A closed-loop test calibration system for the steering synchronization angle of a van-type semi-trailer

Abstract

The present application belongs to the technical field of vehicle steering control, and relates to a van semi-trailer steering synchronization angle closed-loop test calibration system. First, a static steering test is performed on an empty combined vehicle to construct a mapping reference table of the saddle space position and the ideal hinged angle. Then, a multi-condition dynamic steering test is performed on a loaded combined vehicle, and frame deformation data is collected in real time. The deformation data is converted into a hinged angle compensation value through a pre-trained deformation mechanics model, and selective correction of the measured hinged angle is performed based on a deformation compensation threshold value, so as to generate a calibration target to inversely analyze the target saddle position, effectively solving the problem that the prior art fails to fully consider and distinguish the elastic structure deformation that necessarily occurs in the actual dynamic load condition of the van semi-trailer and the real mechanical steering asynchronization fault. Subsequently, multi-condition driving verification data is collected to iteratively update the deformation mechanics model parameters and the compensation threshold value generation strategy, so as to realize closed-loop optimization.

Description

Technical Field

[0001] This invention belongs to the field of vehicle steering control technology and relates to a closed-loop test and calibration system for steering synchronization angle of a box semi-trailer. Background Technology

[0002] Box semi-trailers, with their large load capacity and multi-axle structure, have become core equipment in modern heavy logistics transportation. However, during steering, the synchronization of the steering angles between the tractor and the semi-trailer is a key factor affecting driving safety and handling stability. Inaccurate steering angles can lead to vehicle deviation, abnormal tire wear, and in severe cases, even loss of control and other safety accidents. Therefore, implementing precise closed-loop testing and calibration of steering synchronization angles is crucial.

[0003] Currently, some technical solutions have been attempted to be applied to the closed-loop calibration of steering synchronization angle for box semi-trailers. However, existing technical solutions have significant limitations: they fail to fully consider and effectively distinguish between the elastic structural deformation that inevitably occurs in box semi-trailers under actual dynamic load conditions and the actual mechanical steering asynchrony fault. Specifically, this manifests in the following ways: 1. Existing technologies rely on unloaded static conditions for testing and adjustment, ignoring the elastic deformation of the frame caused by load, road surface, and steering force during actual fully loaded dynamic operation, resulting in the calibration parameters becoming invalid during actual vehicle operation.

[0004] 2. Some existing technologies can recognize that the structure such as the frame may deform and attempt to filter out the influence of frame deformation. However, their core calibration is still based on the assumption that the vehicle is an ideal rigid body. It is easy to mistakenly identify the reasonable elastic deformation of the structure under load as a lack of coordination between the mechanical connection or actuator between the tractor and the trailer. This error compensation not only fails to eliminate the real deviation, but also destroys the inherent correct synchronization relationship, introduces new errors and deteriorates steering performance. Summary of the Invention

[0005] In view of this, in order to solve the problems mentioned in the background technology, a closed-loop test and calibration system for steering synchronization angle of box semi-trailers is proposed.

[0006] The objective of this invention can be achieved through the following technical solution: This invention provides a closed-loop test and calibration system for the steering synchronization angle of a box semi-trailer, comprising: a benchmark establishment module, a dynamic monitoring module, an intelligent decision-making module, a calibration execution module, and a closed-loop optimization module.

[0007] The benchmark establishment module is connected to the dynamic monitoring module, the dynamic monitoring module is connected to the intelligent decision-making module, the intelligent decision-making module is connected to the calibration execution module, and the calibration execution module is connected to the closed-loop optimization module.

[0008] The benchmark establishment module performs static steering tests on unloaded combined vehicles and constructs a mapping benchmark table between the saddle spatial position and the ideal hinge angle by monitoring and calibrating the articulation angle between the tractor and the semi-trailer.

[0009] The dynamic monitoring module performs multi-condition dynamic steering tests on the load-bearing vehicle. It collects deformation data in real time based on the frame strain sensor group and converts the deformation data into hinge angle compensation values ​​through a pre-trained deformation mechanics model.

