points of the rise edge are determined by a preset ratio of the current signal reaching the baseline value to reaching the peak value, thereby avoiding misjudgment caused by noise points. The impact pulse count is the number of pulses within the short-time window that meet the condition that "the peak amplitude exceeds the preset peak judgment threshold and the duration falls within the preset pulse width range". The impact pulse count is used to describe the frequency of impact events and participates in the construction of hard impact intensity index in subsequent steps to evaluate the degree of impact mitigation. In this embodiment, the identification of "peaks" adopts the same judgment criteria: the steady-state statistics of the current signal within a short time window are used as a reference, and the impact pulse is identified based on the dual conditions of amplitude threshold and duration, so that the current peak amplitude, the rate of change of current rise edge and the impact pulse count are consistent and mutually verifiable.

[0026] The load characteristics within the intermediate time window (used to characterize persistent stagnation and fluctuations) are as follows: The root mean square (RMS) value of the current is an energy-type statistical measure of the current signal within the intermediate time window, reflecting the average intensity of the motor output load at the intermediate time scale. The RMS value of the current participates in the construction of the viscous stagnation intensity index in subsequent steps and is used to characterize the degree of viscous stagnation mitigation. The speed fluctuation rate is the degree of fluctuation of the speed signal relative to its mean within the intermediate time window, reflecting the speed instability caused by material adhesion, agglomeration, or local blockage during sample preparation. The speed fluctuation rate participates in the construction of the viscous stagnation intensity index in subsequent steps and is used to characterize the degree of viscous stagnation mitigation. The correlation coefficient (Pearson correlation coefficient) is a measure of the correlation between the "RMS value sequence of current" and the "speed fluctuation rate sequence" on the same intermediate time window update sequence. The RMS value sequence of current is calculated by rolling between adjacent intermediate time windows, and the speed fluctuation rate sequence is calculated by rolling between adjacent intermediate time windows; the two sequences correspond point-by-point under the same update cycle. The correlation coefficient is used to characterize the consistency relationship between "load intensity change" and "speed fluctuation change". In subsequent steps, it serves as a component of the viscous hindrance strength index to enhance the objectivity of working condition identification: when the soil sample exhibits continuous viscous hindrance, the root mean square value of the current and the speed fluctuation rate show synchronous change characteristics in time evolution, thereby providing a stable criterion for hindrance identification by the correlation coefficient.

[0027] Long-term window load characteristics (used to characterize trends and stability): The slope index of the current mean change trend represents the rate of change of the current mean over time within the long-term window, reflecting the continuous increase or decrease of load over a long time scale. This slope index serves as a basis for load evolution trend information in subsequent steps, driving the dynamic adjustment of priority weights and evaluating the degree of improvement in the load evolution trend. The load feedback signal fluctuation amplitude index represents the fluctuation range of the load feedback signal within the long-term window, characterizing the overall fluctuation of the load waveform over a long time scale, reflecting whether the sample preparation process is in a strong disturbance stage or tending to a stable stage. This fluctuation amplitude index is used to evaluate the degree of stability improvement and, together with the load stability index, constitutes the basis for long-term window stability evaluation. The load stability index represents the degree to which the load feedback signal meets the stability criteria within the long-term window. The stability criteria include that the fluctuations of the current signal and speed signal within the long-term window are within a preset range and the trend change does not exceed a preset threshold. The load stability index supports the effect evaluation and update trigger judgment of subsequent steps, and provides a consistent data source for the determination of sample preparation completion of "stability within multiple consecutive long-term windows".

[0028] This embodiment uses the same batch of samples as a background: the soil sample contains a certain proportion of fine-grained cohesive components and a small amount of hard particles. In the initial stage of sample preparation, the current signal shows intermittent spikes with steep rising edges, and the impact pulse count increases within a short time window; as the cohesive aggregates enter the shear zone, the root mean square value of the current increases within a medium time window, and the rotational speed fluctuation rate increases, accompanied by the correlation coefficient showing stable synchronous characteristics; on a longer time scale, the slope index and fluctuation amplitude index of the current mean change trend reflect the evolution process of the load from strong disturbance to convergent stability, and the load stability index meets the stability criterion as the stable stage arrives.

[0029] Step S2: Determine the current working condition mode of the soil sampling process based on the multi-time window load characteristics. The working condition mode includes hard impact working condition, viscous retardation working condition and normal working condition. In this embodiment, the operating condition mode is a discretized representation of the load state during the current sample preparation process. The determination of the operating condition mode uses multi-time window load characteristics as the sole data source and is updated continuously with time windows during the sample preparation process: when new short, medium, and long time window characteristics are updated, the operating condition mode output is refreshed synchronously, thereby ensuring that the operating condition mode is consistent with real-time load changes.

[0030] The determination of the operating mode includes: constructing a hard impact strength index based on the current peak amplitude, the rate of change of the current rising edge, and the impact pulse count; constructing a viscous resistance strength index based on the root mean square value of the current, the rotational speed fluctuation rate, and the correlation coefficient; forming a two-dimensional feature space by combining the hard impact strength index and the viscous resistance strength index; obtaining the decision boundary in the two-dimensional feature space by pre-calibrating typical soil samples; mapping the currently calculated hard impact strength index and viscous resistance strength index to the two-dimensional feature space; and outputting the operating mode as hard impact condition, viscous resistance condition, or normal condition according to the decision boundary.

