A method, system and medium for automatic driving of molten iron ladle cars

Abstract

This invention belongs to the field of automatic driving control technology for molten iron transportation in the steel industry, and specifically relates to an automatic driving method, system, and medium for molten iron ladle cars. The method includes: constructing a multi-mass dynamic model to obtain the nonlinear relationship between the driving parameters of the ladle car and the parameters of the molten iron ladle; generating the desired speed for each driving scenario and collecting the speed error between the driving speed and the desired speed; autonomously learning the speed error to dynamically generate the proportional coefficient and integral coefficient of the PI controller, and using the output of the PI controller as the control signal for the traction and braking force of the molten iron ladle car; predicting the sloshing amplitude of molten iron under different braking forces and optimizing the braking force of the molten iron ladle car; combining the target distance and driving parameters, solving the piecewise affine using the EMPC algorithm, and then locating the state partition corresponding to the piecewise affine through online table lookup to optimize the braking force of the molten iron ladle car, thereby achieving smooth automatic driving of the molten iron ladle car.

Description

Technical Field

[0001] This invention relates to the field of automatic driving control technology for molten iron transportation in the steel industry, and particularly to an automatic driving method, system and medium for molten iron ladle cars. Background Technology

[0002] As the primary equipment for transporting molten iron in steel enterprises, ladle cars operate under various conditions, including empty ladles and molten iron-loaded ladles, as well as on straight tracks, curves, and level crossings. They also require different alignments at the ladle station and transfer station. This places extremely high demands on the stability, precision, and safety of motion control. Traditional ladle cars are manually driven, with drivers relying on experience to control speed and braking. This makes it difficult to accurately match the motion characteristics under different load conditions. When carrying molten iron, there are risks of large mass and molten iron sloshing, and problems such as excessive braking distance, unstable driving, and even molten iron splashing accidents are common. Furthermore, manual driving results in low precision in ladle alignment and stopping. Large errors in aligning the ladle at the outlet can easily lead to molten iron leakage, while deviations at the transfer station affect the efficiency of overhead crane operations.

[0003] In view of this, the present invention is proposed. Summary of the Invention

[0004] To address the problems of poor safety, insufficient stability, and low positioning accuracy of existing manual driving of molten iron ladle cars, as well as the single control target and inability of existing automation solutions to adapt to multi-condition and multi-scenario collaborative control, this invention provides an automatic driving method, system, and storage medium for molten iron ladle cars.

[0005] To address the aforementioned technical problems, this invention provides an autonomous driving method for molten iron ladle cars. The method includes: constructing a multi-mass dynamics model based on molten iron ladle parameters and solid-liquid coupling sloshing characteristics to obtain the nonlinear relationship between the car's driving parameters and the molten iron ladle parameters; correcting the drag coefficient in the multi-mass dynamics model based on the driving scenario; generating the desired speed for each driving scenario based on the nonlinear relationship between the driving parameters, and collecting the speed error between the driving speed and the desired speed during car operation; autonomously learning the speed error using a BP neural network to dynamically generate the proportional and integral coefficients of a PI controller, and using the PI controller's output as the control signal for the molten iron ladle car's traction and braking forces; constructing a Model Predictive Control (MPC) algorithm to predict the molten iron sloshing amplitude under different braking forces based on the molten iron ladle car's driving parameters and driving scenarios, and optimizing the molten iron ladle car's braking force based on the molten iron sloshing amplitude; calculating the target distance between the molten iron ladle car and the target parking position, and solving the piecewise affine using an EMPC algorithm based on the target distance and driving parameters, then locating the corresponding state partition of the piecewise affine using an online lookup table to optimize the molten iron ladle car's braking force.

[0006] In embodiments of the present invention, the parameters of the molten iron ladle include the ladle size, the number of ladles, and the molten iron capacity; the driving parameters include traction force, driving speed, braking force, and braking distance; and the driving scenarios include straight roads, curves, uphill slopes, downhill slopes, and level crossings.

[0007] In an embodiment of the present invention, before the steps of constructing a model predictive control (MPC) algorithm, predicting the sloshing amplitude of molten iron under different braking forces by combining the driving parameters and driving scenarios of the molten iron ladle car, and optimizing the braking force of the molten iron ladle car based on the sloshing amplitude, the method further includes:

[0008] During the movement of the molten iron ladle car, the camera is used to detect the first preset marker information along the route in real time;

[0009] The first distance between the molten iron ladle car and the first marker information is detected by using lidar, and the first absolute coordinates of the current ladle car are calculated by combining the first distance and driving parameters.

[0010] Substitute the first absolute coordinates into the pre-stored transportation map to obtain the driving scenarios that the molten iron tanker will pass through within a preset time period in the future.

[0011] In an embodiment of the present invention, the steps of constructing a model predictive control (MPC) algorithm, predicting the sloshing amplitude of molten iron under different braking forces by combining the driving parameters and driving scenarios of the molten iron ladle car, and optimizing the braking force of the molten iron ladle car based on the sloshing amplitude, further include:

[0012] When the molten iron ladle car approaches the driving scene it is about to pass through, the decision to activate MPC control is made based on the speed error between the ladle car's current speed and the expected speed.

[0013] When the speed error exceeds the error threshold, MPC control is activated, and multiple braking forces are simulated and input into the multi-mass dynamics model to predict the sloshing amplitude of molten iron under each braking force, and the braking force corresponding to the smallest sloshing amplitude is used as the braking force output.

[0014] When the speed error is less than or equal to the threshold, the output of the PI controller continues to be used as the control signal for the traction and braking force of the molten iron ladle car.

[0015] In an embodiment of the present invention, before the steps of calculating the target distance between the molten iron ladle car and the target parking position, combining the target distance and driving parameters, solving the piecewise affine using the EMPC algorithm, and then locating the state partition corresponding to the piecewise affine by online table lookup to optimize the braking force of the molten iron ladle car, the method further includes:

[0016] The camera is used to detect the second marker information preset in front of the tank station or transshipment station in real time.

[0017] The second distance between the molten iron ladle car and the second marker information is detected by using lidar, and the second absolute coordinates of the current ladle car are calculated by combining the second distance and driving parameters.

[0018] In an embodiment of the present invention, the steps of calculating the target distance between the molten iron ladle car and the target parking position, combining the target distance and driving parameters, solving the piecewise affine using the EMPC algorithm, and then locating the state partition corresponding to the piecewise affine by online table lookup to optimize the braking force of the molten iron ladle car, further include:

[0019] The first target distance between the molten iron ladle car and the target parking position is calculated by combining the second absolute coordinates and the coordinates of the target parking position;

[0020] Using the driving parameters, molten iron ladle parameters, and the distance to the first target as inputs, the first segmented affine is solved using the EMPC algorithm.

