Automatic tension adjustment control system for rat-proof net laying

Abstract

This invention relates to the field of automation control technology, specifically to an automatic tension adjustment control system for rodent-proof netting. The system includes: a multi-source data acquisition module that simultaneously acquires current, displacement, and vibration spectrum data, and generates a deployment state feature set after Kalman filtering for noise reduction; a state analysis and calculation module that inputs the feature set into a netting deformation mechanics model to calculate the lateral shrinkage rate and identifies the properties of the contact medium based on wave characteristics; and an adaptive impedance control module that dynamically adjusts virtual impedance parameters based on the shrinkage rate and medium properties, generating drive control commands to adjust the output torque and position response. This invention solves the problem of bottom detachment of the netting due to the Poisson effect, achieving tension while maintaining fit.

Description

Technical Field

[0001] This invention relates to the field of automation control technology, specifically to an automatic rodent-proof net tension adjustment control system. Background Technology

[0002] For applications involving the automatic deployment of flexible protective facilities, rodent-proof nets and other nets need to be tensioned and laid on complex terrain by means of motor-driven mechanisms. The working environment is often accompanied by undulating ground and interference from foreign objects.

[0003] Existing deployment control schemes generally employ traditional proportional-integral-derivative (PID) adjustment or tension control strategies based on simplified linear models, lacking deep decoupling from the physical properties of flexible materials. Due to the significant Poisson effect in flexible nets, when longitudinal tension is applied for stretching, the nets inevitably undergo lateral contraction, causing the bottom edge of the net to retract upwards and detach from the ground, compromising the seal integrity. Furthermore, existing technologies struggle to accurately distinguish between normal terrain friction and obstruction from abnormal obstacles using single-dimensional sensor data. Additionally, the original linear model suffers from physical inconsistencies in angle definition and contraction rate calculation when handling large deformation conditions, preventing the system from sensing the net's true deformation and contact state in real time. This often results in localized stress concentration, plastic damage to the net, or seal failure gaps due to excessive contraction during deployment, making it difficult to maintain constant tension while ensuring bottom adhesion.

[0004] Therefore, how to establish a deformation observation mechanism that conforms to physical laws, and dynamically coordinate the torque and position response of the actuator accordingly, so as to balance the tension and sealing effectiveness, has become an urgent technical problem to be solved. Summary of the Invention

[0005] The purpose of this invention is to provide an automatic deployment control system for flexible protective facilities. Given the limitations of existing flexible protective facility deployment control technologies in terms of real deformation perception under complex terrain interference and large deformation conditions of flexible materials, as well as in the coordinated control of tension and sealing fit based on the Poisson effect, there is an urgent need to propose an automatic deployment control system that can establish a deformation observation mechanism that conforms to physical laws and dynamically coordinate the torque and position response of the actuator, so as to take into account both deployment tension and bottom sealing effectiveness.

[0006] To solve the above-mentioned technical problems, the present invention provides an automatic tension adjustment control system for rodent-proof netting. Specifically, the technical solution of the present invention includes:

[0007] The multi-source data acquisition module is used to acquire motor drive current data, encoder displacement data and end vibration spectrum data of the deployed actuator, and to perform time-series alignment and noise reduction on the motor drive current data, encoder displacement data and end vibration spectrum data to generate a deployment state feature set.

[0008] The state analysis and calculation module is connected to the multi-source data acquisition module. It is used to input the deployment state feature set into the preset net deformation mechanical model to calculate the lateral shrinkage rate of the net. Based on the fluctuation characteristics of the end vibration spectrum data and motor drive current data, it identifies the current contact medium properties, including normal terrain friction state and abnormal obstacle jamming state.

[0009] The adaptive impedance control module is used to dynamically adjust the virtual impedance parameters based on the lateral shrinkage rate of the mesh and the properties of the contact medium, and generate drive control commands. The drive control commands are used to adjust the output torque and position response of the deployment actuator to achieve tension while maintaining the fit of the bottom of the mesh.

[0010] In some implementations, the state analysis and solution module includes:

[0011] The deformation prediction unit is used to store the material mechanical parameters of the laid net, including Young's modulus and Poisson's ratio, and calculates the lateral shrinkage rate of the net based on the tension value mapped from the motor drive current data and the material mechanical parameters.

[0012] The state observation unit is used to construct a tension-displacement observer, calculate the deviation between the theoretically set displacement and the actual encoder displacement, and determine the properties of the contact medium by combining the high-frequency component characteristics of the end vibration spectrum data.

[0013] The risk assessment unit is configured to compare the lateral shrinkage rate of the net with a preset effective coverage redundancy threshold. If the lateral shrinkage rate of the net is greater than the effective coverage redundancy threshold, a high-risk edge shrinkage risk indicator is generated. If the lateral shrinkage rate of the net is less than or equal to the effective coverage redundancy threshold, a low-risk edge shrinkage risk indicator is generated. The edge shrinkage risk indicator is used to characterize whether the edge of the net tends to detach from the ground due to excessive stretching.

[0014] In some implementations, the adaptive impedance control module includes:

[0015] A variable stiffness decision unit is used to adjust the target stiffness coefficient based on the edge retraction risk indicator in response to the contact medium properties being in a normal terrain friction state.

[0016] When the risk of edge retraction is identified as low, the target stiffness coefficient is set as the first stiffness value to implement a force control priority strategy and maintain constant tension.

[0017] When the risk of edge retraction is identified as high, the target stiffness coefficient is set to a second stiffness value, which is less than the first stiffness value, to implement a position control priority strategy, allowing the net to elastically retract to conform to the ground depression.