[0031] In this embodiment, the meanings and generation processes of the above-mentioned "hard impact strength index", "viscous resistance strength index", "two-dimensional feature space" and "decision boundary" are further defined as follows.

[0032] The hard impact strength index is a scalar index composed of short-time-window load characteristics. Its input data includes current peak amplitude, current rise time rate of change, and impact pulse count. The hard impact strength index is used to uniformly characterize the impact intensity and frequency within a short time scale. The current peak amplitude and current rise time rate of change characterize the impact amplitude and steepness of a single impact event, while the impact pulse count characterizes the frequency of impact events within a short time window. To ensure the comparability of sample preparation data from different batches, this embodiment uses the same dimensional processing method for the above three input quantities: a reference range for each input quantity is obtained during the pre-calibration stage, and normalization or piecewise mapping is performed accordingly. Then, the processed three quantities are combined according to a preset synthesis rule to generate the hard impact strength index, ensuring that the hard impact strength index maintains consistent numerical semantics across different soil sample batches and different operational stages.

[0033] The viscous hindrance strength index is a scalar index composed of load characteristics within a medium time window. Its input data include the root mean square (RMS) value of current, rotational speed fluctuation rate, and correlation coefficient. This index is used to uniformly characterize the characteristics of continuous load increase, intensified speed fluctuation, and hindrance synchronicity. The RMS value of current characterizes the load intensity level at a medium time scale, the rotational speed fluctuation rate characterizes the degree of operational instability at a medium time scale, and the correlation coefficient characterizes the consistency between the RMS current value sequence and the rotational speed fluctuation rate sequence during the rolling update process. To ensure that the correlation coefficient has a clear meaning in operating condition identification, in this embodiment, the calculation of the correlation coefficient is strictly based on two sequences of data under the same medium time window update cycle, and it uses the same time alignment and preprocessing caliber as the RMS current value and rotational speed fluctuation rate to avoid correlation deviations caused by sequence asynchrony. The viscous hindrance strength index also obtains a reference range and completes normalization or segmented mapping processing during the pre-calibration stage, and is then generated according to preset synthesis rules, maintaining a consistent discrimination scale across different soil sample batches.

[0034] The two-dimensional feature space is composed of the hard impact strength index and the viscous resistance strength index, with their coordinates corresponding to the magnitudes of the impact strength and resistance strength, respectively. Each operating condition mode judgment corresponds to a mapping point in the two-dimensional feature space, which is jointly determined by the hard impact strength index and the viscous resistance strength index obtained from the current rolling calculation.

[0035] The decision boundary is obtained through pre-calibration of typical soil samples. During pre-calibration, several representative soil sample types are selected and prepared using a standardized process on a sample preparation device. The corresponding multi-time-window load characteristics are collected and calculated. Further calculations are performed to obtain the hard impact strength index and viscous resistivity index at each sampling time. The pre-calibrated samples are then assigned operating condition labels (hard impact condition, viscous resistivity condition, or normal condition), thus forming a labeled sample distribution in a two-dimensional feature space. Based on this sample distribution, a decision boundary is generated to classify the three operating condition modes. This decision boundary is stored in a form directly applicable to online decision-making, including a boundary parameter set or a boundary lookup table data structure. During online operation, the decision boundary serves as the basis for judgment, without introducing additional data sources.

[0036] At any decision point during the sample preparation process, the current hard impact strength index and viscous resistance strength index are combined to form a mapping point and mapped onto the two-dimensional feature space. Then, the positional relationship between the mapping point and the decision boundary is determined: when the mapping point is located in the hard impact region defined by the decision boundary, the output operating condition mode is hard impact operating condition; when the mapping point is located in the viscous resistance region, the output operating condition mode is viscous resistance operating condition; when the mapping point is located in the normal region, the output operating condition mode is normal operating condition. To ensure consistency between the decision output and subsequent steps, in this embodiment, the operating condition mode output and the running state switching criterion for entering S3 use the same update cycle. The entry counting condition for S3 is triggered when the operating condition mode output results maintain consistency within a continuous time window, thereby achieving a closed-loop match between "operating condition mode judgment - running state switching" on the time scale.

[0037] For example, when a small amount of hard particles in the soil sample enters the sample preparation chamber and collides with the cutting tool or grinding mechanism, the current peak amplitude increases, the rate of change of the current rising edge increases, and the number of impact pulses increases within a short time window, thus increasing the hard impact strength index. When fine-grained viscous components agglomerate in the shear zone and cause continuous hindrance, the root mean square value of the current increases, the rotational speed fluctuation rate increases, and the correlation coefficients maintain a consistent relationship within a medium time window, thus increasing the viscous hindrance strength index. The above two types of strength indices together form mapping points in a two-dimensional feature space, and the corresponding working condition mode is output based on the decision boundary obtained by pre-calibration.

