Railway and highway combined transport coil steel turnover clamping and conveying system
Patent Information
- Application Number
- CN202611438543.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-16
- Publication Date
- 2026-10-23
Smart Images
Figure CN122962934A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intermodal steel coil conveying technology, specifically to a railway-road intermodal steel coil turning and clamping conveying system. Background Technology
[0002] With the widespread application of multimodal transport in the steel logistics sector, coiled steel, as a bulk commodity, frequently needs to be transshipped between railway flatcars and road trailers. Traditional coiled steel transshipment operations mainly rely on bridge cranes, gantry cranes, or forklifts in conjunction with wire rope slings for manual rigging and hooking. In existing flipping and clamping conveyor equipment, to achieve precise docking with railway flatcars (such as the NX70 model) and road trailers, magnetic scales or Gray line busbars are usually laid along both sides of the running track as position detection sensors. At the same time, two-dimensional LiDAR or visual contour recognition cameras are mounted on the car body to scan the position of the car body blocks. The running positioning control is achieved by redundantly verifying the visual scanning results with the magnetic scale readings.
[0003] Existing technologies lack a way to resolve the conflict between the absolute displacement measurement benchmark and the visual positioning result caused by uneven settlement of the track foundation and the meshing obstruction of the running gear rack due to uneven settlement of the track foundation, and the synchronous deformation of the foundation with the ground, without damaging the existing ballast subgrade, so as to ensure accurate positioning and attitude conversion control during the steel coil flipping, clamping and conveying process.
[0004] The traveling mechanism includes a segmented rack laid along the track foundation, a traveling gear in the vehicle body, a servo motor, a reducer, and a floating preload mechanism. The segmented rack is mounted on the track foundation via adjustable supports, ball joints, and elastic pads. The adjustable supports compensate for rack height changes caused by uneven settlement of the track foundation. The traveling gear uses drum-shaped or modified teeth and is equipped with a tooth backlash adjustment mechanism. The control device drives the servo motor based on the traveling drive output from the dynamic pose field and the settlement baseline field, ensuring stable meshing between the traveling gear and the rack and preventing meshing jamming due to foundation settlement. Conflicts between the absolute displacement measurement reference and the visual positioning results are uniformly processed in the dynamic pose field through multi-source data fusion using covariance adaptive scheduling. Summary of the Invention
[0005] The purpose of this invention is to provide a railway-road intermodal steel coil turning, clamping, and conveying system to solve the problems mentioned above.
[0006] The objective of this invention can be achieved through the following technical solutions:
[0007] The rail-road intermodal steel coil turning and clamping conveyor system includes:
[0008] The baseline construction module collects the visual target coordinates, vehicle inertial attitude, and wheel mileage on the fixed structure, and associates the visual target coordinates, vehicle inertial attitude, and wheel mileage to construct a settlement baseline field.
[0009] The fusion solution module performs multi-source data fusion with covariance adaptive scheduling on the settlement baseline field based on the real-time collected visual target coordinates, inertial attitude and wheel mileage to generate a dynamic pose field;
[0010] The driving solution module, based on the target landing coordinates obtained by laser scanning, inverts and solves the traversal driving amount in the dynamic pose field. The traversal driving amount is superimposed with the displacement pre-adjustment value recursively compensated for the historical settlement sequence.
[0011] The feedback correction module collects the target landing coordinates twice before the steel coil descends, solves the pose feedback amount between the two collected coordinates and the pose corresponding to the travel drive amount, and obtains the final positioning error through multi-source confidence entropy weight filtering.
[0012] The execution feedback module performs travel compensation based on the final positioning deviation, and drives the clamping and flipping execution sequentially according to the final position after compensation, so that the steel coil is clamped, locked and its attitude is changed and then placed on the target transport vehicle. At the same time, the positioning deviation is fed back to the settlement baseline field to complete the baseline update.
[0013] As a further aspect of the present invention: the process of constructing the settlement baseline field is as follows:
[0014] Using the wheel mileage collected at the current moment as the travel index, the inertial attitude acquired at the same time is decomposed into the spatial orientation of the visual target, and the correspondence between the instantaneous pose of the vehicle body and the coordinates of the fixed target is established.
[0015] The correspondences corresponding to multiple consecutive travel indices are accumulated, and the vertical offset at each index point is inverted using the fixed target coordinates as spatial anchor points.
[0016] The vertical offsets are arranged sequentially according to their travel index to form a settlement baseline field distributed along the travel direction.
[0017] As a further aspect of the present invention: the generation of the dynamic pose field specifically includes:
[0018] The real-time collected visual target coordinates, inertial attitude, and wheel mileage are compared with the pre-stored values at the corresponding travel index in the settlement baseline field.
[0019] Based on the relative magnitude of the deviations obtained from each comparison, each data source at the current moment is assigned its own covariance scheduling coefficient;
[0020] When updating the pose at the travel index corresponding to the settlement baseline field, the data sources are introduced sequentially according to the priority established by the scheduling coefficients of each data source for successive correction, thereby generating the dynamic pose field at the current moment.
[0021] As a further aspect of the present invention: assigning each data source its own covariance scheduling coefficient at the current moment specifically includes:
[0022] Take multiple historical poses adjacent to the current travel index in the settlement baseline field, and calculate the historical dispersion amplitude of the visual target coordinates, inertial attitude and wheel mileage at each pose.
[0023] The current deviation of each data source is obtained by subtracting the visual target coordinates, inertial attitude, and wheel mileage collected at the current moment from the pre-stored pose corresponding components at the current travel index in the settlement baseline field.
[0024] The current deviation of each data source is compared with its historical dispersion, and the covariance scheduling coefficient of each data source is determined based on the comparison results. Data sources in the current trip index Covariance scheduling coefficient at the location Determine using the following formula: ; ;
[0025] in, The length of the history index window is set to 20. To prevent the constant from being zero, we take 0.01; For the first The data source in the first The difference between the measured value at each travel index and the pre-stored value of the settlement baseline field; For the first The data source in the first The measured value at the first travel index is the same as the first... The difference between the measured values at each trip index, the larger the deviation relative to the historical dispersion of the data source, the smaller the scheduling coefficient value.