[0018] In some implementations, the adaptive impedance control module further includes:

[0019] The abnormal response unit is used to immediately interrupt the current tension gain logic and generate a reverse release command in response to the abnormal obstacle jamming state of the contact medium. The reverse release command is used to drive the motor to output a reverse torque to eliminate local stress concentration in the net.

[0020] In some implementations, the multi-source data acquisition module includes:

[0021] The spectrum feature extraction unit is used to perform fast Fourier transform on the end vibration spectrum data, extract the energy value of the feature frequency band, calculate the similarity index between the energy value and the preset tension state spectrum template, and correct the tension evaluation value in the deployment state feature set based on the similarity index.

[0022] In some implementations, the system further includes:

[0023] The visual feedback compensation module is used to acquire image data of the bottom edge of the net in the deployment area, calculate the physical gap value between the net edge and the ground, and feed the physical gap value back to the adaptive impedance control module.

[0024] The adaptive impedance control module is also configured to: if the physical gap value exceeds the preset sealing threshold, reduce the damping coefficient in the virtual impedance parameter to enhance the system's ability to respond to terrain undulations; if the physical gap value does not exceed the preset sealing threshold, maintain the current damping coefficient.

[0025] In some implementations, the adaptive impedance control module is configured with a staged control strategy:

[0026] In response to the encoder displacement data being less than the preset deployment length threshold, it is determined to be the deployment start-up stage, and the virtual impedance parameter is configured to a low stiffness mode to adapt to the initial deployment inertia of the mesh.

[0027] When the encoder displacement data reaches the unfolded length threshold and the fluctuation range of the motor drive current data is within the preset stable range, it is determined to be in the deployment locking stage, and the virtual impedance parameter is configured to high stiffness mode.

[0028] Among them, the stiffness coefficient of the high stiffness mode is greater than that of the low stiffness mode. In the high stiffness mode, the lateral shrinkage rate of the mesh is used as the upper limit of the saturation constraint of the output torque to prevent plastic damage to the mesh caused by static over-tension.

[0029] In some implementations, the system further includes:

[0030] The cloud-based data management module is used to receive deployment status feature sets and drive control commands, and to record the tension change curves and terrain coupling characteristics during the deployment process based on time series.

[0031] The model optimization unit is used to iteratively correct the deformation mechanics model of the mesh body based on the coupling characteristics of tension change curve and terrain, so as to update the material mechanical parameters.

[0032] Compared with the prior art, the present invention has at least the following beneficial effects:

[0033] 1. This invention effectively solves the technical problem of the bottom edge of a flexible mesh detaching from the ground due to lateral contraction during the stretching process by the synergistic effect of a multi-source data acquisition module, a state analysis and calculation module, and an adaptive impedance control module. The system inputs the deployment state feature set into a preset mesh deformation mechanical model, accurately calculates the lateral contraction rate of the mesh, and compares the lateral contraction rate of the mesh with a preset effective coverage redundancy threshold to generate an edge retraction risk indicator that represents the tendency to detach from the ground.

[0034] Based on this identifier, the system dynamically adjusts the virtual impedance parameters: when the edge retraction risk identifier is low risk, the target stiffness coefficient is set to the first stiffness value to implement a force control priority strategy and maintain constant tension; when the edge retraction risk identifier is high risk, the target stiffness coefficient is set to the second stiffness value to implement a position control priority strategy, allowing the net to elastically retract to conform to the ground depression; this adaptive adjustment mechanism, which dynamically adjusts the virtual impedance parameters based on the lateral shrinkage rate of the net and the properties of the contact medium, breaks the open-loop traction of traditional control schemes without feedback state perception, and maximizes the fit and sealing effectiveness of the bottom of the net while maintaining the deployment tension;

[0035] 2. This invention achieves accurate perception and proactive safety protection of interference states in complex operating environments; based on the fluctuation characteristics of end vibration spectrum data and motor drive current data, the system can accurately identify the current contact medium properties, including normal terrain friction state and abnormal obstacle jamming state; in response to the contact medium property being the abnormal obstacle jamming state, the abnormal response unit can immediately interrupt the current tension gain logic and generate a reverse release command to quickly eliminate local stress concentration in the net body;

[0036] Meanwhile, the visual feedback compensation module dynamically reduces or maintains the damping coefficient in the virtual impedance parameters by calculating the physical gap between the edge of the net and the ground, which greatly enhances the system's ability to respond to terrain undulations. Combined with the mechanism in the phased control strategy that uses the lateral contraction rate of the net as the upper limit of the output torque saturation constraint, this invention effectively prevents plastic damage to the net caused by static over-tension, and significantly improves the deployment safety and adaptability of the system in complex terrain. Attached Figure Description

[0037] The following is a brief description of the invention. The accompanying drawings are used to illustrate the preferred embodiments of the invention and are for illustrative purposes only and not for limitation.

[0038] Figure 1 This is a structural diagram of an automatic rodent-proof net tension adjustment control system according to the present invention. Detailed Implementation

[0039] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0040] Example 1:

[0041] like Figure 1 As shown, an automatic rodent-proof net tension adjustment control system includes:

[0042] The multi-source data acquisition module is used to acquire motor drive current data, encoder displacement data and end vibration spectrum data of the deployed actuator, and to perform time-series alignment and noise reduction on the motor drive current data, encoder displacement data and end vibration spectrum data to generate a deployment state feature set.