[0038] Step S3: Preset multiple operating states corresponding to each working mode. Each operating state has entry and exit conditions based on multi-time window load characteristics. Drive the switching between operating states according to the real-time determined working mode. In this embodiment, the operating state is a discretized organization of the motor control strategy, used to establish an executable correspondence between "operating mode" and "control execution mode". The operating state does not correspond to a single operating condition label, but rather uses the operating mode as the main index and is further subdivided based on the changes in load characteristics at different time scales. This allows multiple operating states to correspond to the same operating mode to cover load performance of different severity or stages. In this method flow, the operating state serves as the management layer data structure for control execution. Its function is as follows: after S2 outputs the operating mode, it selects the corresponding set of operating states from the preset mapping relationship based on the operating mode, and completes the switching of operating states within or between sets through constraints of entry conditions, exit conditions, and minimum dwell time, thereby providing S4 with a clear switching trigger timing and a source of basic control commands.

[0039] In this embodiment, both the entry and exit conditions are based on the "consistency of the operating condition mode sequence" and the "current value or trend of the multi-time window load characteristics." The operating condition mode sequence consists of the operating condition modes output by S2 in each update cycle, and the multi-time window load characteristics are output by S1 in the same update cycle. To ensure the reproducibility of the operating state switching process, the entry and exit conditions remain fixed within the same sample preparation batch and do not adaptively change during online operation.

[0040] The switching between operating states includes: setting entry and exit conditions for each operating state. The entry condition is triggered when the operating mode remains consistent within a preset number of consecutive time windows. The exit condition is triggered when the operating mode does not meet the entry condition of the operating state within a preset number of consecutive time windows. After entering any operating state, a minimum dwell time is set, and switching to other operating states is prohibited within the minimum dwell time.

[0041] In this embodiment, the aforementioned "continuous preset number of time windows" uses the operating mode update cycle of S2 as the counting unit to form entry and exit counts. The entry count is the cumulative number of consecutive times "the operating mode remains consistent," and when the cumulative value reaches a preset number, it triggers entry into the corresponding operating state. The exit count is the cumulative number of consecutive times "the operating mode does not meet the entry conditions," and when the cumulative value reaches a preset number, it triggers exit from the corresponding operating state. The setting of these entry and exit counts enables the switching of operating states to be performed based on the continuous consistency criterion.

[0042] The minimum dwell time is a time constraint that the running state must remain unchanged after being triggered to enter the running state. Its timing starts at the triggering moment of entering the running state, and the triggering of other running states is blocked before the timing ends.

[0043] In this embodiment, "triggering of running state switching" is defined as: an event signal that the running state switches from the previous running state to the target running state. The event signal includes at least a target running state identifier and a trigger time identifier. The trigger inputs the current multi-time window load characteristics into the dynamic coordinator. Thus, running state switching is not only used for management and control strategies, but also serves as the sole trigger source for the dynamic coordinator's intervention in decision-making, enabling conflict arbitration logic and priority weight adjustment to be executed centrally at the state switching node.

[0044] For example: when the short-time-window impact characteristic caused by hard particles outputs a hard impact condition for multiple consecutive update cycles, the entry count reaches a preset number and triggers entry into the operating state corresponding to the hard impact condition; when the medium-time-window stagnation characteristic caused by viscous agglomeration outputs a viscous stagnation condition for multiple consecutive update cycles, the entry count reaches a preset number and triggers entry into the operating state corresponding to the viscous stagnation condition; when the operating mode output no longer meets the entry conditions and continues to reach the exit count, it triggers exit from the corresponding operating state; after any operating state is entered, the minimum residence time constraint keeps the operating state unchanged for a preset time.

[0045] Step S4: When the running state is switched, the current multi-time window load characteristics are input into the dynamic coordinator. According to the predefined conflict arbitration logic, the degree of conflict of control requirements for the two abnormal operating conditions, hard impact and viscous resistance, is evaluated. Based on the long time window load characteristics, the priority weight of each control target is dynamically adjusted. The basic control instructions corresponding to each running state are weighted and fused to generate composite control instructions. In this embodiment, the "trigger of operating state switching" is an event signal indicating that the operating state has switched from the previous operating state to the target operating state. Upon receiving the event signal, the dynamic coordinator reads the multi-time-window load characteristics corresponding to the trigger time within the same update cycle, using this as input data for the current decision. The dynamic coordinator is a decision-making unit that coordinates the control objectives of two types of abnormal operating conditions: impact and stall. Its function is to convert the multi-time-window load characteristics into executable control decision parameters and establish a unified weight allocation mechanism among the three control objectives of impact protection, stall clearing, and stable operation, thereby avoiding the mutual cancellation or control oscillation caused by fixed strategies based on a single operating condition in abnormally coupled scenarios.

[0046] In this embodiment, the "predefined conflict arbitration logic" is a fixed set of mapping rules. Its inputs are the impact control demand index and the stall control demand index formed by multi-time window load characteristics, and its output is the control demand conflict level. The control demand conflict level is a measure of the mutual constraint strength proposed by two types of abnormal operating conditions on the control objective at the same time. The conflict level includes at least a first level, a second level, and a third level, which are used to select different weight update rule groups in subsequent priority weight adjustments.

[0047] In this embodiment, the "control objectives" are limited to three categories: current limiting control objectives, speed maintenance control objectives, and torque smoothing control objectives. The current limiting control objective is used to constrain the peak current risk under impact conditions; the speed maintenance control objective is used to support continuous propulsion and clearing efficiency under stall conditions; and the torque smoothing control objective is used to suppress mechanical shocks and load fluctuation amplification caused by sudden changes in control commands. The priority weights are a set of weight parameters for the three types of control objectives in this composite control decision, used to quantify the relative priority and contribution ratio of each control objective when generating composite control commands.