[0021] The first state partition corresponding to the first segment affine mapping is located by looking up the table online.

[0022] The first braking force output is calculated directly based on the first state partition.

[0023] In an embodiment of the present invention, after the step of directly calculating the first braking force output based on the first state partition, the method further includes:

[0024] Millimeter-wave radar was used to detect the distance between a second target and the target parking position of a molten iron ladle car.

[0025] Using the driving parameters, molten iron ladle parameters, and the distance to the second target as inputs, the second piecewise affine is solved using the EMPC algorithm.

[0026] The second state partition corresponding to the second segment affine is located by looking up the table online.

[0027] The optimal braking force output is calculated directly based on the second state partition.

[0028] In an embodiment of the present invention, the automatic driving method for molten iron ladle cars further includes:

[0029] During the movement of the molten iron ladle car, cameras are used to monitor unexpected situations in real time.

[0030] When an emergency is detected, lidar is used to detect the third distance between the molten iron ladle car and the emergency.

[0031] When the third distance is greater than the preset safety distance, MPC control is activated, and multiple braking forces are simulated and input into the multi-mass dynamics model to predict the sloshing amplitude of molten iron under each braking force, and the braking force corresponding to the smallest sloshing amplitude is used as the braking force output.

[0032] Emergency braking is triggered when the third distance is less than or equal to the preset safe distance.

[0033] To solve the above-mentioned technical problems, the present invention also provides an automatic driving system for molten iron ladle cars, the automatic driving system for molten iron ladle cars including a modeling unit, an error calculation unit, an autonomous control unit, a scene control unit, and a parking control unit;

[0034] The modeling unit is configured to: construct a multi-mass dynamic model based on the parameters of the molten iron ladle and the solid-liquid coupling sloshing characteristics, so as to obtain the nonlinear relationship between the driving parameters of the ladle car and the parameters of the molten iron ladle, and correct the drag coefficient in the multi-mass dynamic model in combination with the driving scenario;

[0035] The error calculation unit is configured to: generate the expected speed for each driving scenario based on the nonlinear relationship between driving parameters, and collect the speed error between the driving speed and the expected speed during the operation of the tanker truck;

[0036] The autonomous control unit is configured to: autonomously learn the speed error through a BP neural network, dynamically generate the proportional coefficient and integral coefficient of the PI controller, and use the output of the PI controller as the control signal for the traction and braking force of the molten iron ladle car.

[0037] The scenario control unit is configured to: construct a model predictive control (MPC) algorithm, combine the driving parameters of the molten iron ladle car and the driving scenario to predict the sloshing amplitude of molten iron under different braking forces, and optimize the braking force of the molten iron ladle car based on the sloshing amplitude of molten iron.

[0038] The parking control unit is configured to: calculate the target distance between the molten iron ladle car and the target parking position; combine the target distance and driving parameters; solve the piecewise affine using the EMPC algorithm; and then locate the corresponding state partition of the piecewise affine by online table lookup in order to optimize the braking force of the molten iron ladle car.

[0039] To solve the above-mentioned technical problems, the present invention also provides a computer storage medium, wherein the computer-readable medium stores a computer program, wherein the computer program is configured to execute the above-mentioned automatic driving method for molten iron ladle cars when running.

[0040] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0041] The automatic driving method for molten iron ladle cars of the present invention can comprehensively consider the sloshing characteristics of molten iron, changes in driving scenarios, and parking accuracy requirements during the transportation of molten iron, so as to achieve smooth, safe, and precise automatic driving of molten iron ladle cars, avoid the risk of splashing or overflowing of molten iron due to excessive sloshing during transportation, and improve the overall operational efficiency of molten iron transportation.

[0042] Other features and advantages of the embodiments of the present invention will be described in the following detailed embodiments section. Attached Figure Description

[0043] To more clearly illustrate the specific embodiments of this application or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0044] Figure 1 This is a table showing the correspondence between the maximum speed and braking distance requirements for autonomous driving provided in the embodiments of this application.

[0045] Figure 2 This is a flowchart illustrating the overall steps of the automatic driving method for molten iron ladle cars provided in the embodiments of this application;

[0046] Figure 3 This is a schematic diagram of the multi-mass dynamics model in the automatic driving method for molten iron ladle cars provided in the embodiments of this application;

[0047] Figure 4 This is a schematic diagram of the BP-PI control model in the automatic driving method for molten iron ladle cars provided in the embodiments of this application;

[0048] Figure 5 This is a schematic diagram of the MPC algorithm model in the automatic driving method for molten iron ladle cars provided in the embodiments of this application;

[0049] Figure 6 This is a schematic diagram of the EMPC algorithm (online and offline) model in the automatic driving method for molten iron ladle cars provided in the embodiments of this application;

[0050] Figure 7 This is a module framework diagram of the automatic driving system for molten iron ladle cars provided in an embodiment of this application;

[0051] 1. Autonomous driving system; 11. Modeling unit; 12. Error calculation unit; 13. Autonomous control unit; 14. Scene control unit; 15. Parking control unit. Detailed Implementation

[0052] Unless otherwise specified, the terms “second direction,” “first direction,” “third direction,” “inner,” and “outer” used in the following descriptions, indicating orientation or positional relationships, are understood to be based on the orientation or positional relationships shown in the accompanying drawings. They are used only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this application.

[0053] Furthermore, features specified with "first" or "second" for descriptive purposes only should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Features specified with "first" or "second" may explicitly or implicitly include at least one of the specified features. The description of "multiple" generally means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0054] In this application, unless otherwise explicitly specified and limited, terms such as "installation," "connection," "joining," and "fixing" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can be a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components. Those skilled in the art can understand the specific meaning of the above terms in this application based on the specific circumstances.

[0055] In the description of this specification, the terms "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that the specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0056] Based on the foregoing, and according to overall safety requirements, the control requirements for tank cars vary depending on different load conditions, railway conditions, and levels of urgency. One specific requirement is as follows: Figure 1 As shown.

[0057] This invention provides an automatic driving method, system, and medium for molten iron ladle cars. During the transportation of molten iron, the method comprehensively considers the sloshing characteristics of molten iron, changes in driving scenarios, and parking accuracy requirements to achieve smooth, safe, and precise automatic driving of the molten iron ladle cars. This avoids the risk of splashing or overflowing of molten iron due to excessive sloshing during transportation, while improving the overall operational efficiency of molten iron transportation.