[0026] As a further aspect of the present invention: the process of obtaining the travel drive amount is as follows:
[0027] Extract the vertical settlement values at each historical travel index along the travel direction in the settlement baseline field, and convert the vertical settlement values into equivalent settlement components along the travel direction based on the installation inclination angle of the track or rack at each index point: ;
[0028] in, For the first Vertical settlement values at each index point Pre-adjust the equivalent displacement component in the direction of travel at that point. To calibrate the conversion factors, , For the first The angle of inclination of the track or rack mounting surface relative to the horizontal plane at each index point; then, according to the travel direction... By recursively summing up each item, the cumulative settlement trend value is obtained and used as the displacement pre-adjustment value;
[0029] The target landing coordinates in the dynamic pose field are corrected in reverse by the displacement pre-adjustment value, and the difference between the corrected mapping position and the current vehicle pose in the direction of travel is used as the basic travel.
[0030] Extract the travel drive amount corresponding to the basic stroke, and add the displacement pre-adjustment value to the travel drive amount in the same direction to form the travel drive amount after superimposing the displacement pre-adjustment value.
[0031] As a further aspect of the present invention: the process of obtaining the basic travel is as follows:
[0032] Read the target's landing coordinates in the dynamic pose field and apply the displacement pre-adjustment value in the opposite direction of the travel direction to the corresponding mapping position to obtain the corrected mapping position after settlement compensation.
[0033] Based on the dynamic pose field at the current moment, extract the current coordinates of the vehicle body pose in the direction of travel;
[0034] The difference between the corrected mapping position's coordinates in the travel direction and the current coordinates is used as the base travel distance.
[0035] As a further aspect of the present invention: the screening to obtain the final positioning deviation specifically includes:
[0036] Before the steel coil descends, the target landing coordinates are initially collected, and the first feedback amount is obtained by subtracting the pose corresponding to the initial collected coordinates and the travel drive amount.
[0037] After a preset interval, the target's landing coordinates are collected a second time, and the second feedback amount is obtained by subtracting the pose corresponding to the second collected coordinates from the walking drive amount.
[0038] The first and second feedback quantities are compared with the historical feedback quantity distributions at the corresponding travel indexes in the settlement baseline field to determine their degree of difference. The confidence level of each feedback quantity is determined based on the degree of difference. The contribution of each feedback quantity to the final positioning deviation is then distributed in reverse order of confidence level, and the final positioning deviation is obtained by combining the results.
[0039] As a further aspect of the present invention: the step of placing the steel coil onto the target transport vehicle after clamping, locking, and attitude conversion specifically includes:
[0040] The current travel drive is corrected in reverse by the final positioning deviation, and travel compensation is performed according to the corrected travel drive to make the vehicle reach the final position after compensation.
[0041] The final pose is decomposed into lateral position components, longitudinal position components, and heading angle components;
[0042] The clamping and locking are executed when the lateral and longitudinal position components reach preset thresholds.
[0043] After clamping and locking are completed, the target rotation angle around the horizontal rotation axis is obtained from the preset attitude conversion table according to the target attitude of the steel coil, and the rotation platform of the rotation mechanism is aligned according to the heading angle component, and the rotation actuator is driven to rotate around the horizontal rotation axis until the target rotation angle is reached.
[0044] After placement, the deviation between the measured placement coordinates and the target placement coordinates is taken as the placement deviation and backfilled into the corresponding travel index in the settlement baseline field to complete the baseline update.
[0045] As a further aspect of the present invention: the step of correcting the current travel drive amount in reverse with the final positioning deviation specifically includes:
[0046] The final positioning deviation is decomposed into a travel deviation component and a lateral deviation component along the travel direction. The travel deviation component is then superimposed onto the current travel drive amount to obtain the initial travel drive amount.
[0047] The lateral deviation component is added in the same direction to the lateral drive portion of the initial repair travel drive to obtain the repair travel drive.
[0048] Longitudinal travel compensation is performed using the travel drive portion of the repair travel drive quantity. After the longitudinal travel compensation is in place, lateral travel compensation is performed using the lateral movement drive portion of the repair travel drive quantity, so that the vehicle body reaches the final position after compensation.
[0049] The beneficial effects of this invention are:
[0050] (1) By constructing a settlement baseline field and using multi-source data fusion with covariance adaptive scheduling to generate a dynamic pose field, the travel positioning can still obtain accurate real-time pose information when uneven settlement occurs in the ballast foundation. At the same time, by collecting the target landing coordinates twice and combining them with the historical feedback distribution for confidence entropy weight screening, the problem of positioning jump caused by interference factors such as vibration and obstruction in a single measurement is solved, thereby improving the positioning reliability and stability when the transport vehicle docks with the transport vehicle.
[0051] (2) The displacement pre-adjustment value of the historical settlement sequence is superimposed on the travel drive amount, and the positioning deviation is fed back to the settlement baseline field after the steel coil is placed to complete the baseline update, so that the settlement baseline field gradually approaches the real track settlement distribution state with the operation cycle; at the same time, the clamping locking and flipping are executed sequentially according to the final position after compensation, to ensure that the clamping point is accurately aligned during the attitude conversion of the steel coil, to avoid collision or scratching between the steel coil and the car block due to positioning deviation, and to protect the integrity of the steel coil surface. Attached Figure Description
[0052] The invention will now be further described with reference to the accompanying drawings.
[0053] Figure 1 This is a system block diagram of the present invention;
[0054] Figure 2 This is a flowchart of the process of constructing the settlement baseline field in this invention;
[0055] Figure 3 This is a flowchart of the process for generating a dynamic pose field in this invention. Detailed Implementation
[0056] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0057] Please see Figure 1 As shown, this invention is a railway-road intermodal steel coil turning, clamping, and conveying system, comprising:
[0058] The baseline construction module collects the visual target coordinates, vehicle inertial attitude, and wheel mileage on the fixed structure, and associates the visual target coordinates, vehicle inertial attitude, and wheel mileage to construct a settlement baseline field.
[0059] The fusion solution module performs multi-source data fusion with covariance adaptive scheduling on the settlement baseline field based on the real-time collected visual target coordinates, inertial attitude and wheel mileage to generate a dynamic pose field;
[0060] The driving solution module, based on the target landing coordinates obtained by laser scanning, inverts and solves the traversal driving amount in the dynamic pose field. The traversal driving amount is superimposed with the displacement pre-adjustment value recursively compensated for the historical settlement sequence.
[0061] The feedback correction module collects the target landing coordinates twice before the steel coil descends, solves the pose feedback amount between the two collected coordinates and the pose corresponding to the travel drive amount, and obtains the final positioning error through multi-source confidence entropy weight filtering.
[0062] The execution feedback module performs travel compensation based on the final positioning deviation, and drives the clamping and flipping execution sequentially according to the final position after compensation, so that the steel coil is clamped, locked and its attitude is changed and then placed on the target transport vehicle. At the same time, the positioning deviation is fed back to the settlement baseline field to complete the baseline update.