[0010] The intelligent decision-making module dynamically generates a deformation compensation activation threshold based on load distribution and road environment parameters. Based on the threshold, it selectively corrects the measured articulation angle in multi-condition dynamic steering tests and generates a calibration target.

[0011] The calibration execution module queries the mapping reference table based on the calibration target, reverse analyzes the target saddle position, and drives the three-dimensional actuator to perform spatial pose adjustment.

[0012] The closed-loop optimization module collects multi-condition driving verification data and iteratively updates the deformation mechanics model parameters and compensation threshold generation strategy.

[0013] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention collects the strain data of the frame in real time under multiple working conditions and combines the pre-trained deformation mechanics model to convert the physical deformation into the hinge angle compensation value, which completely overcomes the drawback of the prior art relying on no-load static calibration. It directly models and compensates the elastic deformation of the frame caused by cargo distribution, road surface excitation, etc. in actual full-load operation, ensuring that the calibration parameters truly reflect the steering geometry relationship under dynamic driving environment.

[0014] (2) The present invention constructs a nonlinear deformation-rotation mapping based on strain sensor data and neural network, physically acknowledging the elastic deformation properties of the frame, and using a dynamically generated deformation compensation activation threshold to activate compensation only when the deformation interference exceeds the critical value, thereby avoiding the erroneous logic of the prior art that misjudges reasonable elastic deformation as mechanical asynchrony and preventing damage to the inherent synchronicity of the system.

[0015] (3) The present invention uses multi-condition driving verification data to iteratively update deformation model parameters and compensation threshold strategy, giving it self-evolution capability to improve the robustness and life cycle accuracy of calibration system. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0017] Figure 1 This is a schematic diagram of the system module connections of the present invention.

[0018] Figure 2 This is a logical schematic diagram illustrating the construction of the deformation mechanics model of this invention.

[0019] Figure 3 This is a logical diagram illustrating the reverse analysis of the target saddle position in this invention. Detailed Implementation

[0020] The following description, in conjunction with the implementation of the present invention, is merely an example and illustration of the concept of the present invention. Those skilled in the art may make various modifications or additions to the specific embodiments described, or adopt similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined in the claims, and all such modifications or additions shall fall within the protection scope of the present invention.

[0021] Please see Figure 1 As shown, the objective of this invention can be achieved through the following technical solution: This invention provides a closed-loop test and calibration system for the steering synchronization angle of a box semi-trailer, comprising: a benchmark establishment module, a dynamic monitoring module, an intelligent decision-making module, a calibration execution module, and a closed-loop optimization module.

[0022] The benchmark establishment module is connected to the dynamic monitoring module, the dynamic monitoring module is connected to the intelligent decision-making module, the intelligent decision-making module is connected to the calibration execution module, and the calibration execution module is connected to the closed-loop optimization module.

[0023] The benchmark establishment module performs static steering tests on unloaded combined vehicles and constructs a mapping benchmark table between the saddle spatial position and the ideal hinge angle by monitoring and calibrating the articulation angle between the tractor and the semi-trailer.

[0024] In a preferred embodiment of the present invention, the static steering test includes the following steps: parking the unloaded combined vehicle on a level hard surface, locking the semi-trailer wheels and releasing the tractor brake, and positioning the saddle to a preset initial spatial position.

[0025] It should be noted that the above-mentioned preset initial spatial position is based on the saddle design coordinates provided by the tractor manufacturer. The spatial intersection of the center plane of the traction pin and the symmetrical plane of the frame is measured by laser, and the saddle height is adjusted so that the unloaded ground distance meets the standard value. The three-dimensional coordinates at this moment are locked as the preset initial spatial position.

[0026] The steering wheel angle of the tractor is controlled according to the preset classification rules. Based on the wheelbase of the tractor and the distance from the saddle to the rear axle axis of the semi-trailer, the ideal articulation angle corresponding to each steering wheel angle state is quantified.