[0048] In this embodiment, the "basic control commands" are preset control command sets corresponding to each operating state. These preset control command sets include at least basic command components for three control objectives: current limiting, speed maintenance, and torque smoothing. Their parameter forms include, but are not limited to, current limiting thresholds, target speeds or speed adjustment amounts, and torque change rate limits. The basic control command sets corresponding to each operating state are determined by the operating state mapping relationship and serve as the basic input for weighted fusion by the dynamic coordinator.

[0049] The dynamic coordinator includes: a feature input module, used to receive the operating state switching trigger signal and the current multi-time window load characteristics, and obtain the load evolution trend information characterized by the long-time window load characteristics; a conflict assessment module, used to form impact control demand indicators and damping control demand indicators based on the multi-time window load characteristics according to the predefined conflict arbitration logic, and output the degree of control demand conflict for two abnormal operating conditions, hard impact and viscous damping; a priority scheduling module, used to dynamically calculate and update the priority weight of each control target according to the degree of control demand conflict and the load evolution trend information; and an instruction fusion module, used to perform weighted fusion of the basic control instructions corresponding to each operating state according to the priority weight, generate composite control instructions, and output them to the motor control terminal.

[0050] In this embodiment, the input and output data and meanings of the above modules are further defined as follows.

[0051] The feature input module receives the operating state switching trigger signal and locks the multi-time window load characteristics corresponding to the trigger time, including short-time window load characteristics, medium-time window load characteristics, and long-time window load characteristics. Based on the long-time window load characteristics, the feature input module extracts load evolution trend information. This load evolution trend information is a set of trend data characterizing the load's "continuous increase, continuous relief, or tendency to stabilize" over a long time scale, including at least the slope index of the average current change trend, the load feedback signal fluctuation amplitude index, and the load stability index. This load evolution trend information is used in the priority scheduling module to determine the direction and magnitude constraint of weight updates, ensuring that weight adjustments not only respond to the current impact or resistance intensity but also reflect the phased characteristics of load evolution over time.

[0052] Within the same update cycle, the conflict assessment module generates impact control demand indicators based on short-time-window load characteristics and slack control demand indicators based on medium-time-window and long-time-window load characteristics. The impact control demand indicator quantifies the degree of impact risk, and its input data includes at least the current peak amplitude, current rise time rate, and impact pulse count, enabling it to simultaneously reflect impact amplitude, impact steepness, and impact frequency. The slack control demand indicator quantifies the degree of slack clearing demand, and its input data includes at least the root mean square current value, speed fluctuation rate, correlation coefficient, and trend and stability data for the long-time window, enabling it to simultaneously reflect sustained load levels, speed fluctuations, slack synchronization, and whether slack exhibits a continuously worsening trend. By mapping short-window and medium-to-long-time-window data to impact and slack demands respectively, the conflict assessment module ensures that both types of demand indicators have clear data source boundaries, avoiding semantic ambiguity caused by mixing features from different time scales.

[0053] The conflict assessment module maps the impact control requirement index and the damping control requirement index to conflict severity levels based on predefined conflict arbitration logic. These conflict severity levels are discrete hierarchical results, used to characterize the mutual constraint strength when the two types of control requirements coexist, and are used to select weight update rule groups in the priority scheduling module. Each weight update rule group includes at least three preset weight update rules, corresponding to the first, second, and third levels of conflict severity, respectively. Each preset weight update rule group includes the weight update direction and update amount constraint for the current limiting control target, speed maintenance control target, and torque smoothing control target, ensuring a fixed executable path and reproducibility for the weight update process.

[0054] The control of the degree of demand conflict includes: generating an impact demand index based on short-time window load characteristics, and generating a stall demand index based on medium-time window load characteristics and long-time window load characteristics; mapping the impact demand index and stall demand index to a conflict degree level according to a predefined conflict arbitration logic, wherein the conflict degree level includes a first level, a second level, and a third level; establishing a correspondence between the conflict degree level and the priority weight update rules, wherein when the conflict degree level is a first level, a second level, or a third level, the corresponding first group, second group, or third group of preset weight update rules are called respectively to update the priority weights of the current limiting control target, the speed maintenance control target, and the torque smoothing control target.

[0055] The priority scheduling module receives conflict level and load evolution trend information, and calculates and updates priority weights under the constraints of the corresponding preset weight update rule group. The priority weights are a set of weight parameters that correspond one-to-one with the three types of control objectives. The update process satisfies two types of constraints: "weight boundary" and "weight update rate limit". The weight boundary is used to limit the value range of each weight parameter, so that the weight update is kept within the preset range, and the weight value is used to prevent the control objective from completely failing or being excessively amplified. The weight update rate limit is used to limit the amount of weight change in adjacent update cycles, so that the priority weight changes smoothly in consecutive update cycles, and prevents the composite control command from jumping drastically due to the transition of conflict level in adjacent cycles.