[0063] Please see Figure 1 As shown, in the baseline construction module, the visual target coordinates, vehicle inertial attitude, and wheel mileage on the fixed structure are collected, and the visual target coordinates, vehicle inertial attitude, and wheel mileage are correlated to construct a settlement baseline field, specifically including:
[0064] When collecting the coordinates of visual targets on fixed structures, visual targets are pre-embedded on the surface of fixed structures on both sides of the railway freight yard. The spatial coordinates of each target are pre-calibrated by a total station and stored in the control device. During the movement of the flipping clamping conveyor, an industrial camera installed on the vehicle body collects target images at a sampling frequency of ten times per second. The horizontal and vertical coordinates of the vehicle body relative to each target are obtained through image processing and calculation as the visual target coordinates.
[0065] When acquiring the vehicle's inertial attitude, the inertial measurement unit (IMU), installed at the vehicle's center of gravity, outputs the angular velocity values of the vehicle around its three spatial axes at a sampling frequency of 100 times per second. The angular velocity values output at each moment are successively accumulated according to the sampling sequence to convert them into the vehicle's pitch, roll, and yaw angles at that moment. These three values together constitute the vehicle's inertial attitude. The IMU performs initial alignment when the vehicle is stationary: the initial pitch and roll angles are determined using the gravity vector output from the accelerometer, and the initial yaw angle is determined using the visual target coordinates or a magnetometer. During the stationary period, gyro bias is estimated and zero-bias compensation is applied to the angular velocity output. During operation, the angular velocity values are output at a sampling frequency of 100 times per second and successively accumulated to obtain the attitude angles. Simultaneously, using the visual target coordinates and wheel mileage as external observations, the pitch, roll, and yaw angles are periodically corrected using complementary filtering or Kalman filtering. Zero-speed correction is performed during docking or zero-speed periods to suppress integral drift and ensure the long-term accuracy of the settlement baseline field.
[0066] When collecting wheel travel distance, the incremental encoder installed on the output shaft of the running drive motor outputs a pulse sequence as the motor rotates. The control device accumulates the number of pulses and multiplies the accumulated pulse count by the equivalent travel value corresponding to each revolution of the running wheel to obtain the wheel travel distance value calculated from the starting position.
[0067] When constructing the settlement baseline field, the wheel mileage collected at the current moment is first used as the travel index. The numerical resolution of this travel index is 0.1 meters, that is, every 0.1 meters traveled corresponds to an index point. The pitch angle value in the inertial attitude collected at the same moment is decomposed to the spatial orientation of the visual target. The decomposition method is to geometrically superimpose the pitch angle value with the azimuth angle value of the current visual target relative to the vehicle body to obtain the correspondence between the instantaneous pose of the vehicle body and the coordinates of the fixed target.
[0068] The correspondences corresponding to multiple consecutive travel indexes are accumulated, and the accumulated number is the travel length range covered by one hundred index points. The vertical offset at each index point is inverted using the fixed target coordinates as spatial anchor points. The inversion method is to take the spatial vertical projection value of the instantaneous vehicle pose at each index point relative to the fixed target coordinates, and use the difference between the projection value and the preset standard vertical projection value as the vertical offset at the index point.
[0069] The vertical offsets obtained at each index point are arranged in ascending order of their travel index values to form a settlement baseline field evenly distributed at 0.1-meter intervals along the travel direction. The initial value of the vertical offset at each index point in the settlement baseline field is set to zero, and the vertical offset at the corresponding index point is replaced and updated according to the actual placement deviation after each travel operation.
[0070] Please see Figure 2 As shown, in the fusion solution module, based on the real-time acquired visual target coordinates, inertial attitude, and wheel mileage, multi-source data fusion with covariance adaptive scheduling is performed on the settlement baseline field to generate a dynamic pose field, specifically including:
[0071] Using the wheel mileage collected at the current moment as the travel index, the inertial attitude acquired at the same time is decomposed into the spatial orientation of the visual target, and the correspondence between the instantaneous pose of the vehicle body and the coordinates of the fixed target is established.
[0072] The correspondences corresponding to multiple consecutive travel indices are accumulated, and the vertical offset at each index point is inverted using the fixed target coordinates as spatial anchor points.
[0073] The vertical offsets are arranged sequentially according to their travel index to form a settlement baseline field distributed along the travel direction.
[0074] When comparing the real-time collected visual target coordinates, inertial attitude, and wheel mileage with the pre-stored values at the corresponding travel indexes in the settlement baseline field, the wheel mileage value collected at the current moment is used as the search condition to find the travel index in the settlement baseline field that is equal to the wheel mileage value. Then, the pre-stored visual target coordinate values, inertial attitude values, and wheel mileage values at the travel indexes are extracted respectively. The difference between the values of each data source collected at the current moment and the corresponding pre-stored values is calculated one by one to obtain the visual target coordinate deviation, inertial attitude deviation, and wheel mileage deviation.
[0075] When assigning covariance scheduling coefficients to each data source at the current moment, the top twenty historical poses adjacent to the current travel index in the settlement baseline field are taken, and the historical dispersion amplitude of the visual target coordinates, inertial attitude, and wheel mileage at each pose is calculated. The historical dispersion amplitude of a certain data source is calculated in the following way:
[0076] The measured values of the data source at twenty adjacent historical poses are taken, and the difference between the maximum and minimum values is taken as the historical dispersion amplitude of the data source. The visual target coordinates, inertial attitude, and wheel mileage collected at the current moment are subtracted from the corresponding components of the pre-stored pose at the current travel index in the settlement baseline field to obtain the current deviation amplitude of each data source, specifically the current deviation amplitude of the visual target coordinates, the current deviation amplitude of the inertial attitude, and the current deviation amplitude of the wheel mileage. The current deviation amplitude of each data source is compared with its respective historical dispersion amplitude, and the covariance scheduling coefficient of each data source is determined based on the comparison results. The scheduling coefficient of the data source with a larger deviation amplitude relative to the historical dispersion amplitude is smaller.
[0077] In the specific calculation method for assigning covariance scheduling coefficients to each data source at the current moment, the first... Data sources in the current trip index Covariance scheduling coefficient at the location Determine using the following formula: ; ;
[0078] in, The length of the history index window is set to 20. To prevent the constant from being zero, we take 0.01; For the first The data source in the first The difference between the measured value at each travel index and the pre-stored value of the settlement baseline field; For the first The data source in the first The measured value at the first travel index is the same as the first... The difference between the measured values at each trip index
[0079] When the data sources are introduced sequentially for correction according to the priority established by the scheduling coefficients of each data source, the scheduling coefficients of visual target coordinates, inertial attitude, and wheel mileage are arranged in descending order of value. The data source with the largest scheduling coefficient is introduced for correction first, and the data source with the smallest scheduling coefficient is introduced for correction last.