[0027] It should be noted that the aforementioned preset grading rule specifically refers to allocating sampling density according to the rate of change of steering angle within the effective steering wheel angle range, and forcibly including the maximum steering angle point in the sampling sequence. The sampling density can be exemplarily allocated as follows: The interval shall not exceed For the sampling interval, in The interval shall not exceed For the sampling interval, in The above range shall not exceed The sampling interval is denoted as .

[0028] It should also be noted that the specific quantification process of the ideal articulation angle corresponding to each of the above-mentioned steering wheel angle states is mainly based on Ackermann steering geometry calculation, and its calculation formula can be exemplified as follows: ,in These represent the wheelbase of the tractor unit and the distance from the saddle to the rear axle axis of the semi-trailer, respectively. This refers to the steering wheel angle.

[0029] The steering wheel angle of the fixed tractor is used to simultaneously collect the measured articulation angle between the tractor and the semi-trailer. The saddle is controlled to adjust its position in three-dimensional space until the measured articulation angle equals the ideal articulation angle. The spatial displacement vector of the saddle relative to its initial position is recorded when the ideal articulation angle is reached.

[0030] In a preferred embodiment of the present invention, the mapping reference table between the saddle position and the ideal hinge angle includes the following associated data items: different steering wheel angle states under the graded control of the tractor.

[0031] The measured hinge angle and the ideal hinge angle corresponding to each rotation state.

[0032] The saddle spatial displacement vector that makes the measured hinge angle match the ideal hinge angle under each rotation angle condition.

[0033] The dynamic monitoring module performs multi-condition dynamic steering tests on the load-bearing vehicle, collects deformation data in real time based on the frame strain sensor group, and converts the deformation data into hinge angle compensation values ​​through a pre-trained deformation mechanics model.

[0034] In a preferred embodiment of the present invention, the multi-condition dynamic steering test includes the following: setting up a combination of conditions including different road surface adhesion conditions, load mass distribution patterns and driving speed parameters.

[0035] It should be noted that the above-mentioned different road surface adhesion conditions include low adhesion, variable adhesion and high adhesion road surface types. Low adhesion road surface can be exemplified as snow and ice covered road surface, waterlogged asphalt road surface or loose gravel road surface. Variable adhesion road surface can be exemplified as a mixed road surface with half hardened and half muddy or a bridge joint transition area. High adhesion road surface can be exemplified as a dry asphalt road surface or concrete road surface.

[0036] The load mass distribution pattern includes axial off-center loading, longitudinal off-center loading, and combined off-center loading patterns under various load weights. The axial off-center loading pattern refers to the cargo mass being concentrated in the front or rear axle area of ​​the semi-trailer. The longitudinal off-center loading pattern refers to the cargo mass being unevenly distributed along the length of the trailer, such as front heavy and rear light or front light and rear heavy. The combined off-center loading pattern refers to the asymmetrical distribution of axial and longitudinal off-center loading.

[0037] The driving speed parameters include low-speed, medium-speed, and high-speed ranges. The low-speed range is used to simulate parking conditions, with an example of 0-20 km / h. The medium-speed range is used to simulate turning conditions on urban roads, with an example of 30-50 km / h. The high-speed range is used to simulate lane changing conditions on highways, with an example of 60-80 km / h.

[0038] Under combined working conditions, the steering wheel angle of the tractor is controlled in stages, and the measured articulation angle between the tractor and the semi-trailer is collected simultaneously under each steering wheel angle condition.

[0039] Please see Figure 2 As shown, in a preferred embodiment of the present invention, the process of constructing the deformation mechanics model is as follows:

[0040] a) Apply multiple sets of controllable deformation loads to the frame structure through a deformation mode generator, simultaneously collect deformation data from the strain sensor group, and use a displacement measuring device to calibrate the relative rotation angle between the rear frame section of the tractor and the front frame section of the semi-trailer as the true value of the hinge angle compensation. Divide the deformation data and the compensation true value of the strain sensor group into a training set and a validation set according to a preset ratio.