[0056] The instruction fusion module receives the updated priority weights and performs weighted fusion on the basic control instructions corresponding to the current operating state to generate composite control instructions. This weighted fusion combines the instruction components corresponding to current limiting, speed maintenance, and torque smoothing from the basic control instructions according to their priority weights. This ensures that the composite control instructions simultaneously contain constraint information for all three control objectives within the same control cycle, thereby achieving coordinated protection and propulsion in scenarios involving both impact and stall. After generation, the composite control instructions are output to the motor control terminal and then executed and evaluated in S5.

[0057] The generation of the composite control command includes: taking the degree of demand conflict and the load evolution trend characterized by the load characteristics of a long time window as inputs, setting corresponding priority weights for the current limiting control target, the speed maintenance control target, and the torque smoothing control target, and updating the priority weights according to the degree of demand conflict and the load evolution trend; applying the priority weights to the basic control commands corresponding to each operating state, and weighting and fusing the basic control commands to obtain the composite control command; setting weight boundaries and weight update rate limits for the priority weights so that the priority weights change within a preset range and the change in adjacent update cycles does not exceed a preset upper limit.

[0058] For example, when the impact pulse count increases and the current peak amplitude is high within a short time window, while the root mean square value of the current and the speed fluctuation rate increase within a medium time window, and the correlation coefficients maintain a consistent relationship, the conflict assessment module outputs impact control demand indicators based on short window characteristics and stagnation control demand indicators based on medium-to-long window characteristics, and maps the two to conflict degree levels according to conflict arbitration logic; the priority scheduling module updates the priority weights of three control objectives—current limiting, speed maintenance, and torque smoothing—under the constraints of the preset weight update rules corresponding to this level, and combines the load evolution trend information represented by the long window slope index, fluctuation amplitude index, and stability index; the instruction fusion module weights and fuses the basic control instructions corresponding to the current operating state to generate composite control instructions, so that the control instructions maintain the clearing and advancement while suppressing the risk of impact peaks, and suppress the impact of instruction mutations on the mechanical system through torque smoothing constraints, providing consistent data input and a comparable control execution window for the evaluation of the processing effect.

[0059] Step S5: Control the motor in coordination according to the composite control command, collect the load feedback signal after control, compare the load characteristic changes before and after control in multiple time windows, and evaluate the processing effect of the composite control command on the current working condition. The evaluation of the processing effect includes: calculating the corresponding multi-time-window load characteristics before and after executing the composite control command; evaluating the degree of mitigation of hard impact conditions based on the impact pulse count and current spike amplitude of the short-time-window load characteristics; evaluating the degree of mitigation of viscous stagnation conditions based on the root mean square value of current and speed fluctuation rate of the medium-time-window load characteristics; evaluating the load evolution trend and stability improvement degree based on the slope index of the current mean change trend, the load feedback signal fluctuation amplitude index, and the load stability index of the long-time-window load characteristics; and generating an evaluation result of the processing effect of the composite control command on the current operating condition based on the degree of mitigation of hard impact conditions, the degree of mitigation of viscous stagnation conditions, and the degree of stability improvement.

[0060] Step S6: If the processing effect does not meet the preset conditions, adjust the control parameters and update the composite control command; The evaluation results are generated by comparing load characteristics across multiple time windows before and after control, and include at least three types of evaluation information: the degree of mitigation of hard impact conditions, the degree of mitigation of viscous resistance conditions, and the degree of improvement in load evolution trends and stability. The preset conditions are a set of rules used to determine whether the processing effect of this composite control command meets the standards. This set of rules corresponds one-to-one with the three types of evaluation information in the evaluation results, ensuring that the determination of "not meeting the preset conditions" has a clear data input source and a clear judgment criterion, and serves as a trigger signal for updating the composite control command in this step.

[0061] In this embodiment, the "control parameters" are a set of parameters used to generate and constrain composite control commands. These control parameters include at least: limiting parameters for current limiting control targets, speed control parameters for speed maintenance control targets, smoothing constraint parameters for torque smoothing control targets, and weight boundary and weight update rate limiting parameters for priority weight updates. These control parameters participate in the process of invoking the conflict level to the weight update rule and affect the generation result of the composite control command during the weighted fusion process. Therefore, the adjustment of the control parameters in this step directly affects the output of the composite control command in subsequent control cycles, making the update process form an executable closed-loop correction path.

[0062] The update of the composite control command includes: when the evaluation result indicates that the composite control command does not meet the preset conditions, recalculating the current multi-time window load characteristics based on the load feedback signal after control, and updating the demand conflict degree or updating the priority weight based on the multi-time window load characteristics; re-weighting and fusing the basic control commands corresponding to each operating state according to the updated priority weight, updating and outputting the composite control command.

[0063] In this embodiment, the above update process is further defined as follows.

[0064] After executing the composite control command, the controlled load feedback signal is acquired. This controlled load feedback signal shares the same data type and acquisition caliber as the original load feedback signal, both consisting of the motor's current and speed signals, and forming a sequence of data with a fixed sampling period. The phrase "recalculating the current multi-time-window load characteristics" refers to using the same short, medium, and long time windows as S1, and the same preprocessing caliber, to calculate the current peak amplitude, current rise time rate of change, impact pulse count, current root mean square value, speed fluctuation rate, correlation coefficient, slope index of the current mean change trend, fluctuation amplitude index, and load stability index on the controlled load feedback signal. This ensures that the characteristic data before and after the update remain consistent in statistical intervals, update cycle, and physical meaning, meeting the comparability requirements of subsequent update decisions.