[0080] When updating the pose at the travel index corresponding to the settlement baseline field, the measured values of each data source at the current moment are introduced into the correction in the order mentioned above. The correction process is to move the pose component stored at the current travel index toward the measured value of the data source, and the magnitude of the movement is controlled by the scheduling coefficient of the data source.
[0081] When generating the dynamic pose field at the current moment, the corrected pose values obtained after all data sources have completed a round of correction in priority order are the dynamic pose field at the current moment.
[0082] The range of values for the covariance scheduling coefficients is set as follows: arrive Between, when the ratio of the current deviation of a certain data source to its historical dispersion is less than At that time, the scheduling coefficient of the data source is taken as This means the data source is fully trusted for correction; when the ratio is... to During this period, the scheduling coefficient is used according to the value calculated by the above formula, and this value is within... to Between; when the ratio is greater than and less than or equal to At that time, the scheduling coefficient of the data source is taken as When the ratio is greater than At that time, the scheduling coefficient of the data source is taken as .
[0083] The criteria are as follows: when the deviation is less than half of the historical dispersion, it indicates that the data source is stable and reliable and does not require weight reduction; when the deviation exceeds half of the historical dispersion, it indicates that interference is beginning to appear and weight reduction is necessary; when the deviation exceeds the historical dispersion... A value of 10 times indicates that the data source has deviated significantly from the normal range and retains only the minimum level of reference value.
[0084] In the specific process of sequentially introducing each data source into the priority order established by the scheduling coefficient of each data source for successive correction, the first correction uses the pre-stored pose at the current travel index in the settlement baseline field as the correction starting value. The measured value of the first data source in the priority order is used to correct the starting value. The correction method is to move the corresponding component of the starting value in the direction of the measured value. The amount of movement is the scheduling coefficient of the data source multiplied by the current deviation of the data source to obtain the first correction pose.
[0085] During the second correction, the first correction pose is used as the corrected value. The measured value of the second data source in the priority order is used to correct the corrected value. The correction method is to move the corresponding component of the corrected value towards the measured value of the second data source. The amount of movement is the scheduling coefficient of the second data source multiplied by the current deviation of the second data source to obtain the second correction pose.
[0086] In the third correction, the second correction pose is used as the corrected value. The measured value of the third data source in the priority order is used to correct the corrected value. The correction method is to move the corresponding component of the corrected value in the direction of the measured value of the third data source. The amount of movement is the scheduling coefficient of the third data source multiplied by the current deviation of the third data source to obtain the third correction pose. The third correction pose is output as the final correction pose and marked as the dynamic pose field at the current time.
[0087] Please see Figure 3 As shown, in the drive solution module, based on the target landing coordinates obtained by laser scanning, the traversal drive is inverted and solved in the dynamic pose field. The traversal drive is superimposed with displacement pre-adjustment values recursively compensated for historical settlement sequences, specifically including:
[0088] The inversion solution process of the travel drive uses the dynamic pose field at the current moment as the reference system. This dynamic pose field contains the estimated values of the vehicle body pose at each travel index point along the travel direction. The target landing coordinates are obtained by scanning the position of the block inside the carriage of the target transport vehicle by a laser scanning device installed on the flipping clamping conveyor. The target landing coordinates indicate the lateral and longitudinal positions where the steel coil should be placed on the carriage floor.
[0089] When extracting the vertical settlement values at each historical travel index arranged along the travel direction in the settlement baseline field, the settlement baseline field stores the vertical offset value at each index point from the initial travel index to the current travel index. This vertical offset represents the settlement height of the track foundation at that index point relative to the initial state. The vertical settlement values at each index point are read sequentially from the initial index along the travel direction, and the read vertical settlement values are recursively accumulated item by item. The meaning of item-by-item recursion and accumulation is as follows: at the initial index, the vertical settlement value is directly used as the cumulative value at that point. At each subsequent index point, the vertical settlement value at that point is added to the cumulative value of the previous index point to obtain the cumulative settlement trend value at that point. As the travel distance increases, the cumulative settlement trend value reflects the total algebraic sum of all settlement from the initial position to the current position. The cumulative settlement trend value at the current travel index is used as the displacement pre-adjustment value. This displacement pre-adjustment value has a directional attribute, and its positive or negative sign corresponds to the vertical settlement direction.
[0090] When using the displacement pre-adjustment value to correct the target landing coordinates in the dynamic pose field, first find the corresponding mapping position in the dynamic pose field. This mapping position represents the spatial point where the steel coil should be placed under ideal conditions without considering track settlement. Apply the displacement pre-adjustment value in the reverse direction of travel to this mapping position. The application method is as follows: if the displacement pre-adjustment value is positive, indicating track subsidence, move the mapping position in the reverse direction of travel by a distance equal to the displacement pre-adjustment value; if the displacement pre-adjustment value is negative, indicating track elevation, move the mapping position in the forward direction of travel by a distance equal to the absolute value of the displacement pre-adjustment value. The spatial point obtained after the movement is the corrected mapping position after settlement compensation.
[0091] When extracting the current coordinates of the vehicle body in the travel direction based on the current dynamic pose field, the dynamic pose field contains the pose estimate of the vehicle body at the point corresponding to the travel index at the current moment. The coordinate value of the geometric center of the vehicle body in the travel direction is extracted from the pose estimate as the current coordinate. The current coordinate is the actual distance the vehicle body moves relative to the starting position along the travel direction.
[0092] When subtracting the coordinates of the corrected mapping position in the travel direction from the current coordinates, the difference between the current coordinates and the travel direction coordinates of the corrected mapping position is the base travel. This base travel represents the net travel distance required to move along the travel direction from the current vehicle body position to the target landing position after settlement compensation. This net travel distance has eliminated the offset effect of track settlement on the target position.
[0093] When extracting the travel drive quantity corresponding to the basic stroke, the travel drive quantity is the speed setpoint and position setpoint output by the control device to the travel drive motor. The speed setpoint and position setpoint are directly proportional to the value of the basic stroke. The proportionality coefficient is calculated based on the rated speed of the travel drive motor and the diameter of the travel wheel. The conversion method is to divide the basic stroke by the effective circumference of the travel wheel and then multiply it by the rated speed of the drive motor. The resulting value is the speed setpoint in the travel drive quantity. At the same time, the value of the basic stroke is directly used as the position setpoint in the travel drive quantity.