[0041] b) Construct a residual neural network with deformation data from the strain sensor group as input and discrete node curvature as output. Substitute the curvature value output by the neural network into the integral formula of the segmented beam to quantify the relative rotation angle between the rear frame section of the tractor and the front frame section of the semi-trailer, and use it as the network output compensation value.

[0042] It should be noted that the execution logic of the above segmented beam integral formula is defined as follows: along the length direction of the rear frame of the tractor, the discrete node curvature values ​​output by the neural network are spatially accumulated. The spatial accumulation operation represents the angular contribution of the curvature distribution to the frame deformation, and the accumulated result is the rotation angle of the rear frame segment of the tractor relative to the saddle connection point.

[0043] Similarly, along the length of the front frame of the semi-trailer, the same spatial accumulation operation is performed on the curvature values ​​of discrete nodes, and the accumulation result represents the rotation angle of the front frame section of the semi-trailer relative to the saddle connection point.

[0044] Calculate the algebraic difference between the turning angle of the rear frame section of the tractor and the turning angle of the front frame section of the semi-trailer. The difference is the hinge angle deviation caused by deformation, which is used as the final output hinge angle compensation value.

[0045] c) Train the neural network using the training set data, and construct the damage function based on the mean square error between the network output compensation value and the compensation true value, as well as the curvature distribution smoothness constraint.

[0046] It should be noted that the above curvature distribution smoothness constraint is achieved by penalizing abrupt changes in the curvature values ​​of adjacent nodes.

[0047] d) Use validation set data to validate the pre-trained neural network. When the damage error is below a preset permissible threshold, freeze the network parameters to complete the deformation mechanics model construction.

[0048] This invention, through real-time acquisition of chassis strain data under multiple load conditions, and the conversion of physical deformation into hinge angle compensation values ​​by combining a pre-trained deformation mechanics model, completely overcomes the drawbacks of existing technologies that rely on no-load static calibration. It directly models and compensates for the elastic deformation of the chassis caused by cargo distribution, road surface excitation, etc. during actual full-load operation, ensuring that the calibration parameters truly reflect the steering geometry under dynamic driving conditions.

[0049] The intelligent decision-making module dynamically generates a deformation compensation activation threshold based on load distribution and road environment parameters. Based on the threshold, it selectively corrects the measured articulation angle in the multi-condition dynamic steering test and generates a calibration target.

[0050] In a preferred embodiment of the present invention, the process of generating the deformation compensation activation threshold includes: analyzing the redistribution features of multiple combined working conditions, road adhesion features, and vehicle motion features respectively.

[0051] It should be noted that the above-mentioned load distribution characteristics are the coupling characteristics of the lateral offset of the center of gravity and the longitudinal position on the frame stiffness, the road adhesion characteristics are the vehicle instability characteristics under the conditions of road surface adhesion coefficient and road curvature radius, and the vehicle motion characteristics are the abnormal fluctuation detection characteristics of vehicle longitudinal acceleration and yaw rate.

[0052] Based on feature correlation, load-sensitive factors, road-sensitive factors, and motion-inhibiting factors are generated.

[0053] It should be noted that the above-mentioned load-sensitive factor refers to the adjustment command value generated to reduce the compensation threshold based on the degree of increase in the lateral offset of the center of gravity or the forward shift of the longitudinal position. The road-sensitive factor refers to the correction command value generated to tighten the compensation threshold when the road surface adhesion coefficient decreases or the road curvature radius decreases. The motion inhibition factor refers to the inhibition command value generated to temporarily increase the threshold if the longitudinal acceleration or yaw rate exceeds the system's preset critical range.