[0058] like Figure 2 As shown, to solve the above-mentioned technical problems, the present invention provides an automatic driving method for molten iron ladle cars, the automatic driving method for molten iron ladle cars comprising:

[0059] Step S1: Construct a multi-mass dynamic model based on the parameters of the molten iron ladle and the solid-liquid coupling sloshing characteristics to obtain the nonlinear relationship between the driving parameters of the ladle car and the parameters of the molten iron ladle, and correct the drag coefficient in the multi-mass dynamic model in combination with the driving scenario;

[0060] Step S2: Generate the expected speed for each driving scenario based on the nonlinear relationship between driving parameters, and collect the speed error between the driving speed and the expected speed during the operation of the tanker truck;

[0061] Step S3: The speed error is autonomously learned through the BP neural network, and the proportional coefficient and integral coefficient of the PI controller are dynamically generated. The output of the PI controller is used as the control signal for the traction and braking force of the molten iron ladle car.

[0062] Step S4: Construct a model predictive control (MPC) algorithm, combine the driving parameters and driving scenarios of the molten iron ladle car to predict the sloshing amplitude of molten iron under different braking forces, and optimize the braking force of the molten iron ladle car based on the sloshing amplitude of molten iron.

[0063] Step S5: Calculate the target distance between the molten iron ladle car and the target parking position. Combining the target distance and driving parameters, solve the piecewise affine using the EMPC algorithm. Then, locate the corresponding state partition of the piecewise affine by looking up the table online to optimize the braking force of the molten iron ladle car.

[0064] By constructing a multi-mass dynamics model, the frequency and amplitude of molten iron sloshing within the ladle can be calculated during the operation of the ladle car under different driving parameters. The nonlinear relationship between the car's driving parameters and the ladle's parameters can then be obtained. This nonlinear relationship can be referenced in subsequent autonomous driving to ensure the smooth operation of the ladle car. Furthermore, the drag coefficient in the multi-mass dynamics model can be modified according to the driving scenario, such as increasing the slope drag coefficient when going uphill. This ensures that the multi-mass dynamics model accurately reflects the sloshing characteristics of the molten iron within the ladle under different driving scenarios, further guaranteeing the driving stability of the ladle car and avoiding the risk of molten iron splashing or overflowing.

[0065] In one specific embodiment, a multi-mass dynamics model is as follows: Figure 3 The example provided illustrates a multi-mass-solid-liquid coupling dynamics model for a molten iron ladle car. The train is simplified to five rigid mass points (the locomotive and four molten iron ladles), connected by spring dampers to simulate the buffering characteristics of the coupler. Each ladle is considered a rigid body, and the molten iron inside is simplified as an equivalent swaying mass point, constructing a solid-liquid coupling model. When fully loaded, the molten iron is considered as a swaying equivalent mass block, connected to the ladle body via lateral / longitudinal spring dampers to simulate the nonlinear relationship between swaying force and displacement. When unloaded, swaying is ignored, and only the rigid mass of the ladle body is considered.

[0066] Table 1

[0067] m0 Locomotive (including driver's cab) m0 Connected to the M1 coupler (spring damping) No molten iron, only rigid body mass m1 No. 1 molten iron ladle mc1+ms1 Connected to m0 and m2 couplers Tank body MC1, molten iron ms1, sway stiffness ks1, damping CS1 m2 No. 2 molten iron ladle mc2+ms2 Connected to M1 and M3 couplers Tank body MC2, molten iron ms2, sway stiffness ks2, damping CS2 m3 No. 3 molten iron ladle mc3+ms3 Connected to M2 and M4 couplers Tank body MC3, molten iron ms3, sway stiffness ks3, damping CS3 m4 No. 4 molten iron ladle mc4+ms4 Connected to M3 coupler The tank body is made of MC4 steel, the molten iron is made of MS4 steel, the sway stiffness is made of KS4 steel, and the damping is made of CS4 steel.

[0068] Longitudinal model:

[0069] Taking the i-th mass point of the tank (i=1,2,3,4) as an example, its longitudinal forces include: the spring force and damping force of the front and rear couplers, the longitudinal inertial force and damping force generated by the sloshing of molten iron, the rolling resistance of the wheels, and the braking force / traction force.

[0070] The equation of motion for the mass of the tank, mi = mci + msi:

[0071] (mci+msi) i=F c,i−1 -F c,i +Fs,iF f,i -F b,i

[0072] F c,i−1 =k c (x i−1 -x i )+c c ( i−1 - i )

[0073] F c,i =k c (x i -x i+1 )+c c ( i - i+1 )

[0074] F s,i =k si (x si -x i )+c si ( si - i )

[0075] F f,i =f r (m ci +m si )g

[0076] Where, x i Longitudinal displacement of the tank; x si : Longitudinal displacement of equivalent sloshing particles in molten iron; F c,i : The force of the i-th coupler; Fs,i : The sloshing force of molten iron on the ladle; F f,i Rolling resistance (f) r (This refers to the rolling resistance coefficient, i.e., the resistance coefficient mentioned above); F b,i Braking force / traction force.

[0077] molten iron sloshing particle m si Equation of motion: m s si =−F s,i ;

[0078] Therefore, by combining the equations, we can obtain: (m ci +m si ) i =k c (x i−1 -2x i +x i+1 )+c c ( i−1 -2 i + i+1 )+k si (x si -x i )+c si ( si - i )−F f,i -F b,i ;

[0079] m si si =−k si (x si -x i )−c si ( si - i )

[0080] Lateral model (curving / acceleration / deceleration scenario):

[0081] The lateral aspect mainly reflects the influence of molten iron sloshing on the ladle's posture: m ci i =k c (y i−1 -2y i +y i+1 )+c c ( i−1 -2 i + i+1 )+k sy,i (y si -y i )+c sy,i ( si −y˙ i )+m si a yi

[0082] m si si =−k sy,i (y si -y i )−c sy,i ( si− i )−m si a yi

[0083] Among them, a yi k is the lateral acceleration of the vehicle (centrifugal or steering acceleration in curves). sy,i c sy,i For lateral sway stiffness and damping.

[0084] Furthermore, the actual resistance experienced by molten iron ladle cars varies depending on the driving scenario (straight road, curve, level crossing, parallel parking, etc.), making a fixed "resistance coefficient" unsuitable for all scenarios. Therefore, the resistance coefficient in the model is dynamically adjusted based on the current road conditions, vehicle speed, load, and other factors, making the dynamic calculations more closely resemble real-world conditions. In other words, the resistance coefficient of the multi-mass dynamics model is dynamically corrected based on whether the vehicle is on a straight road, curve, level crossing, or parallel parking, thus improving the accuracy of the dynamics model.