[0094] When the displacement pre-adjustment value is added to the travel drive in the same direction, the same direction superposition means judging whether the sign of the displacement pre-adjustment value is consistent with the sign of the basic travel. If the two signs are consistent, the absolute value of the displacement pre-adjustment value is added to the position setpoint in the travel drive to obtain the new position setpoint. If the two signs are opposite, the absolute value of the displacement pre-adjustment value is subtracted from the position setpoint in the travel drive to obtain the new position setpoint. At the same time, the value obtained by dividing the displacement pre-adjustment value by the effective circumference of the travel wheel and multiplying it by the rated speed of the drive motor is added or subtracted in the same direction with the speed setpoint in the travel drive to form the travel drive after superimposing the displacement pre-adjustment value. The superimposed travel drive includes both the motion requirements corresponding to the basic travel and the additional motion requirements required for track settlement compensation.
[0095] During the acquisition of the basic travel, when reading the target landing coordinates in the dynamic pose field, the target landing coordinates obtained by the laser scanning device are used as input parameters. Coordinate transformation operation is performed in the dynamic pose field. This coordinate transformation operation transforms the target landing coordinates from the measurement coordinate system of the laser scanning device to the global coordinate system of the dynamic pose field. The global coordinates obtained by the transformation are the mapping positions.
[0096] When the displacement pre-adjustment value is applied to the mapped position in the opposite direction of the travel direction, the travel direction coordinate value of the mapped position is subtracted from the displacement pre-adjustment value. If the displacement pre-adjustment value is positive, the travel direction coordinate value of the mapped position decreases; if the displacement pre-adjustment value is negative, the travel direction coordinate value of the mapped position increases. The subtraction yields a new travel direction coordinate value, which, together with the lateral coordinate value of the original mapped position, constitutes the corrected mapped position.
[0097] When subtracting the coordinates of the corrected mapping position in the travel direction from the current coordinates, the difference between the coordinates of the corrected mapping position in the travel direction and the current coordinates is taken as the base travel. The value of the base travel can be positive or negative. A positive value indicates that the vehicle needs to move in the positive direction of the travel direction, and a negative value indicates that the vehicle needs to move in the opposite direction of the travel direction.
[0098] The starting point for the accumulation of displacement pre-adjustment values during the accumulation process is set as the starting position of track settlement monitoring. The vertical settlement value at this starting position is preset to zero. The ending point for accumulation is set as the position corresponding to the current travel index. The accumulation range covers the entire travel interval from the starting position to the current position. The accumulation method is linear accumulation, that is, the settlement values of each point are directly added without attenuation processing. The basis for this setting is that the settlement deformation of the track foundation has irreversible cumulative characteristics. The settlement generated at each point in the early stage contributes to the current cumulative settlement trend and should not be attenuated or forgotten over time. Therefore, the equal weighted accumulation method of the entire process from the starting position to the current position is adopted.
[0099] When performing reverse correction, the reverse application coefficient of the displacement pre-adjustment value is set to 1.0, i.e., completely reverse application. The basis for this setting is that the sign and magnitude of the displacement pre-adjustment value accurately reflect the actual deformation amount and direction of the track foundation along the vertical direction. The track foundation is equipped with racks along the running direction, and there is an installation inclination angle between the rack mounting surface and the horizontal plane. The first settlement baseline field The vertical offset at each index point is Because uneven settlement of the track foundation will cause the track and rack to tilt, and the wheel travel distance accumulates along the track surface, while the target landing coordinates are defined under the design horizontal datum, there is a projection conversion relationship between the two: ;
[0100] in, For the first The travel direction conversion coefficient at each index point is obtained through on-site calibration using a total station or laser tracker. The displacement preset value is: ;
[0101] The target's landing coordinates in the dynamic pose field are corrected in reverse using the displacement pre-adjustment value: ;
[0102] Therefore, the vertical settlement is not directly and equally converted into horizontal displacement, but is geometrically converted through the installation angle and calibration coefficient of the track or rack.
[0103] When the displacement pre-adjustment value is superimposed in the same direction, the superposition coefficient is set to 1.0, which means that it is superimposed in the same direction. The basis for this setting is that the position setpoint in the travel drive only represents the net travel required to reach the corrected mapping position from the current position. This net travel already includes the settlement compensation correction amount. The displacement pre-adjustment value is superimposed in the travel drive to compensate for the mechanical travel loss caused by settlement. This loss amount is equal to the settlement amount, so the superposition coefficient of 1.0 can achieve complete compensation.
[0104] In the feedback correction module, the target landing coordinates are acquired twice before the steel coil descends. The pose feedback amount between the two acquired coordinates and the pose corresponding to the travel drive is calculated. The final positioning error is obtained by filtering through multi-source confidence entropy weights, specifically including:
[0105] Before the steel coil descends, when the target landing coordinates are initially collected, the tilting and clamping conveyor stops moving after reaching the preset steel coil landing position. The laser scanning device installed on the vehicle body performs the first scan of the landing area inside the target transport vehicle. The scanning time is recorded as the initial collection time. The acquired target landing coordinates include the lateral coordinate value and the longitudinal coordinate value of the landing point. The initial collected coordinates are compared with the vehicle body pose corresponding to the driving amount in the same coordinate system. The lateral coordinate value of the initial collected coordinates is subtracted from the lateral coordinate value of the vehicle body pose to obtain the first lateral difference value. The longitudinal coordinate value of the initial collected coordinates is subtracted from the longitudinal coordinate value of the vehicle body pose to obtain the first longitudinal difference value. The first lateral difference value and the first longitudinal difference value together constitute the first feedback amount.
[0106] When acquiring the target's location coordinates for the second time after a preset interval, the timer starts from the initial acquisition time and waits for the preset time interval to arrive before performing the second scan. This time interval is set to 0.5 seconds based on the single scan cycle of the laser scanning device and the data processing time. The acquired coordinates are compared with the vehicle body pose corresponding to the same driving quantity. The lateral coordinate value of the acquired coordinates is subtracted from the lateral coordinate value of the vehicle body pose to obtain the second lateral difference. The longitudinal coordinate value of the acquired coordinates is subtracted from the longitudinal coordinate value of the vehicle body pose to obtain the second longitudinal difference. The second lateral difference and the second longitudinal difference together constitute the second feedback quantity.