[0054] The load sensitivity factor, road sensitivity factor, and motion inhibition factor are implemented using a combination of hierarchical interval mapping and instruction assignment, as detailed below:

[0055] The spatial coordinates of the semi-trailer's center of gravity under multiple working conditions are collected by the on-board weighing device. By comparing the coordinates with the standard center of gravity coordinates of the semi-trailer, the lateral offset and longitudinal forward movement of the center of gravity are determined. The numerical ranges corresponding to the lateral slight offset, lateral moderate offset, and lateral significant offset levels are defined by the absolute value of the lateral offset, and the numerical ranges corresponding to the longitudinal backward, neutral, and forward levels are defined by the longitudinal forward movement. The lateral and longitudinal level attributes corresponding to the center of gravity offset under multiple working conditions are retrieved. The combination of lateral slight offset and longitudinal backward level, lateral moderate offset and longitudinal neutral level, and lateral significant offset and longitudinal forward level are marked as the first, second, and third level combinations, respectively. The absolute values ​​of the adjustment instructions to reduce the compensation threshold are increased sequentially, for example, -0.1, -0.3, and -0.5.

[0056] The vehicle's road adhesion coefficient is sensed by tire sensors under multiple working conditions, and the road curvature radius under multiple working conditions is identified by digital map. Both the road adhesion coefficient and the road curvature radius have low, medium and high risk level ranges. The absolute values ​​of the correction instructions for tightening the compensation threshold are assigned in descending order of monitoring double high risk, single high risk, double medium risk and other combined risks, which can be exemplified as -0.5, -0.3, -0.2 and -0.1.

[0057] The longitudinal acceleration and yaw rate of the semi-trailer are collected under multiple working conditions. The overshoot of the longitudinal acceleration and yaw rate relative to the preset critical range of the system is quantified. The sum of the overshoots determines the vehicle motion hazard level, including first-order hazard level, second-order hazard level and third-order hazard level. The higher the hazard level, the larger the temporary threshold suppression command value, which can be 0.1, 0.3 and 0.5 for example.

[0058] A deformation compensation decision function is constructed by integrating multiple factors, and the deformation compensation activation threshold corresponding to each combination of working conditions is output.

[0059] It should also be noted that the process of constructing a deformation compensation decision function by integrating multiple factors can include any of the following mechanisms: i. a linear weighted fusion mechanism, in which the linear fusion weights of each factor are allocated according to industry experience to perform weighted calculation.

[0060] ii. Logic gate cascading mechanism: By establishing a factor priority sequence, when a higher-level factor is activated, the influence of lower-level factors is shielded. For example, if the motion inhibition factor triggers an increase threshold instruction, the load and road factors are ignored.

[0061] iii. Fuzzy membership mechanism: Convert each factor instruction into fuzzy linguistic variables, synthesize a fuzzy decision table based on the membership function, and output it after defuzzification.

[0062] Taking the linear weighted fusion mechanism as an example, according to industry experience, the priority of factor weight allocation in linear weighted fusion should follow the order of highest weight for motion inhibition factors, second highest weight for road sensitivity factors, and lowest weight for load sensitivity factors, with a cumulative weight value of 1. The deformation compensation decision function can be expressed as follows: ,in To preset the basic deformation compensation activation threshold, For the first Linear allocation of weights and instruction values ​​for each factor. Number each factor. .

[0063] Assuming a certain semi-trailer truck, under a specific operating condition, exhibits a first-level centroid shift, a double-high-risk road environment, and a second-level vehicle motion hazard, the corresponding command values ​​for the load sensitivity factor, road sensitivity factor, and motion inhibition factor are -0.1, -0.5, and 0.3, respectively, with corresponding weight allocations of 0.2, 0.3, and 0.5. The specific calculation process for the deformation compensation activation threshold is as follows: .

[0064] In a preferred embodiment of the present invention, the selective correction of the measured articulation angle in the multi-condition dynamic steering test based on the threshold includes the following: extracting the measured articulation angle between the tractor and the semi-trailer, the articulation angle compensation value, and the deformation compensation activation threshold in each steering wheel angle state of the dynamic steering test under each working condition.