[0043] The state analysis and calculation module is connected to the multi-source data acquisition module. It is used to input the deployment state feature set into the preset net deformation mechanical model to calculate the lateral shrinkage rate of the net. Based on the fluctuation characteristics of the end vibration spectrum data and motor drive current data, it identifies the current contact medium properties, including normal terrain friction state and abnormal obstacle jamming state.

[0044] The adaptive impedance control module is used to dynamically adjust the virtual impedance parameters based on the lateral shrinkage rate of the mesh and the properties of the contact medium, and generate drive control commands. The drive control commands are used to adjust the output torque and position response of the deployment actuator to achieve tension while maintaining the fit of the bottom of the mesh.

[0045] This embodiment details the hardware architecture and core control logic of the above system. The system aims to solve the technical problem that the bottom edge of the flexible mesh detaches from the ground due to the lateral contraction caused by the Poisson effect during the stretching process. The multi-source data acquisition module realizes the synchronous acquisition of heterogeneous sensor data through a field-programmable gate array. The module synchronously acquires motor drive current data representing the output torque, encoder displacement data representing the deployment length, and end vibration spectrum data representing the tension mode and contact medium characteristics of the mesh at a sampling rate of 1kHz.

[0046] For the noise reduction step, to ensure the reproducibility of signal processing, the system uses a discrete linear Kalman filter algorithm to process the current data; the system state vector is defined as:

[0047]

[0048] in, This is an estimated value of the actual current. The rate of change of current; The transpose symbol for a matrix or vector is used; the state transition equation and observation equation are constructed as follows:

[0049]

[0050]

[0051] in, The sampling interval is 1ms. and These are process noise and observation noise, respectively. For observation vectors or system observations, their covariance matrix is... and The system is initialized based on the noise floor parameter in the sensor manual; the system obtains the smoothed optimal estimate through recursive updates. As motor drive current data, and using timestamps to align displacement data and vibration data, a standardized deployment state feature set is generated;

[0052] The state analysis and solution module, as the core logic operation module, constructs a network deformation mechanics model based on a simplified finite element analysis. Specifically, this model employs the lumped parameter method to discretize the continuous network into a truss network structure composed of series and parallel viscoelastic mechanical elements. For any node in the network... Establish the dynamic equilibrium equations, and their calculation formulas are as follows:

[0053]

[0054] in, Let j be the acceleration vector of node j. Let j be the velocity vector of node j. Let k be the displacement vector of the neighboring node k. This is the chaining operator for traversing nodes within a neighborhood set. For nodal displacement vectors, For nodes The set of connected neighboring nodes. Let be the equivalent stiffness coefficient of the connection element, i.e., its calculation formula is:

[0055]

[0056] in, For Young's modulus, The cross-sectional area of ​​the wire is... For the unit length, For nodes The equivalent lumped mass is obtained by allocating half of the mass of all elements connected to the node to that node using the lumped mass method. The calculation formula is as follows:

[0057]

[0058] in, Density of the mesh material; For nodes The structural damping coefficient is estimated using a simplified Rayleigh damping form. External loads mapped to boundary nodes based on tension estimates; tension estimates The calculation formula is as follows, derived from the motor current equation:

[0059]

[0060] in, The torque constant of the motor. The reduction ratio of the speed reducer. Let be the drum radius; the mapping relationship is defined as the tension mapping distribution function, and its calculation formula is:

[0061]

[0062] in, The total number of boundary nodes. The unit vector for setting the traction direction;

[0063] The system employs the explicit Euler integral method to solve this system of equations, and sets the convergence criterion as: when the norm of displacement increments of all nodes in adjacent time steps... When the value is less than the preset convergence threshold, the steady-state displacement field is obtained, the iteration stops, and the lateral displacement components of the boundary nodes are aggregated to calculate the lateral shrinkage rate of the mesh, thereby realizing the mapping of sensor data to a physical state that cannot be directly measured.

[0064] Based on this, the specific quantitative calculation logic for identifying the current contact medium properties, its Gaussian distribution characteristics of current fluctuations, and the broadband noise characteristics of the vibration spectrum is as follows: Current Gaussian characteristic calculation: The system for a length of... Current data within the sliding window The formula for calculating statistical kurtosis is:

[0065]

[0066] in, The statistical kurtosis of the current data. For the first in the window One current sampling point, This is the arithmetic mean of the current data within this window;

[0067] Spectral bandwidth characteristic calculation: The spectral flatness of the spectrum is calculated using the following formula:

[0068]

[0069] in, This is the chain multiplication operator. For power spectral density, For discrete frequency points, use index numbers. The total number of valid frequency points involved in the calculation; The calculated spectral flatness;

[0070] Attribute discrimination logic: If The tolerance range is ,and The preset threshold of 0.6 indicates that the current fluctuation conforms to a normal distribution and the vibration signal is close to white noise, which is judged as a normal terrain friction state.

[0071] like For example, a value greater than 10 indicates the presence of a spike pulse, or The preset threshold of 0.2 indicates the presence of significant single-frequency resonance, which is determined to be an abnormal obstacle jamming state.

[0072] If the calculation results do not meet the explicit conditions for the normal terrain friction state and the abnormal obstacle jamming state mentioned above, i.e. and The system executes a state preservation strategy, which maintains the contact medium property determination result of the previous moment to prevent frequent switching of control logic due to signal jitter in the critical state.

[0073] The threshold is obtained as follows: the system pre-collects vibration data of the net body during dragging on a standard flat surface, calculates its spectral flatness distribution characteristics, and takes the lower bound of the 95% confidence interval as the threshold. Simultaneously, vibration data were collected during collisions and friction between the net and typical obstacles, and the upper bound of its spectral flatness distribution characteristics was taken as... This allows for the quantitative differentiation of the properties of the contact medium.