[0107] When comparing the first and second feedback quantities with the historical feedback quantity distributions at the corresponding travel indexes in the settlement baseline field, the historical feedback quantity values recorded in the most recent work cycles stored at the current travel index are extracted from the settlement baseline field using the current travel index as the retrieval condition. Each historical feedback quantity includes historical lateral differences and historical longitudinal differences. The average and standard deviation of all extracted historical lateral differences and historical longitudinal differences are calculated. The difference of the first feedback quantity is obtained as follows: the average of the first lateral difference in the first feedback quantity and the historical lateral difference at the travel index are compared. The mean values are subtracted and the absolute value is taken. This absolute value is then divided by the standard deviation of the historical lateral differences to obtain the lateral difference ratio. The first longitudinal difference in the first retransmission is subtracted from the average of the historical longitudinal differences at the travel index and the absolute value is taken. This absolute value is then divided by the standard deviation of the historical longitudinal differences to obtain the longitudinal difference ratio. The larger of the lateral difference ratio and the longitudinal difference ratio is taken as the difference degree of the first retransmission. The difference degree of the second retransmission is obtained in the same way, except that the second lateral difference and the second longitudinal difference in the second retransmission are calculated in the same way with the corresponding historical average and standard deviation, and the larger value is taken as the difference degree of the second retransmission.
[0108] When determining the confidence level of each feedback quantity based on its degree of difference, the smaller the degree of difference, the closer the feedback quantity is to the historical stable distribution, and the higher its reliability, thus the higher the confidence level. The confidence level is determined as follows: when the degree of difference of a feedback quantity is less than 0.5, the confidence level of the feedback quantity is 1.0; when the degree of difference is between 0.5 and 1.5, the confidence level is calculated by subtracting the degree of difference from 1.5 and then dividing by 1.0, that is, the confidence level decreases linearly from 1.0 to 0.0; when the degree of difference is greater than 1.5, the confidence level is 0.1. The basis for this setting is that a degree of difference exceeding 1.5 indicates that the feedback quantity has deviated significantly from the historical distribution and retains only the minimum reference value.
[0109] When assigning the contribution of each feedback quantity to the final positioning deviation in reverse order based on confidence level, the confidence levels of the first and second feedback quantities are first arranged in descending order of value. The feedback quantity with higher confidence level is assigned a larger contribution, and the feedback quantity with lower confidence level is assigned a smaller contribution. The contribution is assigned by summing the two confidence levels to obtain the total confidence level, dividing the confidence level of the first feedback quantity by the total confidence level to obtain the first contribution coefficient, and dividing the confidence level of the second feedback quantity by the total confidence level to obtain the second contribution coefficient. The sum of the two contribution coefficients equals 1.
[0110] When calculating the final positioning deviation, the first contribution value is obtained by multiplying the first contribution coefficient by the first lateral difference value in the first feedback quantity, and the first longitudinal difference value in the first feedback quantity is obtained by multiplying the first contribution coefficient by the first longitudinal difference value in the first feedback quantity; the second contribution value is obtained by multiplying the second contribution coefficient by the second lateral difference value in the second feedback quantity, and the second longitudinal difference value in the second feedback quantity is obtained by multiplying the second contribution coefficient by the second longitudinal difference value in the second feedback quantity; the first lateral contribution value and the second lateral contribution value are added together to obtain the final lateral positioning deviation, and the first longitudinal contribution value and the second longitudinal contribution value are added together to obtain the final longitudinal positioning deviation. The final lateral positioning deviation and the final longitudinal positioning deviation together constitute the final positioning deviation.
[0111] The extraction range for historical feedback distribution is set to the feedback values recorded in the most recent 50 historical work cycles at the current travel index. This setting is based on the fact that a sample size of 50 can provide statistically stable estimates of the mean and standard deviation, while avoiding the dilution effect of outdated data on the representativeness of the current working conditions. The standard deviation in the variance calculation adopts an unbiased estimation method, that is, each historical difference is subtracted from the mean, squared, and then summed. The summation is divided by the sample size minus one to obtain the variance, and the square root of the variance is the standard deviation.
[0112] The confidence thresholds of 0.5 and 1.5 are set as follows: a difference of less than 0.5 indicates that the feedback falls within one standard deviation of the historical distribution and belongs to the high confidence interval; a difference of more than 1.5 indicates that the feedback falls outside three standard deviations of the historical distribution and belongs to the low confidence interval and should be rejected; when the difference is between the two, the confidence level is allocated in a linear decreasing manner to maintain a smooth transition.
[0113] In the execution feedback module, travel compensation is performed based on the final positioning deviation, and the clamping and flipping executions are driven sequentially according to the compensated final pose, so that the steel coil is clamped, locked, and its attitude is changed before being placed on the target transport vehicle. At the same time, the positioning deviation is fed back to the settlement baseline field to complete the baseline update, specifically including:
[0114] When correcting the current travel drive amount in reverse with the final positioning deviation, the final positioning deviation is orthogonally decomposed in the horizontal plane along the travel direction and perpendicular to the travel direction to obtain the travel deviation component and the lateral deviation component. The travel deviation component represents the displacement of the vehicle body from the target position along the travel direction, and the lateral deviation component represents the displacement of the vehicle body from the target position in the lateral direction perpendicular to the travel direction.
[0115] When the travel deviation component is superimposed onto the current travel drive, the sign of the travel deviation component is determined. If the travel deviation component is positive, it means that the vehicle is currently in front of the target position. In this case, the absolute value of the travel deviation component is subtracted from the longitudinal travel setpoint in the current travel drive. If the travel deviation component is negative, it means that the vehicle is currently behind the target position. In this case, the absolute value of the travel deviation component is added to the longitudinal travel setpoint in the current travel drive. After the above calculation, the initial travel drive is obtained.
[0116] When the lateral deviation component is added to the lateral drive portion of the initial repair travel drive in the same direction, the sign of the lateral deviation component is determined. If the lateral deviation component is positive, it means that the vehicle body needs to move in the positive lateral direction. In this case, a value equal to the absolute value of the lateral deviation component is added to the lateral setpoint in the initial repair travel drive. If the lateral deviation component is negative, it means that the vehicle body needs to move in the negative lateral direction. In this case, a value equal to the absolute value of the lateral deviation component is subtracted from the lateral setpoint in the initial repair travel drive. After the above calculation, the repair travel drive is obtained.
[0117] When longitudinal travel compensation is performed using the travel drive portion of the repair travel drive quantity, the control device outputs the longitudinal travel setpoint of the repair travel drive quantity to the speed controller of the travel drive motor, driving the vehicle body to move along the travel direction. After the longitudinal travel compensation is in place, lateral travel compensation is performed using the lateral travel drive portion of the repair travel drive quantity. The control device outputs the lateral travel setpoint of the repair travel drive quantity to the speed controller of the lateral travel drive motor, driving the vehicle body to move along the lateral travel direction, so that the vehicle body reaches the final position after compensation.