[0065] If the hinge angle compensation value is greater than or equal to the deformation compensation activation threshold, it indicates that the measured hinge angle is subject to strong deformation interference and needs to be corrected. The correction process involves superimposing the measured hinge angle and the hinge angle compensation value in the direction of driving the measured hinge angle towards the ideal hinge angle convergence, and using the corrected measured hinge angle as the calibration target.

[0066] Conversely, if the hinge angle compensation value is less than the deformation compensation activation threshold, the measured hinge angle will be directly used as the calibration target.

[0067] The calibration execution module queries the mapping reference table based on the calibration target, reverse analyzes the target saddle position, and drives the three-dimensional actuator to perform spatial pose adjustment.

[0068] Please see Figure 3 As shown, in a preferred embodiment of the present invention, the reverse analysis process of the target saddle position includes: analyzing the unit hinge angle deviation-saddle displacement conversion coefficient based on the mapping reference table. The conversion coefficient is calibrated by the mapping relationship between the deviation value of the measured hinge angle and the ideal hinge angle in the static steering test and the saddle displacement vector.

[0069] Based on the conversion coefficient, the deviation between the calibration target and the measured hinge angle under the same steering angle state in the mapping reference table is converted into the saddle space displacement optimization vector.

[0070] Extract the original saddle spatial displacement vectors from the mapping reference table under the same steering angle state, and superimpose the optimized vectors to generate the target displacement vector.

[0071] The target displacement vector is adjusted based on the preset initial spatial position of the saddle to determine the target saddle position.

[0072] The embodiments of the present invention construct a nonlinear deformation-rotation mapping based on strain sensor data and neural network, physically acknowledging the elastic deformation properties of the frame, and using a dynamically generated deformation compensation activation threshold to initiate compensation only when the deformation disturbance exceeds the critical value. This fundamentally avoids the erroneous logic of existing technologies that misjudge reasonable elastic deformation as mechanical asynchrony, and prevents damage to the inherent synchronicity of the system.

[0073] The closed-loop optimization module collects multi-condition driving verification data and iteratively updates the deformation mechanics model parameters and compensation threshold generation strategy.

[0074] In a preferred embodiment of the present invention, the multi-condition driving verification data collection process is as follows: after the three-dimensional actuator completes the saddle posture adjustment, the deviation value between the measured hinge angle and the ideal hinge angle after the saddle adjustment is collected and used as driving verification data.

[0075] In a preferred embodiment of the present invention, the iterative update strategy for deformation mechanics model parameters and compensation threshold generation includes:

[0076] The percentage of driving conditions in multi-condition driving verification data where the calibrated articulation angle exceeds the permissible range is statistically analyzed. When the percentage of such driving conditions is higher than the preset proportion of non-minor driving conditions, an update command is triggered.

[0077] It should be noted that the above-mentioned permissible range is a two-way tolerance range set with the ideal hinge angle in the benchmark establishment module as the benchmark value. Its boundary is determined by vehicle dynamics stability constraints, mainly involving various physical parameters such as vehicle wheelbase, center of gravity height, and tire lateral stiffness. The proportion of non-minority operating conditions refers to the proportion of failure operating conditions that can affect the safety of the system. The specific values ​​of the permissible range and the proportion of non-minority operating conditions are all manually calibrated before the system is developed.

[0078] The pre-trained deformation mechanics model is incrementally trained using the newly added deformation dataset until the validation set error meets the accuracy requirements.

[0079] The load sensitivity factor, road sensitivity factor, and motion inhibition factor in the compensation threshold generation strategy are dynamically adjusted based on the misjudgment rate.

[0080] It should be noted that the above process of dynamically adjusting the compensation threshold generation strategy based on the misjudgment rate may include the following steps: using the ratio of the number of decisions that require deformation compensation but are not activated to the total number of compensation decision triggers as the misjudgment rate.

[0081] The difference between the false positive rate and the preset benchmark false positive rate is used as the numerator, and the preset benchmark false positive rate is used as the denominator. A ratio calculation is performed, and the result of the calculation is used as the relative deviation of the false positive rate.