[0074] The adaptive impedance control module replaces the traditional proportional-integral-derivative control. Based on the above calculation results, it dynamically adjusts the stiffness coefficient and damping coefficient, and generates drive control commands to adjust the torque and position response of the actuator.

[0075] Example 2:

[0076] The state analysis and calculation module includes: a deformation prediction unit, which stores the material mechanical parameters of the laid net body, including Young's modulus and Poisson's ratio, and calculates the lateral shrinkage rate of the net body based on the tension value mapped from the motor drive current data and the material mechanical parameters;

[0077] The state observation unit is used to construct a tension-displacement observer, calculate the deviation between the theoretically set displacement and the actual encoder displacement, and determine the properties of the contact medium by combining the high-frequency component characteristics of the end vibration spectrum data.

[0078] The risk assessment unit is configured to compare the lateral shrinkage rate of the net with a preset effective coverage redundancy threshold. If the lateral shrinkage rate of the net is greater than the effective coverage redundancy threshold, a high-risk edge shrinkage risk indicator is generated. If the lateral shrinkage rate of the net is less than or equal to the effective coverage redundancy threshold, a low-risk edge shrinkage risk indicator is generated. The edge shrinkage risk indicator is used to characterize whether the edge of the net tends to detach from the ground due to excessive stretching.

[0079] This embodiment further specifies the internal logic of the state analysis and solution module. The specific solution process of the mesh deformation mechanics model is as follows: The deformation prediction unit pre-stores material mechanical parameters such as Young's modulus and Poisson's ratio of the mesh, and calculates the lateral shrinkage rate of the mesh based on the tension value mapped from the motor drive current data using an iterative algorithm based on the geometric topology of a rhombic mesh. The specific calculation logic is as follows:

[0080] Wire axial strain calculation: based on current tension value The axial elongation of the individual wires constituting the grid is calculated using the following formula:

[0081]

[0082] in, The axial elongation of the wire. The number of mesh openings on the stressed section;

[0083] Mesh geometry update: Let the initial angle between the mesh edge and the horizontal axis be the angle between the mesh and the horizontal line on the ground. Using the geometric compatibility equations in structural mechanics, the angle after deformation is updated based on the Poisson effect. The formula is constructed based on cosine projection contraction logic:

[0084]

[0085] in, For the amplitude limiting function, when Time to take ,when Time to take Otherwise take ; This is the equivalent Poisson's ratio of the grid structure, and its value is related to the material's Poisson's ratio. This embodiment The conversion formula is:

[0086]

[0087] in, To find the minimum value function, For example, a predefined numerical stability regularization term. rad, This is a preset physical saturation limit, for example, 1.5. This is the Poisson's ratio parameter of the mesh material itself; its selection principle is to be slightly larger than the machine precision of the system's floating-point operations to prevent computational overflow caused by a zero denominator; this value is derived based on the geometric limit when the mesh structure is completely flattened, and is used to prevent calculated values ​​from violating topological constraints due to sensor noise; the system performs angle-to-radian conversion before substituting into the calculation; this formula introduces a regularization term. Effectively avoids Approaching 0 The computational deadlock problem caused by a value of 0;

[0088] This correction ensures that, under small-angle conditions, the model can be based on reality. Calculate the Poisson's ratio that satisfies the physical limit and trigger it correctly. Saturation mechanism; Shrinkage rate calculation: The macroscopic lateral shrinkage rate is calculated based on the lateral projection dimensions before and after deformation. The calculation formula is as follows:

[0089]

[0090] in, This refers to the macroscopic horizontal contraction rate;

[0091] The state observation unit constructs a tension-displacement observer, and calculates the deviation between the theoretically set displacement and the actual encoder displacement through integral calculation. This calculation is based on an ideal pure damping reference model, and the calculation formula is as follows:

[0092]

[0093] in, This is the displacement deviation value, in meters (m). The equivalent viscous damping coefficient of the system is derived from the model's preset value; For time variables, This represents the real-time tension value of the netting.

[0094] The specific offline step response identification method is as follows: under the unconstrained suspension state of the net body, a step torque is applied. Measure steady-state velocity ,in accordance with Definition of viscous damping, calculation:

[0095]

[0096] Under the typical operating conditions of this embodiment, the system equivalent viscous damping coefficient obtained after identification The value range is usually between 10 and 50. between; The actual encoder displacement, sourced from sensor feedback, is expressed in meters (m). Furthermore, the above integral formula is explicitly implemented in the digital controller as a discrete accumulation with initial conditions.

[0097]

[0098] in, For the first Tension value at time; lower limit of integration corresponding to deployment start time, initial deviation: Based on this, the system combines the displacement deviation value with the high-frequency component characteristics of the end vibration spectrum data to determine the properties of the contact medium; the specific determination logic is as follows: calculate the high-frequency component energy characteristics, the calculation formula is:

[0099]

[0100] in, High-frequency component energy characteristics, and These are the start and end frequency index values ​​corresponding to the preset high-frequency bands. For the first in the characteristic frequency band spectral amplitude at discrete frequency points; Operators for solving the modulus or absolute value of complex numbers;

[0101] like That is, the preset deviation threshold is 0.05m, and If the static friction noise threshold is preset, it is determined that the motor is in a stalled state and no relative sliding has occurred, which is confirmed as an abnormal obstacle jamming state.