[0118] When the final pose is decomposed into lateral position components and longitudinal attitude components, the coordinate values of the vehicle body in the traveling direction are extracted from the compensated final pose as the longitudinal position component, and the rotation angle value of the vehicle body around the vertical axis is extracted as the heading angle component.
[0119] The clamping and locking condition is that both the lateral position component and the longitudinal position component reach a preset threshold. The preset threshold is triggered when the absolute value of the deviation between the lateral position component and the target lateral position is less than or equal to 0.5 mm. The setting is based on the fact that this deviation range can ensure that the clamping jaws apply a uniform clamping force to both sides of the steel coil, and avoid the steel coil from lateral sliding or local extrusion deformation during the clamping process due to lateral deviation.
[0120] After clamping and locking are completed, the target rotation angle around the horizontal rotation axis is obtained from the preset attitude conversion table according to the target posture of the steel coil. For example, the target rotation angle is 90° when rotating from vertical to horizontal. The rotation mechanism is installed on the slewing platform. First, the slewing platform is adjusted according to the yaw angle component so that the rotation axis is longitudinally aligned with the target transport vehicle. Then, the rotation actuator is driven to rotate around the horizontal rotation axis to the target rotation angle.
[0121] After placement, the deviation between the measured placement coordinates and the target placement coordinates is taken as the current placement deviation and backfilled into the corresponding travel index in the settlement baseline field to complete the baseline update. Specifically, after placement, a laser scanning device is used to measure the actual position of the placed steel coil to obtain the measured placement coordinates. The target placement coordinates are subtracted from the measured placement coordinates to obtain the lateral and longitudinal placement deviation values. The lateral and longitudinal deviation values are combined to form the current placement deviation. Using the current travel index as the search condition, the corresponding storage location is found in the settlement baseline field. After placement, a laser scanning device is used to measure the actual position of the placed steel coil to obtain the measured placement coordinates; the target placement coordinates are subtracted from the measured placement coordinates to obtain the lateral and longitudinal placement deviation values. The longitudinal placement deviation value is converted into a vertical settlement correction amount using an inverse conversion relationship. ;
[0122] in, This represents the longitudinal placement deviation value. For the first The conversion coefficients at each index point are then used. The vertical offset at that index point is then updated using a recursive filtering method. ;
[0123] in, To update the gain, we set it to 0.2. The lateral positioning deviation value is stored as an auxiliary parameter at the same index.
[0124] The reason for performing longitudinal and lateral travel compensation in sequence is as follows: the travel direction is the main direction of movement of the vehicle body, with a large moment of inertia and a long stroke. Therefore, longitudinal travel compensation should be completed first to avoid motion interference caused by the longitudinal position not being in place during lateral compensation. Lateral travel compensation has a shorter stroke and a smaller moment of inertia. Performing lateral compensation after the longitudinal position is locked can ensure that the positioning accuracy of the final pose is not affected by the residual error of longitudinal movement.
[0125] The lateral position component preset threshold of 0.5 mm is set based on the following: the allowable deviation range of the clamping jaws is determined by the outer diameter tolerance of the steel coil and the matching requirements of the jaw contact surface curvature. The threshold of 0.5 mm can ensure that the contact pressure difference between the two sides of the jaws and the outer surface of the steel coil is less than five percent of the interlayer friction of the steel coil, thereby avoiding the misalignment of the strip layers during clamping.
[0126] The longitudinal attitude component is directly used as the basis for the target angle position of the flipping mechanism: the angle control of the flipping actuator is based on the feedback of the absolute encoder as the closed-loop control basis. The longitudinal attitude component value is directly used as the angle setpoint to eliminate the accumulated error of the intermediate conversion link and ensure that the flipping angle control accuracy meets the clearance fit requirements between the steel coil and the bottom plate of the transport vehicle when the attitude is changed.
[0127] The basis for replacing the vertical offset with the longitudinal offset value during the backfilling process is that the longitudinal offset directly reflects the longitudinal positioning error of the steel coil caused by the deformation of the track foundation at the index of the stroke. This error is numerically related to the track settlement height. Therefore, replacing the original vertical offset with the longitudinal offset value can make the settlement baseline field gradually approach the real track settlement distribution state.
[0128] The working principle of this invention is as follows: A settlement baseline field is constructed by collecting the visual target coordinates on a fixed structure, the vehicle's inertial attitude, and wheel mileage. Based on the real-time collected visual target coordinates, inertial attitude, and wheel mileage, multi-source data fusion with covariance adaptive scheduling is performed on the settlement baseline field to generate a dynamic pose field. Based on the target landing coordinates obtained by laser scanning, the travel drive is inverted and solved in the dynamic pose field. The travel drive is superimposed with the displacement pre-adjustment value recursively compensated for the historical settlement sequence. Before the steel coil descends, the target landing coordinates are collected twice, and the pose feedback between the two collected coordinates and the pose corresponding to the travel drive is solved. The final positioning deviation is obtained by multi-source confidence entropy weighting. Travel compensation is performed based on the final positioning deviation, and the clamping, locking, and flipping are sequentially driven according to the compensated final pose, so that the steel coil lands on the target transport vehicle after completing the attitude transformation. At the same time, the landing deviation is fed back to the settlement baseline field to complete the baseline update.
[0129] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.
Claims
1. A railway-road intermodal steel coil turning, clamping, and conveying system, characterized in that, include: The baseline construction module collects the visual target coordinates, vehicle inertial attitude, and wheel mileage on the fixed structure, and associates the visual target coordinates, vehicle inertial attitude, and wheel mileage to construct a settlement baseline field. The fusion solution module performs multi-source data fusion with covariance adaptive scheduling on the settlement baseline field based on the real-time collected visual target coordinates, inertial attitude and wheel mileage to generate a dynamic pose field; The driving solution module, based on the target landing coordinates obtained by laser scanning, inverts and solves the traversal driving amount in the dynamic pose field. The traversal driving amount is superimposed with the displacement pre-adjustment value recursively compensated for the historical settlement sequence. The feedback correction module collects the target landing coordinates twice before the steel coil descends, solves the pose feedback amount between the two collected coordinates and the pose corresponding to the travel drive amount, and obtains the final positioning error through multi-source confidence entropy weight filtering. The execution feedback module performs travel compensation based on the final positioning deviation, and drives the clamping and flipping execution sequentially according to the final position after compensation, so that the steel coil is clamped, locked and its attitude is changed and then placed on the target transport vehicle. At the same time, the positioning deviation is fed back to the settlement baseline field to complete the baseline update.