[0085] In the above embodiments, the parameters of the molten iron ladle include the ladle size, the number of ladles, and the molten iron capacity; the driving parameters include traction force, driving speed, braking force, and braking distance; the driving scenarios include straight roads, curves, uphill slopes, downhill slopes, and level crossings.

[0086] The aforementioned nonlinear relationship generates the desired speed for each driving scenario. For example, when a molten iron ladle is transporting four fully loaded molten iron ladles with a capacity of 180 tons each, the desired speed on a straight road is 15 km / h, and on a curve, the desired speed is 5 km / h. Similarly, when transporting six empty molten iron ladles with a capacity of 180 tons each, the desired speed on a straight road is 20 km / h, and on a curve, the desired speed is 15 km / h. This ensures both transportation efficiency and a smooth driving experience during molten iron transport, preventing splashing or spillage. During actual operation, the molten iron ladle's CAN bus driving parameters are used to calculate the speed error between the actual speed and the desired speed, providing precise input variables for the subsequent PI controller.

[0087] In one example, such as Figures 4 to 6 As shown, a BP neural network is used to autonomously learn the speed error. When the speed error is large, the BP neural network automatically increases the proportional coefficient to speed up the response; when the error fluctuates frequently, the integral coefficient is adjusted appropriately to avoid integral saturation. The output of the PI controller is used as the control signal for the traction and braking force of the molten iron ladle car. For example, a positive output signal indicates an increase in traction force, and a negative output signal indicates the application of braking force. This achieves adaptive adjustment of the PI controller parameters. Compared with fixed-parameter PI control, it can better adapt to load changes and road condition fluctuations, ensuring the speed tracking speed and stability.

[0088] Autonomous driving via BPNN and PI, as shown in the diagram. r The ideal velocity curve is the reference velocity; v i The actual operating speed of the locomotive is represented by ui, which is the system's output signal; ui is the locomotive's traction / braking force (the resultant force of air braking and electric braking), which is the system's control signal; and error is the error between the locomotive's actual speed and the ideal target speed. The locomotive speed tracking controller is a closed-loop feedback controller. Compared to an open-loop control system without feedback, a closed-loop control system increases the system's anti-interference capability and reduces its sensitivity to noise. Through a PI control algorithm based on a BP neural network, the system error between the locomotive's actual operating speed and the target speed is learned, dynamically generating the control parameters of the PI controller to enable the locomotive to track the ideal speed target curve and achieve smooth locomotive operation.

[0089] Further as Figure 5This application can selectively predict the amplitude of molten iron sloshing through the Model Predictive Control (MPC) algorithm, and optimize the braking force of the molten iron ladle car based on the amplitude of molten iron sloshing. This enables the automatic driving method of the molten iron ladle car to have forward prediction capability, realize the active suppression of molten iron sloshing caused by scene changes (such as curves, uphill or downhill), prevent molten iron from splashing due to excessive or insufficient braking, and ensure the safety of the transportation process.

[0090] In this application, the target parking positions are set as the ladle station and the transfer station, respectively. When the ladle car approaches the target parking position, the target distance (e.g., 50m) between the ladle car and the target parking position is calculated. Then, the ATO (Automatic Train Operation) algorithm based on EMPC (Explicit Model Predictive Control) is used. The MPC optimization problem is transformed into a linear function corresponding to multiple state partitions offline. This allows the corresponding piecewise affine function to be retrieved directly based on the target distance and driving parameters when solving the piecewise affine function, improving the calculation speed. Then, the state partition to which the piecewise affine function belongs is located by online table lookup, and the optimal braking force output is directly calculated. This greatly reduces the computational burden during the parking phase, while ensuring braking accuracy and the suppression of molten iron sloshing, ensuring that the ladle car stops smoothly and accurately at the target position.

[0091] Understandably, based on a general inventive concept of the embodiments of this application, and addressing the precise parking requirements of molten iron ladle cars at ladle stations and transfer stations, the EMPC-ATO algorithm is adopted. Through "offline pre-calculation and online table lookup," it reduces the online computational load while ensuring control accuracy and suppressing molten iron sloshing, achieving smooth and precise parking. Specifically:

[0092] The ladle station and transfer station serve as precise parking target points for molten iron ladle cars. Both have stringent requirements for parking alignment accuracy (directly affecting the safety and efficiency of molten iron transfer operations). When the molten iron ladle car travels to a distance of 50m from the target parking position, the precise parking control process is triggered, entering the braking alignment stage.

[0093] The core mechanism of EMPC (Explicit Model Predictive Control) is designed as follows:

[0094] Offline phase: The optimization problem of Model Predictive Control (MPC) is decomposed into multiple state partitions in advance, and each state partition corresponds to a piecewise affine linear function (that is, the complex nonlinear optimization problem is transformed into a combination of multiple simple linear functions).

[0095] In the online phase: there is no need to solve complex MPC optimization equations in real time. Based on the target distance of the tanker (the remaining distance to the parking point) and driving parameters (such as vehicle speed, load, acceleration, etc.), the system quickly looks up the table to locate the partition to which the current vehicle state belongs, calls the piecewise ray simulacral function of the corresponding partition, and directly calculates the optimal braking force.

[0096] The molten iron ladle car receives the optimal braking force command output by the EMPC algorithm, driving the ladle car braking system to perform corresponding operations, achieving precise speed control and smooth stopping. By using "offline pre-calculation," the time-consuming process of solving complex optimization problems online is avoided, greatly reducing the online computational burden during the stopping phase and ensuring real-time control. At the same time, the precise matching of piecewise ray-like functions can effectively guarantee braking accuracy and suppress molten iron sloshing, ultimately achieving smooth and precise stopping of the ladle car at the target stopping position (ladle station, ladle transfer station), meeting operational requirements.

[0097] In this embodiment, the speed error is autonomously learned by a BP neural network to dynamically generate the proportional coefficient and integral coefficient of the PI controller, and the output of the PI controller is used as the control signal for the traction and braking force of the molten iron ladle car; or in the process of MPC and EMPC, constraints are applied when outputting control quantities, such as traction and braking force limits, workshop force limits, and maximum speed limits.