[0102] During this process, a preset deviation threshold is set. The setting is based on the superposition of the inherent mechanical backlash of the system's transmission chain and the maximum elastic deformation of the mesh material before its yield point; static friction noise threshold It is determined by applying a holding torque to the motor in a static environment, measuring and statistically analyzing the peak value of the high-frequency noise energy fed back by the sensor at this time;

[0103] like If so, it is determined to be a normal terrain friction state;

[0104] The risk assessment unit will calculate the lateral shrinkage rate. Compare with the preset effective coverage redundancy threshold; effective coverage redundancy threshold It is calculated based on the limiting geometric configuration of the mesh, and the calculation formula is as follows:

[0105]

[0106] in, To ensure that the mesh does not undergo irreversible plastic distortion, the default half-angle is set to... The threshold calculated in this way can accurately define the critical point at which a risk indicator is generated at the edge of a high-risk or low-risk level.

[0107] Example 3:

[0108] The adaptive impedance control module includes: a variable stiffness decision unit, used to adjust the target stiffness coefficient according to the edge retraction risk indicator in response to the contact medium properties being in a normal terrain friction state; when the edge retraction risk indicator is low risk, the target stiffness coefficient is set to a first stiffness value to execute a force control priority strategy and maintain constant tension; when the edge retraction risk indicator is high risk, the target stiffness coefficient is set to a second stiffness value, wherein the second stiffness value is less than the first stiffness value, to execute a position control priority strategy and allow the net to elastically retract to conform to the ground depression.

[0109] This embodiment details the control strategy of the variable stiffness decision unit and the underlying impedance control algorithm; the system monitors the current contact medium properties and risk indicators;

[0110] When the contact medium properties are under normal terrain friction conditions and the edge recoil risk is identified as low risk, the variable stiffness decision unit will set the target stiffness coefficient. Set as the first stiffness value, for example Prioritize execution control strategies;

[0111] When the edge retreat risk is marked as high risk, the system will Set as the second stiffness value, for example To implement a location control priority strategy;

[0112] To convert the aforementioned stiffness coefficient into actual motor drive commands, the adaptive impedance control module in this embodiment uses the following discretized impedance control law to generate torque commands, the calculation formula of which is:

[0113]

[0114] in, This is the output torque command at time k, in N·m; is the Jacobian transpose matrix of the transmission system, used to map Cartesian space forces to joint space torques; The damping coefficient is usually taken as... To ensure critical damping characteristics, where, The equivalent inertia parameter preset for the system; and These are the reference speed and the actual speed, respectively. The preset system compensation torque is used to counteract the weight of the deployment mechanism itself and the friction of the foundation. The actual position fed back by the encoder. For reference position;

[0115] This embodiment is... The divergence calculation logic adopted is as follows: Force control priority mode: In this case, constant tension needs to be maintained. The system uses an admittance iteration method with force error compensation to correct the reference position in real time. The calculation formula is:

[0116]

[0117] in, The target tension value preset by the system; The dimensionless force loop compensation gain is taken as 0.5; the third term in the formula is obtained by dividing by the stiffness coefficient. This converts the force deviation value into an equivalent position compensation amount, thereby ensuring the consistency of the physical dimensions of the reference trajectory calculation. For the current moment, the tension is measured or estimated.

[0118] Position control priority mode: In this mode, the mesh retraction must be allowed, and the system cuts off the above force-based correction loop. Locked into the theoretical planning trajectory That is, the ideal position obtained by integrating according to the preset deployment speed; its specific digital iteration logic is as follows:

[0119]

[0120] in, The preset constant deployment speed is used, and the integral is accumulated only during the effective period of the deployment command; in this mode, because It was set to a smaller second stiffness value, and No longer fluctuating with tension, the system exhibits the characteristics of a weakly stiff spring, and its calculation formula is as follows:

[0121]

[0122] in, The output torque or equivalent output force in position control priority mode. This is the low stiffness coefficient, i.e., the second stiffness value, when the system is in position control priority mode; when the mesh needs to retract due to terrain depression, i.e. As the torque is reduced, the output torque decreases flexibly, unlike the force control mode which forcibly pulls in the opposite direction to maintain tension, thus achieving the technical effect of allowing the net to elastically retract.

[0123] Using the above formula, the system can combine stiffness adjustment with reference trajectory correction, thereby achieving a smooth transition between position control and constant tension control modes.

[0124] The adaptive impedance control module also includes an abnormal response unit, which is used to immediately interrupt the current tension gain logic and generate a reverse release command in response to the abnormal obstacle jamming state of the contact medium. The reverse release command is used to drive the motor to output a reverse torque to eliminate local stress concentration in the net.

[0125] This embodiment describes the abnormal response unit in the adaptive impedance control module, which is used to handle unexpected physical interference. The system monitors the contact medium properties output by the state analysis and calculation module in real time. In response to the property being determined to be an abnormal obstacle jamming state, such as the mesh mechanically engaging with a steel bar or stone protruding from the ground, the abnormal response unit immediately takes over control. This unit executes interrupt logic to immediately clear the accumulated terms of the integrator, aiming to prevent integrator saturation and output runaway. The system generates a reverse release command, which is an open-loop trapezoidal torque pulse sequence. The specific generation logic is as follows:

[0126]

[0127] in, This represents the final torque value maintained by the motor at the moment the jamming occurs. The reverse release gain coefficient is preset to 0.5, which means that half of the holding torque is output for reverse relaxation, which can eliminate stress and avoid excessive winding. The pulse duration is preset to 300ms;

[0128] The reverse release command drives the motor to output a reverse torque, actively relaxing the net to eliminate local stress concentration at the hook points, and in... After completion, it enters a zero-torque standby state, waiting for manual reset or the next round of deployment instructions.