2. The railway-road intermodal steel coil turning and clamping conveying system according to claim 1, characterized in that, The process of constructing the settlement baseline field is as follows: Using the wheel mileage collected at the current moment as the travel index, the inertial attitude acquired at the same time is decomposed into the spatial orientation of the visual target, and the correspondence between the instantaneous pose of the vehicle body and the coordinates of the fixed target is established. The correspondences corresponding to multiple consecutive travel indices are accumulated, and the vertical offset at each index point is inverted using the fixed target coordinates as spatial anchor points. The vertical offsets are arranged sequentially according to their travel index to form a settlement baseline field distributed along the travel direction.
3. The railway-road intermodal steel coil turning and clamping conveying system according to claim 1, characterized in that, The generation of the dynamic pose field specifically includes: The real-time collected visual target coordinates, inertial attitude, and wheel mileage are compared with the pre-stored values at the corresponding travel index in the settlement baseline field. Based on the relative magnitude of the deviations obtained from each comparison, each data source at the current moment is assigned its own covariance scheduling coefficient; When updating the pose at the travel index corresponding to the settlement baseline field, the data sources are introduced sequentially according to the priority established by the scheduling coefficients of each data source for successive correction, thereby generating the dynamic pose field at the current moment.
4. The railway-road intermodal steel coil turning and clamping conveying system according to claim 3, characterized in that, Assigning covariance scheduling coefficients to each data source at the current moment specifically includes: Take multiple historical poses adjacent to the current travel index in the settlement baseline field, and calculate the historical dispersion amplitude of the visual target coordinates, inertial attitude and wheel mileage at each pose. The current deviation of each data source is obtained by subtracting the visual target coordinates, inertial attitude, and wheel mileage collected at the current moment from the pre-stored pose corresponding components at the current travel index in the settlement baseline field. The current deviation of each data source is compared with its historical dispersion, and the covariance scheduling coefficient of each data source is determined based on the comparison results. Data sources in the current trip index Covariance scheduling coefficient at the location Determine using the following formula: ； ； in, The length of the history index window is set to 20. To prevent the constant from being zero, we take 0.01; For the first The data source in the first The difference between the measured value at each travel index and the pre-stored value of the settlement baseline field; For the first The data source in the first The measured value at the first travel index is the same as the first... The difference between the measured values at each trip index, the larger the deviation relative to the historical dispersion of the data source, the smaller the scheduling coefficient value.
5. The railway-road intermodal steel coil turning and clamping conveying system according to claim 1, characterized in that, The process for obtaining the travel drive amount is as follows: Extract the vertical settlement values at each historical travel index along the travel direction in the settlement baseline field, and convert the vertical settlement values into equivalent settlement components along the travel direction based on the installation inclination angle of the track or rack at each index point: ； in, For the first Vertical settlement values at each index point Pre-adjust the equivalent displacement component in the direction of travel at that point. To calibrate the conversion factors, , For the first The angle of inclination of the track or rack mounting surface relative to the horizontal plane at each index point; then, according to the travel direction... By recursively summing up each item, the cumulative settlement trend value is obtained and used as the displacement pre-adjustment value; The target landing coordinates in the dynamic pose field are corrected in reverse by the displacement pre-adjustment value, and the difference between the corrected mapping position and the current vehicle pose in the direction of travel is used as the basic travel. Extract the travel drive amount corresponding to the basic stroke, and add the displacement pre-adjustment value to the travel drive amount in the same direction to form the travel drive amount after superimposing the displacement pre-adjustment value.
6. The railway-road intermodal steel coil turning and clamping conveying system according to claim 5, characterized in that, The process of obtaining the basic itinerary is as follows: Read the target's landing coordinates in the dynamic pose field and apply the displacement pre-adjustment value in the opposite direction of the travel direction to the corresponding mapping position to obtain the corrected mapping position after settlement compensation. Based on the dynamic pose field at the current moment, extract the current coordinates of the vehicle body pose in the direction of travel; The difference between the corrected mapping position's coordinates in the travel direction and the current coordinates is used as the base travel distance.
7. The railway-road intermodal steel coil turning and clamping conveying system according to claim 1, characterized in that, The final positioning deviation obtained from the screening process specifically includes: Before the steel coil descends, the target landing coordinates are initially collected, and the first feedback amount is obtained by subtracting the pose corresponding to the initial collected coordinates and the travel drive amount. After a preset interval, the target's landing coordinates are collected a second time, and the second feedback amount is obtained by subtracting the pose corresponding to the second collected coordinates from the walking drive amount. The first and second feedback quantities are compared with the historical feedback quantity distributions at the corresponding travel indexes in the settlement baseline field to determine their degree of difference. The confidence level of each feedback quantity is determined based on the degree of difference. The contribution of each feedback quantity to the final positioning deviation is then distributed in reverse order of confidence level, and the final positioning deviation is obtained by combining the results.
8. The railway-road intermodal steel coil turning and clamping conveying system according to claim 1, characterized in that, The process of placing the steel coil onto the target transport vehicle after clamping, locking, and attitude conversion specifically includes: The current travel drive is corrected in reverse by the final positioning deviation, and travel compensation is performed according to the corrected travel drive to make the vehicle reach the final position after compensation. The final pose is decomposed into lateral position components, longitudinal position components, and heading angle components; The clamping and locking are executed when the lateral and longitudinal position components reach preset thresholds. After clamping and locking are completed, the target rotation angle around the horizontal rotation axis is obtained from the preset attitude conversion table according to the target attitude of the steel coil, and the rotation platform of the rotation mechanism is aligned according to the heading angle component, and the rotation actuator is driven to rotate around the horizontal rotation axis until the target rotation angle is reached. After placement, the deviation between the measured placement coordinates and the target placement coordinates is taken as the placement deviation and backfilled into the corresponding travel index in the settlement baseline field to complete the baseline update.
9. The railway-road intermodal steel coil turning and clamping conveying system according to claim 8, characterized in that, The method of correcting the current driving amount in reverse based on the final positioning deviation specifically includes: The final positioning deviation is decomposed into a travel deviation component and a lateral deviation component along the travel direction. The travel deviation component is then superimposed onto the current travel drive amount to obtain the initial travel drive amount. The lateral deviation component is added in the same direction to the lateral drive portion of the initial repair travel drive to obtain the repair travel drive. Longitudinal travel compensation is performed using the travel drive portion of the repair travel drive quantity. After the longitudinal travel compensation is in place, lateral travel compensation is performed using the lateral movement drive portion of the repair travel drive quantity, so that the vehicle body reaches the final position after compensation.