[0082] The product of the preset learning rate adjustment factor and the relative deviation of the misjudgment rate is used as the factor adjustment amount.

[0083] The sum of the previous generation factor values ​​and the factor adjustment amount is used to generate the next generation factor values.

[0084] The embodiments of the present invention iteratively update the deformation model parameters and compensation threshold strategy based on multi-condition driving verification data, giving it self-evolution capability to improve the robustness and life cycle accuracy of the calibration system.

[0085] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation.

[0086] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0087] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0088] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0089] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0090] Finally, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A closed loop test calibration system for a van semi-truck turning synchronization angle, characterized by, The method comprises the following steps: A reference establishing module performs a static steering test on an empty combined vehicle, monitors and calibrates the articulation angle between the tractor and the semitrailer through intelligent sensors, and constructs a mapping reference table of the saddle spatial position and the ideal articulation angle; A dynamic monitoring module performs a multi-working-condition dynamic steering test on a loaded combined vehicle, collects deformation data in real time based on a frame strain sensor group, and converts the deformation data into articulation angle compensation values through a pre-trained deformation mechanics model; The construction process of the deformation mechanics model is as follows: A deformation mode generator applies multiple groups of controllable deformation loads to the frame structure, synchronously collects deformation data of the strain sensor group, and uses a displacement measuring device to calibrate the relative rotation angle between the rear frame section of the tractor and the front frame section of the semitrailer as the articulation angle compensation true value, and divides the deformation data of the strain sensor group and the compensation true value into a training set and a verification set according to a pre-set ratio; A residual neural network is constructed with the deformation data of the strain sensor group as input and discrete node curvature as output, the curvature value output by the neural network is substituted into the sectional beam integral formula to quantify the relative rotation angle between the rear frame section of the tractor and the front frame section of the semitrailer, and the network output compensation value is obtained; The neural network is trained using the training set data, and a damage function is constructed based on the mean square error of the network output compensation value and the compensation true value and the smoothness constraint of the curvature distribution; The pre-trained neural network is verified using the verification set data, the network parameters are frozen when the damage error is lower than a pre-set permission threshold, and the construction of the deformation mechanics model is completed; An intelligent decision-making module dynamically generates a deformation compensation enabling threshold according to the load distribution and road environment parameters, selectively corrects the measured articulation angle in the multi-working-condition dynamic steering test based on the threshold, and generates a calibration target; A calibration execution module queries the mapping reference table according to the calibration target, reversely analyzes the target saddle position, and drives a three-dimensional execution mechanism to perform spatial pose adjustment; A closed-loop optimization module iteratively updates the deformation mechanics model parameters and the compensation threshold generation strategy by collecting multi-working-condition driving verification data.

2. The closed-loop test calibration system for the synchronization angle of a van-type semi-trailer according to claim 1, characterized in that: The static steering test comprises the following steps: The empty combined vehicle is parked on a horizontal hard road surface, the semitrailer wheels are locked and the tractor brakes are released, and the saddle is positioned to a pre-set initial spatial position; The steering wheel rotation angle of the tractor is controlled according to a pre-set grading rule, the ideal articulation angle corresponding to each steering wheel rotation angle state is quantified based on the wheelbase of the tractor and the distance from the saddle to the rear axle line of the semitrailer; The steering wheel rotation angle of the tractor is fixed, the measured articulation angle between the tractor and the semitrailer is synchronously collected, the saddle performs position adjustment in three-dimensional space until the measured articulation angle is equal to the ideal articulation angle, and the spatial displacement vector of the saddle relative to the initial position when the ideal articulation angle is reached is recorded.

3. The closed loop test calibration system for the synchronization angle of a van-type semi-trailer according to claim 2, characterized in that: The mapping reference table of the saddle position and the ideal articulation angle comprises the following associated data items: Different steering wheel rotation angle states under the control of the tractor; The measured articulation angle and the ideal articulation angle corresponding to each rotation angle state; The saddle spatial displacement vector that matches the measured articulation angle with the ideal articulation angle under each rotation angle state.