[0129] Example 4:

[0130] The multi-source data acquisition module includes: a spectrum feature extraction unit, which performs fast Fourier transform on the end vibration spectrum data, extracts the energy value of the feature frequency band, calculates the similarity index between the energy value and the preset tension state spectrum template, and corrects the tension evaluation value in the deployment state feature set based on the similarity index.

[0131] This embodiment refines the spectral feature extraction unit in the multi-source data acquisition module, introducing characteristic frequency offset determination logic. The spectral feature extraction unit performs a fast Fourier transform on the acquired time-domain vibration signal to extract the spectral amplitude vector within the characteristic frequency band from 100Hz to 800Hz. ;

[0132] The system calculates two key indicators:

[0133] Cosine similarity index :calculate With the preset standard tension template The similarity is calculated using the following formula:

[0134]

[0135] The molecule in the formula This represents the dot product of two vectors, where the denominator contains... This represents the operation of finding the magnitude or L2 norm of a vector;

[0136] Spectral centroid offset Calculate the energy centroid frequency of the current spectrum:

[0137]

[0138] in, For the i-th discrete frequency point within the characteristic frequency band, Given its corresponding spectral amplitude, and compared with the centroid frequency of the template. Find the difference, that is The system corrects the tension assessment values ​​in the deployment state characteristic set based on the above indicators. The corrected correction formula is as follows:

[0139]

[0140] in, The corrected tension assessment value; The initial tension is estimated based on the current. The two parameters are the preset dimensionless correction gain coefficients, which are obtained through offline supervised learning: a sample dataset containing multiple tension states is constructed, with the measurement values ​​of industrial-grade tension sensors as labels, and the objective function of minimizing tension assessment error is iteratively solved to ensure that the correction logic is consistent with the actual physical characteristics. This is the normalized frequency offset factor; This is the sign function, used to determine the correction direction;

[0141] The physical meaning is: if the frequency of the centroid shifts in the positive direction, that is... This indicates that the tension in the mesh is too high, because the string vibration frequency is positively correlated with the tension, and the system needs to be adjusted accordingly. To cause the controller to reduce its output;

[0142] If the centroid frequency shifts negatively, it indicates that the mesh is loose, and the system needs to be adjusted. To prompt the controller to increase its output;

[0143] This logic ensures that regardless of similarity... No matter how the tension is reduced, the system can correctly identify whether the tension is too tight or too loose based on the direction of frequency deviation, thereby achieving logical self-consistency in closed-loop control.

[0144] Example 5:

[0145] The visual feedback compensation module is used to acquire image data of the bottom edge of the netting in the deployment area, calculate the physical gap value between the netting edge and the ground, and feed the physical gap value back to the adaptive impedance control module. The adaptive impedance control module is also configured to: if the physical gap value exceeds the preset sealing threshold, reduce the damping coefficient in the virtual impedance parameter to enhance the system's ability to follow terrain undulations; if the physical gap value does not exceed the preset sealing threshold, maintain the current damping coefficient.

[0146] This embodiment adds a visual feedback compensation module, forming a visual-force fusion control architecture; the visual feedback compensation module acquires the bottom edge image data of the net in the deployment area through an industrial camera installed at the rear of the deployment trolley, and uses the Canni edge detection algorithm to extract the pixel coordinates of the bottom edge of the net and the pixel coordinates of the ground contour.

[0147] For the step of calculating the physical gap between the edge of the mesh and the ground, this embodiment adopts a gap calculation model based on camera pitch angle compensation: Planar projection deviation calculation: using a pre-calibrated ground homography matrix To extract the edge pixels of the mesh in the image and ground contour reference points Mapped to ground projection points in the world coordinate system and And calculate the two in The axial direction, i.e., the Euclidean distance perpendicular to the direction of travel. ;

[0148] High geometric reduction: due to It is the perspective projection of the suspended edge of the net on the ground, its actual physical height, i.e., the physical gap value. , and projection deviation The following trigonometric relationship exists, and its calculation formula is:

[0149]

[0150] in, The camera's optical axis is at the mounting pitch angle relative to the ground plane, for example... The system will use angles during calculations. It is the tangent function; it should be noted that this calculation formula is an engineering approximation based on the pinhole imaging model, and is only applicable to the case where the edge of the mesh is located in the center region of the camera's field of view; if accurate measurement of the entire field of view is required, an inverse perspective transformation algorithm based on the camera's intrinsic parameter matrix needs to be introduced to correct distortion in the non-central region;

[0151] The system will calculate The data is fed back to the adaptive impedance control module in real time. Based on this, the adaptive impedance control module adjusts the damping coefficient in the virtual impedance parameters according to the physical gap value. If the physical gap value exceeds the preset sealing threshold, indicating that the mesh is suspended, the system reduces the damping coefficient. The corrected damping coefficient is then applied. The calculation formula is:

[0152]

[0153] in, Based on the basic damping coefficient, To preset the sealing threshold, such as 5mm, The preset sensitivity gain is usually 2.0. This formula shows that the larger the gap, the more severe the damping attenuation, with a minimum attenuation to 20% of the base value. This reduces the system's resistance to speed changes, enhances its ability to follow terrain undulations, and ensures that the net can quickly follow the settling and fit into the ground depression.