[0065] The "updating demand conflict level" refers to the process in the conflict assessment path described in S4, where the impact demand index and the stall demand index are regenerated based on the recalculated multi-time-window load characteristics, and the conflict level is remapped according to the same predefined conflict arbitration logic. This update path is used to handle scenarios where "the coupling relationship of the operating conditions changes": when the relative strength of the short-time-window impact characteristics and the medium-time-window stall characteristics changes after control, the conflict level is updated accordingly, thereby driving the invocation of different weight update rule groups in S4, so that the priority weight allocation of the next control cycle is consistent with the current load state.

[0066] The "updating priority weights" refers to recalculating and updating the priority weights of the current limiting control target, speed maintenance control target, and torque smoothing control target based on the load evolution trend information represented by the recalculated long-term window load characteristics, when the conflict level remains unchanged or does not require reclassification. This update path is used to handle scenarios where "the load evolution trend changes": when the slope index, fluctuation index, or load stability index of the current mean change trend indicates that the load is in a stage of continuous increase or has not yet stabilized, the priority weights are updated according to the corresponding preset weight update rules; at the same time, the weight boundary and weight update rate limit remain effective during the update process, so that the priority weight update result always falls within the preset range and the change in adjacent update cycles does not exceed the preset upper limit, thereby ensuring that the composite control command remains continuous in adjacent control cycles.

[0067] After updating the degree of demand conflict or priority weight, the basic control instructions corresponding to the current operating state are re-weighted and fused according to the updated priority weight to obtain the updated composite control instructions. The "updating and outputting composite control instructions" means that the updated composite control instructions are used as the control input for the next control cycle in S5 for execution, and the load feedback signal after control is collected again and evaluated in subsequent loops, thus forming a closed-loop iterative process of "evaluation-update-re-execution-re-evaluation" until the evaluation result meets the preset conditions.

[0068] For example, if, after a composite control command is executed, the impact pulse count does not decrease within a short time window and the current peak amplitude remains at a high level, while the root mean square value of the current and the speed fluctuation rate remain highly volatile within a medium time window, and the slope index and fluctuation amplitude index indicate that the load has not entered the convergence and stabilization stage within a long time window, then the generated evaluation result indicates that the preset conditions are not met. This triggers this step to recalculate the multi-time window load characteristics of the controlled load feedback signal. Based on the recalculated results, the conflict evaluation path regenerates the impact demand index and the hindrance demand index and updates the conflict degree level, or updates the priority weight based on the load evolution trend information. Subsequently, the basic control command corresponding to the current operating state is re-weighted and fused to generate an updated composite control command, which is then output to the next control cycle for execution, so that the updated control decision is consistent with the impact-hindrance coupled load state of this batch of soil samples, until the processing effect meets the preset conditions.

[0069] Step S7: When the load feedback signal is stable for several consecutive long time windows and no longer meets the identification conditions for hard impact or viscous hindrance conditions, the soil sample preparation is considered complete.

[0070] The "stable operation within multiple consecutive long-term windows" is determined based on the load characteristics of these long-term windows. These load characteristics include the slope of the average current change trend, the fluctuation amplitude of the load feedback signal, and the load stability index. The load stability index characterizes the degree to which the load feedback signal meets the stability criterion within the long-term window. In this embodiment, the "stability criterion" is defined as follows: within the long-term window, the fluctuation amplitudes of the current signal and the speed signal are within a preset range, and the slope of the average current change trend does not exceed a preset trend threshold, thus ensuring the load is in a convergent and stable state.

[0071] The "multiple consecutive long-term windows" uses the long-term window rolling update cycle as the counting unit. When the load stability index continuously meets the stability criterion within a preset number of consecutive long-term windows, it is determined that the condition of "stability within multiple consecutive long-term windows" is met. This continuity requirement is used to avoid triggering the completion determination based on only a single occasional stable segment of a long-term window, and it is consistent with the continuous counting logic of the entry / exit conditions in S3 in terms of data style.

[0072] The "identification conditions" are consistent with the operating condition mode judgment, derived from the two-dimensional feature space and decision boundary judgment results constructed based on multi-time window load characteristics. Specifically, S2 continuously outputs the operating condition mode as hard impact condition, viscous hindrance condition, or normal condition; when the operating condition mode output result corresponding to the S7 judgment time is neither hard impact condition nor viscous hindrance condition in the continuous multiple long time windows, it is determined that "the identification conditions of hard impact condition or viscous hindrance condition are no longer met". The above constraints ensure that the sample preparation completion judgment simultaneously satisfies two types of constraints: on the one hand, the long time window stability index indicates that the load has entered the convergence and stabilization stage; on the other hand, the operating condition mode output no longer falls into the abnormal operating condition region within the continuous window, thereby avoiding the misjudgment of "the load stabilizes in a short time but there are still intermittent impacts or residual hindrance".