[0098] In summary, the automatic driving method for molten iron ladle cars provided in this application achieves precise adaptation and optimized control to meet the differentiated needs of different transportation and braking scenarios, and has at least the following beneficial effects:

[0099] 1) Achieve precise adaptation for different transportation scenarios to improve scenario adaptability and operational rationality: Combine the differences in road conditions of straight roads, curves, and level crossings in the plant area, as well as the alignment requirements of the iron tapping station and the transfer station, with differentiated control strategies. For example, on straight roads, the vehicle travels at a preset maximum speed to improve efficiency, while slowing down in advance at curves and level crossings to ensure safety; the iron tapping station uses EMPC for precise alignment to meet the requirements of the transfer station, and the transfer station balances accuracy and efficiency. At the same time, it distinguishes between two load scenarios: empty tanks and molten iron tanks. When the tanks are loaded with molten iron, the speed is reduced and the safety distance is increased to suppress the molten iron sloshing, while the speed is increased when the tanks are empty to improve transportation efficiency, thus achieving adaptive control for all transportation scenarios. Furthermore, it achieves optimized control for different braking scenarios, taking into account safety, stability, and accuracy: differentiating between different scenarios such as conventional braking, emergency braking, and precise positioning braking, and adopting layered collaborative control; for example, during conventional driving, BPNN+PI speed tracking control is used to achieve smooth acceleration and deceleration braking, avoiding molten iron sloshing; during emergency braking (such as at level crossings), the braking force is optimized online based on the MPC algorithm to ensure that the braking distance strictly meets the safety threshold (such as ≤9m at the iron-carrying waterway crossing), eliminating the risk of insufficient or over-braking; during precise positioning braking (at the tapping point and the transfer station), the EMPC algorithm is used in conjunction with a progressive braking strategy to first decelerate to a low speed and then make fine adjustments to achieve precise stopping, which not only meets the real-time braking requirements but also ensures a smooth braking process and precise positioning.

[0100] Achieving coordinated adaptation between transportation and braking scenarios to improve overall operational safety and efficiency: Deeply integrating road conditions and load requirements of different transportation scenarios with corresponding braking scenario control strategies, adapting to scenario differences by correcting dynamic model parameters, and matching scenario requirements through differentiated braking control, effectively solving the problems of traditional control being unable to adapt to multiple scenarios, insufficient braking accuracy, and high risk of molten iron sloshing, ensuring that tank cars drive smoothly, brake safely, and are accurately positioned in different scenarios, while taking into account both transportation efficiency and operational safety.

[0101] The following is a specific example:

[0102] In this embodiment of the application, the step S4 is further included as follows:

[0103] Step S041: During the travel of the molten iron ladle car, the camera is used to detect the first preset marker information along the route in real time;

[0104] Step S042: Use lidar to detect the first distance between the molten iron ladle car and the first marker information, and calculate the first absolute coordinates of the current ladle car by combining the first distance and driving parameters;

[0105] Step S043: Substitute the first absolute coordinates into the pre-stored transportation map to obtain the driving scene that the molten iron tanker will pass through in the future within a preset time period.

[0106] For example, a camera can detect the first marker information (such as a QR code target pasted 50m before a curve) along the route in real time. At the same time, a lidar can be used to detect the first distance between the molten iron tank car and the first marker information. The first absolute coordinates of the molten iron tank car can be calculated by combining the first distance and driving parameters. The first absolute coordinates can be substituted into a pre-stored transportation map to obtain the driving scene that the molten iron tank car will pass through in the future within a preset time period. This enables advance prediction of the road conditions ahead, provides sufficient reaction time for the MPC controller, and avoids control lag caused by sudden scene changes.

[0107] In an embodiment of the present invention, step S4 further includes

[0108] Step S401: When the molten iron ladle car approaches the driving scene it will pass through, decide whether to activate MPC control based on the speed error between the ladle car's driving speed and the expected speed.

[0109] Step S402: When the speed error exceeds the error threshold, MPC control is activated, and multiple braking forces are simulated and input into the multi-mass dynamics model to predict the sloshing amplitude of molten iron under each braking force, and the braking force corresponding to the smallest sloshing amplitude is used as the braking force output.

[0110] Step S403: When the speed error is less than or equal to the threshold, continue to use the output of the PI controller as the control signal for the traction and braking force of the molten iron ladle car.

[0111] Understandably, when a molten iron ladle car is about to enter different driving scenarios (such as straight tracks, curves, approaching level crossings / ladle stations), the actual speed of the ladle car is collected in real time and compared with the preset expected speed in that scenario (such as the expected speed of 5 km / h when carrying molten iron on a curve). The speed error between the two is calculated, which serves as the core basis for determining whether to enable MPC control (essentially, determining whether the current PI control can meet the scenario requirements).

[0112] When the calculated speed error exceeds the preset error threshold, it indicates that the current PI control can no longer accurately track the desired speed and may cause increased molten iron sloshing. At this point, MPC control is triggered. Simultaneously, multiple braking forces of different magnitudes (covering a reasonable range within the safety range) are substituted into the previously constructed multi-mass dynamics model (including the solid-liquid coupling sloshing characteristics of molten iron) to simulate and predict the molten iron sloshing amplitude corresponding to each braking force. Finally, the braking force with the "smallest molten iron sloshing amplitude" is selected as the actual output, taking into account both speed correction and sloshing suppression.

[0113] When the speed error is less than or equal to the preset threshold, it means that the current PI control can stably track the desired speed and the molten iron sloshing is within a controllable range. There is no need to enable the more computationally intensive MPC control. The traction / braking force signal output by the PI controller can be used to ensure the simplicity and real-time nature of the control.

[0114] By setting a speed error threshold, the control mode of the molten iron ladle car is determined. This allows the ladle car to only activate the computationally intensive MPC control when necessary (when the speed deviation is large and there is a change in the scene ahead), while using efficient PI control during smooth travel. This achieves a balance between control accuracy and computational resources. MPC control has a higher computational load than PI control. Triggering MPC activation through the "speed error threshold" avoids computational overload caused by using MPC throughout the entire process, ensuring that control commands can respond quickly when the ladle car changes scenes (such as straight sections and curves) or experiences speed fluctuations.

[0115] In an embodiment of the present invention, prior to step S5, the method further includes:

[0116] Step S051: Use a camera to detect the second marker information preset in front of the tank station or transfer station in real time;

[0117] Step S052: Use lidar to detect the second distance between the molten iron ladle car and the second marker information, and calculate the second absolute coordinates of the current ladle car by combining the second distance and driving parameters.

[0118] At the pre-set locations of the tanker station and transfer station (such as a suitable location in front of the target parking point), a second marker is pre-set (which can be understood as a positioning reference, such as a dedicated positioning sign, QR code, specific geometric mark, etc., adapted to the scene and easy for camera recognition). The camera on the tanker truck acquires images of the target area in real time, identifies and captures the second marker, confirming that the tanker truck has entered the alignment preparation area, providing a positioning reference for subsequent distance detection and coordinate calculation. Using the lidar on the tanker truck, the straight-line distance (i.e., the second distance) between the tanker truck and the identified second marker is accurately detected. Simultaneously, combining this second distance with the tanker truck's current driving parameters (such as real-time speed, heading angle, vehicle attitude, etc., from a multi-mass dynamics model or onboard sensors), a coordinate conversion algorithm calculates the tanker truck's current second absolute coordinates—this coordinate, based on the second marker, accurately reflects the tanker truck's real-time spatial position relative to the target parking location at the tanker station / transfer station, eliminating positioning errors.