[0154] When the physical clearance value does not exceed the preset sealing threshold, the system executes maintenance logic to restore or maintain a normal state; specifically, the system checks the current damping coefficient. :

[0155] like This indicates that the device was previously in a suspended, low-damped state and is now in contact with the ground. Therefore, a preset linear slope is applied, for example, increasing by a certain amount per control cycle. Gradually adjust the damping coefficient back to To avoid oscillations caused by step jumps;

[0156] like Then keep Unchanged; through this dynamic maintenance strategy, it is ensured that the system can smoothly recover to a highly damped state with sufficient disturbance resistance after the net body is in contact with the ground;

[0157] like If the system is determined to be in an unexpected overdamped state, the damping coefficient will be gradually reduced to a preset linear slope. This is to ensure that the sensitivity of the system's dynamic response is restored to the design baseline.

[0158] Example 6:

[0159] The adaptive impedance control module is configured with a phased control strategy: when the encoder displacement data is less than the preset deployment length threshold, it is determined to be the deployment start-up phase, and the virtual impedance parameter is configured to a low stiffness mode to adapt to the initial deployment inertia of the mesh; when the encoder displacement data reaches the deployment length threshold and the fluctuation range of the motor drive current data is within the preset stable range, it is determined to be the deployment lock-up phase, and the virtual impedance parameter is configured to a high stiffness mode.

[0160] Among them, the stiffness coefficient of the high stiffness mode is greater than that of the low stiffness mode, and in the high stiffness mode, the lateral shrinkage rate of the mesh is used as the upper limit of the saturation constraint of the output torque to prevent plastic damage to the mesh caused by static over-tension.

[0161] This embodiment describes the phased control strategy of the adaptive impedance control module and the key torque saturation constraint algorithm; the system monitors encoder displacement data in real time;

[0162] When the displacement data is less than the preset deployment length threshold, such as 90% of the target length, the system determines that it has entered the deployment start-up phase and configures the virtual impedance parameter to low stiffness mode. It is designed to absorb the impact during startup;

[0163] When the displacement data reaches the threshold and the fluctuation range of the motor drive current data is within the preset stable range, for example... After 500ms, the system determines that it has entered the deployment locking phase;

[0164] At this stage, the system switches the virtual impedance parameters to high-stiffness mode. And introduce the lateral shrinkage rate of the mesh. This serves as the upper limit of the saturation constraint for the output torque; the specific saturation constraint is implemented through the following dynamic limiting function, the calculation formula of which is:

[0165]

[0166] in, This represents the maximum allowable torque output at the current moment, expressed in N·m. This refers to the rated maximum torque of the motor or system. The lateral shrinkage rate of the mesh is calculated in real time; A safe shrinkage rate threshold, such as 5%, is used within which torque is not restricted; The critical plastic damage shrinkage rate threshold is, for example, 15%, at which point the forced torque is reduced to zero.

[0167] Final motor output command for This ensures torque saturation constraint in high stiffness mode at the physical level, preventing irreversible plastic damage to the mesh due to over-positioning and locking.

[0168] The system also includes: a cloud data management module, which receives deployment status feature sets and drive control commands, and records the tension change curve and terrain coupling characteristics during the deployment process based on time series; and a model optimization unit, which iteratively corrects the mesh deformation mechanics model based on the tension change curve and terrain coupling characteristics to update the material mechanical parameters.

[0169] The cloud-based data management module records data; the model optimization unit performs parameter identification, extracting and analyzing data fragments from the system during the deployment and locking phase, i.e., the high-stiffness mode phase; during this phase, the net body unfolding length... Fixed, with the motor serving as the active loading source;

[0170] The specific parameter update logic is as follows: Data slicing: Extracting the time window during the locking phase where the motor output torque undergoes a step change. ;

[0171] Physical quantity calculation: Stress increment, the calculation formula is:

[0172]

[0173] in, For tension data; The effective cross-sectional area is the sum of the cross-sectional areas of all bearing wires along the direction of force on the mesh, and its calculation formula is:

[0174]

[0175] in, This refers to the number of longitudinal wires on the cross-section that bear the axial tensile load. This definition eliminates the interference of transverse weft wires perpendicular to the direction of force on the calculation of the cross-sectional area, ensuring that the stress calculation accurately reflects the material response under longitudinal tension and avoiding inaccurate modulus identification due to the misunderstanding that the transverse direction is the vertical direction of the arrangement. The calculation formula is as follows:

[0176]

[0177] in, For the equivalent longitudinal strain increment, The encoder feedback is a small elastic tensile displacement. This represents the total length of the deployment that has been locked at that moment.

[0178] Calculation of measured modulus: The stress-strain slope within this segment is obtained by fitting the least squares method. The calculation formula is as follows:

[0179]

[0180] in, This represents the total number of data points within the time window. and The first The stress and strain values ​​at each data point, in the formula above. and These are the arithmetic mean values ​​of the stress and strain data within the window, respectively.

[0181] Parameter iteration: The Young's modulus is updated using the exponentially weighted moving average method, and its calculation formula is as follows:

[0182]

[0183] in, For the updated Young's modulus, The Young's modulus parameters stored in the system for the previous moment or in history. This is the confidence factor, for example, 0.05;

[0184] In addition, to fully update the material mechanical parameters of the embodiment, the model optimization unit also simultaneously executes Poisson's ratio correction logic: based on the aforementioned time window. Changes in the transverse shrinkage rate of the internally recorded mesh Calculate the equivalent transverse strain increment ; Calculate the measured Poisson's ratio = The system also updates the Poisson ratio parameter stored in the system using the same exponentially weighted moving average method. By limiting identification to the locking phase, the system successfully avoids the technical challenge of distinguishing between rigid body displacement and elastic deformation during the net's unfolding motion, thus ensuring... and The accuracy of the calculation.