4. The closed loop test calibration system for the synchronization angle of a van-type semi-trailer according to claim 3, characterized in that: The multi-working-condition dynamic steering test comprises the following contents: Setting a combined working condition including different road surface adhesion conditions, load mass distribution patterns and driving speed parameters; In the combined working condition, the steering wheel angle of the tractor is controlled in stages, and the actual articulation angle between the tractor and the semitrailer is synchronously collected under the steering wheel angle of each steering wheel.

5. The closed loop test calibration system for the synchronization angle of a van-type semi-trailer according to claim 1, characterized in that: The generation process of the deformation compensation enabling threshold value includes: The redistribution characteristics, road adhesion characteristics, and vehicle motion characteristics under multiple combined working conditions are analyzed respectively. Based on the characteristic correlation, load-sensitive factors, road-sensitive factors, and motion inhibition factors are generated. The deformation compensation decision function is constructed by fusing multiple factors, and the deformation compensation enabling threshold value corresponding to each combined working condition is output.

6. The closed loop test calibration system for the synchronization angle of a van-type semi-trailer according to claim 4, characterized in that: The selective correction of the actual articulation angle in the multi-working-condition dynamic steering test based on the threshold value includes the following contents: The actual articulation angle between the tractor and the semitrailer under the steering wheel angle of each steering wheel in the dynamic steering test of each working condition, the articulation angle compensation value, and the deformation compensation enabling threshold value are extracted. If the articulation angle compensation value is greater than or equal to the deformation compensation enabling threshold value, it indicates that the actual articulation angle is strongly disturbed by deformation, and the actual articulation angle needs to be corrected. The correction process is to superimpose the actual articulation angle and the articulation angle compensation value in the direction of driving the actual articulation angle to converge to the ideal articulation angle, and the corrected actual articulation angle is taken as the calibration target. On the contrary, if the articulation angle compensation value is less than the deformation compensation enabling threshold value, the actual articulation angle is directly taken as the calibration target.

7. The closed loop test calibration system for the synchronization angle of a van-type semi-trailer according to claim 6, characterized in that: The target saddle position reverse analysis process includes: Based on the mapping reference table, the unit articulation angle deviation-saddle displacement conversion coefficient is analyzed. The conversion coefficient is calibrated through the mapping relationship between the deviation value of the actual articulation angle and the ideal articulation angle in the static steering test and the saddle displacement vector; Based on the conversion coefficient, the calibration target and the deviation value of the actual articulation angle in the mapping reference table under the same steering angle state are converted into an optimized saddle space displacement vector; The original saddle space displacement vector under the same steering angle state in the mapping reference table is extracted, and the optimized vector is superimposed to generate a target displacement vector; The target displacement vector is adjusted based on the preset initial space position of the saddle to determine the target saddle position.

8. The closed loop test calibration system for the synchronization angle of a van-type semi-trailer according to claim 1, characterized in that: The multi-working-condition driving verification data collection process is: After the three-dimensional actuator completes the adjustment of the saddle position, the deviation value of the actual articulation angle and the ideal articulation angle after the adjustment of the saddle is collected and taken as the driving verification data.

9. The closed loop test calibration system for the synchronization angle of a van-type semi-trailer according to claim 8, characterized in that: The iterative update of the deformation mechanics model parameters and the compensation threshold value generation strategy includes: The proportion of the number of working conditions in which the calibrated articulation angle exceeds the permitted range in the multi-working-condition driving verification data is counted. When the proportion is higher than the preset non-secondary working condition ratio, an update instruction is triggered; The pre-trained deformation mechanics model is incrementally trained using the newly added deformation data set until the error of the verification set meets the accuracy requirement; Based on the misjudgment rate, the load-sensitive factors, road-sensitive factors, and motion inhibition factors in the compensation threshold value generation strategy are dynamically adjusted.

Images