[0119] It is understandable that by combining cameras and LiDAR to obtain the second absolute coordinates of the molten iron ladle car, a preliminary positioning reference under absolute coordinates can be provided for subsequent parking control, avoiding the problem of inaccurate positioning caused by the accumulation of errors in industrial environments when relying solely on inertial navigation or GPS.

[0120] In an embodiment of the present invention, step S5 further includes:

[0121] Step S501: Calculate the first target distance between the molten iron ladle car and the target parking position by combining the second absolute coordinates and the coordinates of the target parking position;

[0122] Step S502: Using the driving parameters, molten iron ladle parameters, and the distance to the first target as input, solve the first segmented affine using the EMPC algorithm;

[0123] Step S503: Locate the first state partition corresponding to the first segment affine by looking up the table online;

[0124] Step S504: Calculate the first braking force output directly based on the first state partition.

[0125] Understandably, based on the second absolute coordinates obtained from the previous fusion positioning, the difference between these coordinates and the preset target parking position coordinates is calculated to obtain the first target distance from the molten iron ladle car to the parking point, which serves as the core positional basis for subsequent precise braking. The vehicle's real-time driving parameters, the molten iron ladle's structure and load parameters, and the first target distance to the parking point are used as inputs. The corresponding first piecewise affine function, i.e., the optimal control law expression in the current state, is obtained through explicit model predictive control (EMPC). In the offline stage, EMPC has divided all feasible states into multiple state partitions and pre-stored them. During online operation, it only needs to look up the table and match the current state information to quickly determine the first state partition to which the current piecewise affine belongs. Based on the located state partition and the corresponding piecewise affine function, the optimal first braking force is directly calculated and output to the actuator to complete precise braking control.

[0126] First, the second absolute coordinates of the molten iron ladle car are acquired using cameras and LiDAR. Then, the first piecewise affine function is solved using the EMPC algorithm, and the first braking force output is directly calculated. EMPC employs offline pre-calculation and online table lookup, avoiding repeated online optimization and solving, significantly reducing computational latency and meeting the real-time braking requirements of the molten iron ladle car. Thus, the first-level parking control is performed based on the initial positioning reference, allowing the molten iron ladle car to slowly approach the target parking position. Furthermore, the same EMPC lookup framework can be used for different workstations such as the ladle station and the transfer station; differentiated and precise alignment can be achieved simply by adjusting the target coordinates, demonstrating strong versatility.

[0127] In an embodiment of the present invention, after step S504, the method further includes:

[0128] Step S5041: Use millimeter-wave radar to detect the distance between the molten iron ladle car and the target parking position;

[0129] Step S5041: Using the driving parameters, molten iron ladle parameters, and the distance to the second target as inputs, solve the second piecewise affine using the EMPC algorithm;

[0130] Step S5041: Locate the second state partition corresponding to the second segment affine by looking up the table online;

[0131] Step S5041: Calculate the optimal braking force output directly based on the second state partition.

[0132] After the initial EMPC braking force output, the distance to the second target between the molten iron ladle car and the target stopping position is measured in real time using millimeter-wave radar. The second piecewise affine control law is then solved again using the updated driving parameters, molten iron ladle parameters, and the second target distance measured by millimeter-wave radar. Based on the latest state, the second state partition corresponding to this piecewise affine is determined by looking up a table, achieving rapid matching. Based on the control law corresponding to the second state partition, the optimal braking force for the next round is directly calculated and output, allowing for fine-tuning and correction of the vehicle's braking.

[0133] The second target distance between the molten iron ladle car and the target parking position is obtained by millimeter-wave radar. Then, the second piecewise affine function is solved by the EMPC algorithm, and the second braking force output is directly calculated. A two-level positioning and two-level control strategy is adopted. Taking advantage of the high precision and high frequency ranging of millimeter-wave radar, a second precise positioning is performed when the molten iron ladle car approaches the target parking position, which improves the accuracy of the final parking position and avoids the blind spot problem that may occur in single visual positioning at close range.

[0134] In an embodiment of the present invention, the automatic driving method for molten iron ladle cars further includes:

[0135] Step S6: During the movement of the molten iron ladle car, use a camera to detect any unexpected situations in real time.

[0136] Step S7: When an emergency is detected, use lidar to detect the third distance between the molten iron ladle car and the emergency.

[0137] Step S8: When the third distance is greater than the preset safety distance, MPC control is activated, and multiple braking forces are simulated and input into the multi-mass dynamics model to predict the sloshing amplitude of molten iron under each braking force, and the braking force corresponding to the smallest sloshing amplitude is used as the braking force output.

[0138] Step S9: When the third distance is less than or equal to the preset safe distance, trigger emergency braking.

[0139] Throughout the tanker's journey, onboard cameras perform real-time image recognition of the driving path and surrounding environment to detect unexpected situations on the factory roads, such as foreign objects on the tracks, personnel intrusions, and obstacles. When the camera detects an emergency, it triggers a lidar system to accurately measure the distance between the tanker and the emergency, obtaining a third distance between the tanker and the emergency, providing a reliable distance basis for subsequent braking decisions. If the third distance is greater than a preset safety distance, it is determined to be a non-emergency dangerous state, and MPC model predictive control is activated. Multiple sets of braking forces are input into the multi-mass dynamics model to predict the molten iron sloshing amplitude corresponding to different braking forces, and the braking force with the smallest sloshing is selected to output, ensuring deceleration and obstacle avoidance while suppressing molten iron splashing. If the third distance is less than or equal to the preset safety distance, it is determined to be an emergency dangerous state, and the maximum braking force is directly triggered to execute emergency braking, stopping the vehicle in the shortest distance and prioritizing driving safety.

[0140] First, the distance between the molten iron ladle car and the emergency situation is obtained through cameras and lidar. Then, the braking method of the molten iron ladle car is determined in combination with the preset safety distance, taking into account both driving safety and the problem of molten iron sloshing. When the distance is sufficient, the minimum sloshing braking is selected to reduce the risk of molten iron splashing. When the distance is insufficient, the emergency braking is decisively performed to ensure safety first.

[0141] like Figure 2 As shown, in order to solve the above-mentioned technical problems, the present invention also provides an automatic driving system 1 for molten iron ladle cars. The automatic driving system 1 for molten iron ladle cars includes a modeling unit 11, an error calculation unit 12, an autonomous control unit 13, a scene control unit 14, and a parking control unit 15.