[0185] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the inventive concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions disclosed in this application.

Claims

1. A tension adjustment and control system for automatic installation of rodent-proof netting, characterized in that, The system includes: The multi-source data acquisition module is used to acquire motor drive current data, encoder displacement data and end vibration spectrum data of the deployed actuator, and to perform time-series alignment and noise reduction processing on the motor drive current data, encoder displacement data and end vibration spectrum data to generate a deployment state feature set. The state analysis and calculation module is connected to the multi-source data acquisition module. It is used to input the deployment state feature set into the preset net deformation mechanical model to calculate the lateral shrinkage rate of the net. Based on the fluctuation characteristics of the end vibration spectrum data and the motor drive current data, it identifies the current contact medium properties, including normal terrain friction state and abnormal obstacle jamming state. An adaptive impedance control module is used to dynamically adjust virtual impedance parameters based on the lateral shrinkage rate of the mesh and the properties of the contact medium, and to generate drive control commands. The drive control commands are used to adjust the output torque and position response of the deployment actuator to achieve tension while maintaining the fit of the bottom of the mesh.

2. The automatic rodent-proof net tension adjustment control system according to claim 1, characterized in that, The state analysis and solution module includes: The deformation prediction unit is used to store the material mechanical parameters of the laid net body, including Young's modulus and Poisson's ratio, and calculate the lateral shrinkage rate of the net body based on the tension value mapped from the motor drive current data and the material mechanical parameters. The state observation unit is used to construct a tension-displacement observer, calculate the deviation between the theoretically set displacement and the actual encoder displacement, and determine the properties of the contact medium by combining the high-frequency component characteristics of the end vibration spectrum data. The risk assessment unit is configured to compare the lateral shrinkage rate of the net body with a preset effective coverage redundancy threshold; if the lateral shrinkage rate of the net body is greater than the effective coverage redundancy threshold, a high-risk edge shrinkage risk identifier is generated; if the lateral shrinkage rate of the net body is less than or equal to the effective coverage redundancy threshold, a low-risk edge shrinkage risk identifier is generated; the edge shrinkage risk identifier is used to characterize whether the edge of the net body tends to detach from the ground due to excessive stretching.

3. The automatic rodent-proof net tension adjustment control system according to claim 2, characterized in that, The adaptive impedance control module includes: A variable stiffness decision unit is used to adjust the target stiffness coefficient according to the edge retraction risk indicator in response to the contact medium property being the normal terrain friction state. When the risk of edge retraction is identified as low, the target stiffness coefficient is set to the first stiffness value to implement a force control priority strategy and maintain constant tension. When the edge retraction risk is identified as high risk, the target stiffness coefficient is set to a second stiffness value, wherein the second stiffness value is less than the first stiffness value, to execute a position control priority strategy, allowing the net to elastically retract to conform to the ground depression.

4. The automatic rodent-proof net tension adjustment control system according to claim 3, characterized in that, The adaptive impedance control module further includes: An abnormal response unit is used to immediately interrupt the current tension gain logic and generate a reverse release command in response to the contact medium property being an abnormal obstacle jamming state. The reverse release command is used to drive the motor to output a reverse torque to eliminate local stress concentration in the net body.

5. The automatic rodent-proof net tension adjustment control system according to claim 1, characterized in that, The multi-source data acquisition module includes: The spectral feature extraction unit is used to perform fast Fourier transform on the end vibration spectrum data, extract the energy value of the feature frequency band, calculate the similarity index between the energy value and the preset tension state spectrum template, and correct the tension evaluation value in the deployment state feature set based on the similarity index.

6. The automatic rodent-proof net tension adjustment control system according to claim 1, characterized in that, The system also includes: The visual feedback compensation module is used to acquire image data of the bottom edge of the net in the deployment area, calculate the physical gap value between the edge of the net and the ground, and feed the physical gap value back to the adaptive impedance control module. The adaptive impedance control module is further configured to: if the physical gap value exceeds a preset sealing threshold, reduce the damping coefficient in the virtual impedance parameter to enhance the system's ability to respond to terrain undulations; if the physical gap value does not exceed the preset sealing threshold, maintain the current damping coefficient.

7. The automatic rodent-proof net tension adjustment control system according to claim 1, characterized in that, The adaptive impedance control module is configured with a staged control strategy: In response to the encoder displacement data being less than a preset deployment length threshold, it is determined to be in the deployment start-up phase, and the virtual impedance parameter is configured to a low stiffness mode to adapt to the initial deployment inertia of the mesh. When the encoder displacement data reaches the unfolded length threshold and the fluctuation range of the motor drive current data is within a preset stable range, it is determined to be in the deployment locking stage, and the virtual impedance parameter is configured to high stiffness mode. The stiffness coefficient of the high-stiffness mode is greater than that of the low-stiffness mode, and in the high-stiffness mode, the lateral shrinkage rate of the mesh is used as the upper limit of the saturation constraint of the output torque to prevent plastic damage to the mesh caused by static over-tension.

8. The automatic rodent-proof net tension adjustment control system according to claim 2, characterized in that, The system also includes: The cloud data management module is used to receive the deployment status feature set and the drive control command, and record the tension change curve and terrain coupling characteristics during the deployment process based on time series. The model optimization unit is used to iteratively correct the deformation mechanics model of the net body based on the coupling characteristics of the tension change curve and the terrain, so as to update the material mechanical parameters.

Images