[0073] When both conditions are met ("stable within multiple consecutive long-term windows") and ("no longer meeting the identification conditions for hard impact or viscous resistance conditions"), the sample preparation completion judgment result is output, and this result is used as the end marker of the sample preparation process for this batch. This end marker is used to terminate the subsequent composite control command update cycle, so that S6 no longer triggers the update and output of composite control commands, thereby forming an objective end mechanism based on load stability and the fading of abnormal conditions.

[0074] For example: In the early and middle stages of sample preparation, the impact pulse count and current spike amplitude fluctuate within a short time window, while the root mean square value of current and speed fluctuation rate show a continuous increase with synchronous characteristics of correlation coefficient within a medium time window, corresponding to outputting hard impact condition or viscous resistance condition; as the composite control command is executed and iteratively updated, the slope of the current mean change trend within a long time window tends to flatten, the load feedback signal fluctuation amplitude index converges, and the load stability index meets the stability criterion within multiple consecutive long time windows, while continuously outputting normal operating condition within the corresponding window; when the above two conditions are met, the sample preparation completion judgment result is output, and the sample preparation process of this batch ends.

[0075] Example 2, Adaptive control system for soil sampling based on load feedback, see [link / reference] Figure 1 As shown, it includes the following modules: Load characteristic module: During the soil sampling process, the current signal and speed signal of the motor are collected as load feedback signals; the load feedback signals are analyzed to obtain multi-time window load characteristics; Operating condition identification module: Determines the operating condition mode of the current soil sampling process based on multi-time window load characteristics; Operation switching module: Presets multiple operating states corresponding to each operating mode, and drives the switching between operating states according to the real-time determined operating mode; Instruction generation module: When triggered by a change in running state, the current multi-time-window load characteristics are input into the dynamic coordinator to generate composite control instructions; Command evaluation module: Based on the composite control command, the motor is controlled in coordination, the load feedback signal after control is collected, the load characteristic changes of multiple time windows before and after control are compared, and the processing effect of the composite control command on the current working condition is evaluated. Instruction update module: If the processing effect does not meet the preset conditions, adjust the control parameters and update the composite control instruction; Sampling determination module: Soil sampling is determined to be complete when the load feedback signal is stable for several consecutive long time windows and no longer meets the identification conditions for hard impact or viscous hindrance conditions.

[0076] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A soil sampling adaptive control method based on load feedback, characterized in that, include: During soil sampling, the current and speed signals of the motor are collected as load feedback signals; the load feedback signals are analyzed within short, medium, and long time windows to obtain multi-time window load characteristics. The working mode of the current soil sampling process is determined based on the load characteristics of multiple time windows. The working mode includes hard impact working mode, viscous retardation working mode and normal working mode. Multiple operating states corresponding to each working condition are preset. Each operating state has entry and exit conditions based on multi-time window load characteristics. The switching between operating states is driven by the real-time determined working condition. When the operation state is switched, the current multi-time window load characteristics are input into the dynamic coordinator. According to the predefined conflict arbitration logic, the degree of conflict of control requirements for the two abnormal operating conditions, hard impact and viscous resistance, is evaluated. Based on the long time window load characteristics, the priority weight of each control objective is dynamically adjusted, and the basic control instructions corresponding to each operation state are weighted and fused to generate composite control instructions. Based on the composite control command, the motor is controlled in coordination. The load feedback signal after control is collected, and the load characteristic changes of multiple time windows before and after control are compared to evaluate the processing effect of the composite control command on the current working condition. If the processing effect does not meet the preset conditions, adjust the control parameters and update the composite control command; Soil sample preparation is considered complete when the load feedback signal remains stable for several consecutive long time windows and no longer meets the identification conditions for hard impact or viscous hindrance conditions.

2. The soil sampling adaptive control method based on load feedback according to claim 1, characterized in that, The multi-time-window load characteristics include: short-time-window load characteristics: current peak amplitude, current rise time change rate, and impact pulse count calculated based on the current signal within the short-time window; medium-time-window load characteristics: current root mean square value, speed fluctuation rate, and correlation coefficient between the current root mean square value sequence and the speed fluctuation rate sequence calculated based on the current signal and speed signal within the medium-time window; and long-time-window load characteristics: slope index of the current mean change trend, load feedback signal fluctuation amplitude index, and load stability index calculated based on the load feedback signal within the long-time window.

3. The soil sampling adaptive control method based on load feedback according to claim 2, characterized in that, The determination of the operating mode includes: constructing a hard impact strength index based on the current peak amplitude, the rate of change of the current rising edge, and the impact pulse count; constructing a viscous resistance strength index based on the root mean square value of the current, the rotational speed fluctuation rate, and the correlation coefficient; forming a two-dimensional feature space by combining the hard impact strength index and the viscous resistance strength index; obtaining the decision boundary in the two-dimensional feature space by pre-calibrating typical soil samples; mapping the currently calculated hard impact strength index and viscous resistance strength index to the two-dimensional feature space; and outputting the operating mode as hard impact condition, viscous resistance condition, or normal condition according to the decision boundary.

4. The soil sampling adaptive control method based on load feedback according to claim 1, characterized in that, The switching between operating states includes: setting entry and exit conditions for each operating state. The entry condition is triggered when the operating mode remains consistent within a preset number of consecutive time windows. The exit condition is triggered when the operating mode does not meet the entry condition of the operating state within a preset number of consecutive time windows. After entering any operating state, a minimum dwell time is set, and switching to other operating states is prohibited within the minimum dwell time.