[0142] The modeling unit 11 is configured to: construct a multi-mass dynamic model based on the parameters of the molten iron ladle and the solid-liquid coupling sloshing characteristics, so as to obtain the nonlinear relationship between the driving parameters of the ladle car and the parameters of the molten iron ladle, and correct the drag coefficient in the multi-mass dynamic model in combination with the driving scenario;

[0143] The error calculation unit 12 is configured to: generate the expected speed corresponding to each driving scenario based on the nonlinear relationship between driving parameters, and collect the speed error between the driving speed and the expected speed when the tanker is running;

[0144] The autonomous control unit 13 is configured to: autonomously learn the speed error through a BP neural network, dynamically generate the proportional coefficient and integral coefficient of the PI controller, and use the output of the PI controller as the control signal for the traction and braking force of the molten iron ladle car.

[0145] The scenario control unit 14 is configured to: construct a model predictive control (MPC) algorithm, combine the driving parameters of the molten iron ladle car and the driving scenario to predict the sloshing amplitude of molten iron under different braking forces, and optimize the braking force of the molten iron ladle car based on the sloshing amplitude of molten iron.

[0146] The parking control unit 15 is configured to: calculate the target distance between the molten iron ladle car and the target parking position; combine the target distance and driving parameters; solve the piecewise affine using the EMPC algorithm; and then locate the state partition corresponding to the piecewise affine by online table lookup in order to optimize the braking force of the molten iron ladle car.

[0147] By constructing a multi-mass dynamics model, the frequency and amplitude of molten iron sloshing within the ladle can be calculated during the operation of the ladle car under different driving parameters. The nonlinear relationship between the car's driving parameters and the ladle's parameters can then be obtained. This nonlinear relationship can be referenced in subsequent autonomous driving to ensure the smooth operation of the ladle car. Furthermore, the drag coefficient in the multi-mass dynamics model can be modified according to the driving scenario, such as increasing the slope drag coefficient when going uphill. This ensures that the multi-mass dynamics model accurately reflects the sloshing characteristics of the molten iron within the ladle under different driving scenarios, further guaranteeing the driving stability of the ladle car and avoiding the risk of molten iron splashing or overflowing.

[0148] In the above embodiments, the parameters of the molten iron ladle include the ladle size, the number of ladles, and the molten iron capacity; the driving parameters include traction force, driving speed, braking force, and braking distance; the driving scenarios include straight roads, curves, uphill slopes, downhill slopes, and level crossings.

[0149] The aforementioned nonlinear relationship generates the desired speed for each driving scenario. For example, when a molten iron ladle is transporting four fully loaded molten iron ladles with a capacity of 180 tons each, the desired speed on a straight road is 15 km / h, and on a curve, the desired speed is 5 km / h. Similarly, when transporting six empty molten iron ladles with a capacity of 180 tons each, the desired speed on a straight road is 20 km / h, and on a curve, the desired speed is 15 km / h. This ensures both transportation efficiency and a smooth driving experience during molten iron transport, preventing splashing or spillage. During actual operation, the molten iron ladle's CAN bus driving parameters are used to calculate the speed error between the actual speed and the desired speed, providing precise input variables for the subsequent PI controller.

[0150] By employing a backpropagation (BP) neural network to autonomously learn speed errors, the BP neural network automatically increases the proportional coefficient to accelerate response when speed errors are large, and adjusts the integral coefficient appropriately to avoid integral saturation when error fluctuations are frequent. The output of the PI controller is used as the control signal for the traction and braking forces of the molten iron ladle car; for example, a positive output signal indicates increased traction, and a negative output signal indicates applied braking force. This achieves adaptive adjustment of the PI controller parameters. Compared to fixed-parameter PI control, it can better adapt to load changes and road condition fluctuations, ensuring rapid and stable speed tracking.

[0151] The Model Predictive Control (MPC) algorithm is used to predict the amplitude of molten iron sloshing, and the braking force of the molten iron ladle car is optimized based on the amplitude of molten iron sloshing. This enables the automatic driving method of the molten iron ladle car to have forward prediction capability, realize the active suppression of molten iron sloshing caused by scene changes (such as curves, uphill or downhill), prevent molten iron from splashing due to excessive or insufficient braking, and ensure the safety of the transportation process.

[0152] The target parking positions are set as the ladle station and the transfer station. When the ladle car approaches the target parking position, the target distance (e.g., 50m) between the ladle car and the target parking position is calculated. The EMPC algorithm is used offline to transform the MPC optimization problem into a linear function corresponding to multiple state partitions. Thus, when solving the piecewise affine, the corresponding piecewise affine can be directly retrieved based on the target distance and driving parameters, improving the calculation speed. Then, the state partition to which the piecewise affine belongs is located by online table lookup, and the optimal braking force output is directly calculated. This greatly reduces the calculation burden during the parking stage, while ensuring braking accuracy and the suppression of molten iron sloshing, ensuring that the ladle car stops smoothly and accurately at the target position.

[0153] To solve the above-mentioned technical problems, the present invention also provides a computer storage medium, wherein the computer-readable medium stores a computer program, wherein the computer program is configured to execute the above-mentioned automatic driving method for molten iron ladle cars when running.

[0154] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Any simple modifications, equivalent changes, and alterations made by those skilled in the art to the above embodiments based on the technical essence of the present invention shall still fall within the protection scope of the present invention.

[0155] The technical features described above can be combined arbitrarily. Although not all possible combinations of these technical features are described, any combination of these technical features should be considered to be covered by this specification, provided that such combination does not contain contradictions.

[0156] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and not to limit them; although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still adjust the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some or all of the technical features; and these adjustments or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A molten iron ladle car automatic driving method characterized by, The automatic driving method for the molten iron ladle car includes: A multi-mass dynamic model is constructed based on the parameters of the molten iron ladle and the solid-liquid coupling sloshing characteristics to obtain the nonlinear relationship between the driving parameters of the ladle car and the parameters of the molten iron ladle. The drag coefficient in the multi-mass dynamic model is then corrected in combination with the driving scenario. Based on the nonlinear relationship between driving parameters, the expected speed for each driving scenario is generated, and the speed error between the driving speed and the expected speed is collected during the operation of the tanker truck. The speed error is autonomously learned by the BP neural network, and the proportional coefficient and integral coefficient of the PI controller are dynamically generated. The output of the PI controller is used as the control signal for the traction and braking force of the molten iron ladle car. A model predictive control (MPC) algorithm is constructed to predict the sloshing amplitude of molten iron under different braking forces by combining the driving parameters and driving scenarios of the molten iron ladle car, and the braking force of the molten iron ladle car is optimized based on the sloshing amplitude of the molten iron. The target distance between the molten iron ladle car and the target parking position is calculated. Combining the target distance and driving parameters, the piecewise affine is solved using the EMPC algorithm. Then, the state partition corresponding to the piecewise affine is located by online table lookup in order to optimize the braking force of the molten iron ladle car.

2. The automatic driving method for molten iron ladle cars according to claim 1, characterized in that, The parameters of the molten iron ladle include the ladle size, the number of ladles, and the molten iron capacity; the driving parameters include traction force, driving speed, braking force, and braking distance; the driving scenarios include straight roads, curves, uphill slopes, downhill slopes, and level crossings.

3. The automatic driving method for molten iron ladle cars according to claim 2, characterized in that, Before the steps of constructing the Model Predictive Control (MPC) algorithm, predicting the sloshing amplitude of molten iron under different braking forces by combining the driving parameters and driving scenarios of the molten iron ladle car, and optimizing the braking force of the molten iron ladle car based on the sloshing amplitude, the following further steps are included: During the movement of the molten iron ladle car, the camera is used to detect the first preset marker information along the route in real time; The first distance between the molten iron ladle car and the first marker information is detected by using lidar, and the first absolute coordinates of the current ladle car are calculated by combining the first distance and driving parameters. Substitute the first absolute coordinates into the pre-stored transportation map to obtain the driving scenarios that the molten iron tanker will pass through within a preset time period in the future.

4. The automatic driving method for molten iron ladle cars according to claim 3, characterized in that, In the steps of constructing the Model Predictive Control (MPC) algorithm, combining the driving parameters and driving scenarios of the molten iron ladle car to predict the sloshing amplitude of molten iron under different braking forces, and optimizing the braking force of the molten iron ladle car based on the sloshing amplitude, the following further steps are included: When the molten iron ladle car approaches the driving scene it is about to pass through, the decision to activate MPC control is made based on the speed error between the ladle car's current speed and the expected speed. When the speed error exceeds the error threshold, MPC control is activated, and multiple braking forces are simulated and input into the multi-mass dynamics model to predict the sloshing amplitude of molten iron under each braking force, and the braking force corresponding to the smallest sloshing amplitude is used as the braking force output. When the speed error is less than or equal to the threshold, the output of the PI controller continues to be used as the control signal for the traction and braking force of the molten iron ladle car.

5. The automatic driving method for molten iron ladle cars according to claim 2, characterized in that, Before the steps of calculating the target distance between the molten iron ladle car and the target parking position, combining the target distance and driving parameters, solving the piecewise affine using the EMPC algorithm, and then locating the corresponding state partition of the piecewise affine through online table lookup to optimize the braking force of the molten iron ladle car, the following further steps are included: The camera is used to detect the second marker information preset in front of the tank station or transshipment station in real time. The second distance between the molten iron ladle car and the second marker information is detected by using lidar, and the second absolute coordinates of the current ladle car are calculated by combining the second distance and driving parameters.

6. The automatic driving method for molten iron ladle cars according to claim 5, characterized in that, The steps of calculating the target distance between the molten iron ladle car and the target stopping position, combining the target distance and driving parameters, solving the piecewise affine using the EMPC algorithm, and then locating the corresponding state partition of the piecewise affine through online table lookup to optimize the braking force of the molten iron ladle car further include: The first target distance between the molten iron ladle car and the target parking position is calculated by combining the second absolute coordinates and the coordinates of the target parking position; Using the driving parameters, molten iron ladle parameters, and the distance to the first target as inputs, the first segmented affine is solved using the EMPC algorithm. The first state partition corresponding to the first segment affine mapping is located by looking up the table online. The first braking force output is calculated directly based on the first state partition.

7. The automatic driving method for molten iron ladle cars according to claim 6, characterized in that, Following the step of directly calculating the first braking force output based on the first state partition, the method further includes: Millimeter-wave radar was used to detect the distance between a second target and the target parking position of a molten iron ladle car. Using the driving parameters, molten iron ladle parameters, and the distance to the second target as inputs, the second piecewise affine is solved using the EMPC algorithm. The second state partition corresponding to the second segment affine is located by looking up the table online. The optimal braking force output is calculated directly based on the second state partition.

8. The automatic driving method for molten iron ladle cars according to claim 4, characterized in that, The automatic driving method for molten iron ladle cars also includes: During the movement of the molten iron ladle car, cameras are used to monitor unexpected situations in real time. When an emergency is detected, lidar is used to detect the third distance between the molten iron ladle car and the emergency. When the third distance is greater than the preset safety distance, MPC control is activated, and multiple braking forces are simulated and input into the multi-mass dynamics model to predict the sloshing amplitude of molten iron under each braking force, and the braking force corresponding to the smallest sloshing amplitude is used as the braking force output. Emergency braking is triggered when the third distance is less than or equal to the preset safe distance.

9. An automatic driving system for molten iron ladle cars, characterized in that, The automatic driving system for the molten iron ladle car includes a modeling unit, an error calculation unit, an autonomous control unit, a scene control unit, and a parking control unit. The modeling unit is configured to: construct a multi-mass dynamic model based on the parameters of the molten iron ladle and the solid-liquid coupling sloshing characteristics, so as to obtain the nonlinear relationship between the driving parameters of the ladle car and the parameters of the molten iron ladle, and correct the drag coefficient in the multi-mass dynamic model in combination with the driving scenario; The error calculation unit is configured to: generate the expected speed for each driving scenario based on the nonlinear relationship between driving parameters, and collect the speed error between the driving speed and the expected speed during the operation of the tanker truck; The autonomous control unit is configured to: autonomously learn the speed error through a BP neural network, dynamically generate the proportional coefficient and integral coefficient of the PI controller, and use the output of the PI controller as the control signal for the traction and braking force of the molten iron ladle car. The scenario control unit is configured to: construct a model predictive control (MPC) algorithm, combine the driving parameters of the molten iron ladle car and the driving scenario to predict the sloshing amplitude of molten iron under different braking forces, and optimize the braking force of the molten iron ladle car based on the sloshing amplitude of molten iron. The parking control unit is configured to: calculate the target distance between the molten iron ladle car and the target parking position; combine the target distance and driving parameters; solve the piecewise affine using the EMPC algorithm; and then locate the corresponding state partition of the piecewise affine by online table lookup in order to optimize the braking force of the molten iron ladle car.

10. A computer storage medium, characterized in that, The computer-readable medium stores a computer program, wherein the computer program is configured to execute the automatic driving method for the molten iron ladle car according to any one of claims 1 to 8 when it is run.

Images