5. The soil sampling adaptive control method based on load feedback according to claim 1, characterized in that, The dynamic coordinator includes: a feature input module, used to receive the operating state switching trigger signal and the current multi-time window load characteristics, and obtain the load evolution trend information characterized by the long-time window load characteristics; a conflict assessment module, used to form impact control demand indicators and damping control demand indicators based on the multi-time window load characteristics according to the predefined conflict arbitration logic, and output the degree of control demand conflict for two abnormal operating conditions, hard impact and viscous damping; a priority scheduling module, used to dynamically calculate and update the priority weight of each control target according to the degree of control demand conflict and the load evolution trend information; and an instruction fusion module, used to perform weighted fusion of the basic control instructions corresponding to each operating state according to the priority weight, generate composite control instructions, and output them to the motor control terminal.

6. The soil sampling adaptive control method based on load feedback according to claim 5, characterized in that, The control of the degree of demand conflict includes: generating an impact demand index based on short-time window load characteristics, and generating a stall demand index based on medium-time window load characteristics and long-time window load characteristics; mapping the impact demand index and stall demand index to a conflict degree level according to a predefined conflict arbitration logic, wherein the conflict degree level includes a first level, a second level, and a third level; establishing a correspondence between the conflict degree level and the priority weight update rules, wherein when the conflict degree level is a first level, a second level, or a third level, the corresponding first group, second group, or third group of preset weight update rules are called respectively to update the priority weights of the current limiting control target, the speed maintenance control target, and the torque smoothing control target.

7. The soil sampling adaptive control method based on load feedback according to claim 6, characterized in that, The generation of the composite control command includes: taking the degree of demand conflict and the load evolution trend characterized by the load characteristics of a long time window as inputs, setting corresponding priority weights for the current limiting control target, the speed maintenance control target, and the torque smoothing control target, and updating the priority weights according to the degree of demand conflict and the load evolution trend; applying the priority weights to the basic control commands corresponding to each operating state, and weighting and fusing the basic control commands to obtain the composite control command; setting weight boundaries and weight update rate limits for the priority weights so that the priority weights change within a preset range and the change in adjacent update cycles does not exceed a preset upper limit.

8. The soil sampling adaptive control method based on load feedback according to claim 2, characterized in that, The evaluation of the processing effect includes: calculating the corresponding multi-time-window load characteristics before and after executing the composite control command; evaluating the degree of mitigation of hard impact conditions based on the impact pulse count and current spike amplitude of the short-time-window load characteristics; evaluating the degree of mitigation of viscous stagnation conditions based on the root mean square value of current and speed fluctuation rate of the medium-time-window load characteristics; evaluating the load evolution trend and stability improvement degree based on the slope index of the current mean change trend, the load feedback signal fluctuation amplitude index, and the load stability index of the long-time-window load characteristics; and generating an evaluation result of the processing effect of the composite control command on the current operating condition based on the degree of mitigation of hard impact conditions, the degree of mitigation of viscous stagnation conditions, and the degree of stability improvement.

9. The soil sampling adaptive control method based on load feedback according to claim 8, characterized in that, The update of the composite control command includes: when the evaluation result indicates that the composite control command does not meet the preset conditions, recalculating the current multi-time window load characteristics based on the load feedback signal after control, and updating the demand conflict degree or updating the priority weight based on the multi-time window load characteristics; re-weighting and fusing the basic control commands corresponding to each operating state according to the updated priority weight, updating and outputting the composite control command.

10. A soil sampling adaptive control system based on load feedback, characterized in that, The system employs the soil sampling adaptive control method based on load feedback as described in any one of claims 1 to 9, including: Load characteristic module: During the soil sampling process, the current signal and speed signal of the motor are collected as load feedback signals; the load feedback signals are analyzed in short time windows, medium time windows and long time windows to obtain multi-time window load characteristics; Working condition identification module: Based on the load characteristics of multiple time windows, determine the working condition mode of the current soil sampling process. The working condition mode includes hard impact working condition, viscous retardation working condition and normal working condition. Operation switching module: It presets multiple operating states corresponding to each operating mode. Each operating state has entry and exit conditions based on multi-time window load characteristics. It drives the switching between operating states according to the real-time determined operating mode. Instruction generation module: When triggered by the switching of running states, the current multi-time window load characteristics are input into the dynamic coordinator. According to the predefined conflict arbitration logic, the degree of conflict of control requirements for two abnormal operating conditions, hard impact and viscous resistance, is evaluated. Based on the long time window load characteristics, the priority weight of each control objective is dynamically adjusted. The basic control instructions corresponding to each running state are weighted and fused to generate composite control instructions. Command evaluation module: Based on the composite control command, the motor is controlled in coordination, the load feedback signal after control is collected, the load characteristic changes of multiple time windows before and after control are compared, and the processing effect of the composite control command on the current working condition is evaluated. Instruction update module: If the processing effect does not meet the preset conditions, adjust the control parameters and update the composite control instruction; Sampling determination module: Soil sampling is determined to be complete when the load feedback signal is stable for several consecutive long time windows and no longer meets the identification conditions for hard impact or viscous hindrance conditions.

Images