Self-balancing carriage control method and control system

Abstract

The invention relates to the technical field of vehicles, in particular to a self-balancing carriage control method and system, and the method comprises the steps: collecting pressure signals of cargoes in a carriage on two sides of the carriage through pressure sensors installed on the two sides of the carriage; the pressure signal is processed, and an oil pump pressure control instruction is obtained; responding to the oil pump pressure control instruction, adjusting the oil pump output pressure to push the balance weight to move along the slide rail, and adjusting the torque on two sides of the carriage to maintain the carriage balance. Pressure signals of goods in the carriage on the two sides of the carriage are collected through the pressure sensors installed on the two sides of the carriage, an oil pump pressure control instruction is obtained based on the pressure signals, so that the oil pump output pressure is adjusted to push the balance weight to move along the sliding rails, torque on the two sides of the carriage is adjusted, and the carriage balance is maintained; a high-pressure air injection device does not need to be installed, an active suspension does not need to be installed to adjust rigidity and damping force, cost is reduced, and the overall structure is simplified.

Description

Technical Field

[0001] The present invention relates to the technical field of vehicles, and in particular, to a self-balancing carriage control method and control system. Background Art

[0002] Currently, for the existing self-balancing carriage control technology, one method is to use a high-pressure jet device to provide a downward pressure on the tilted side to maintain the balance of the vehicle. Some other self-balancing carriage control systems use active suspensions to dynamically adjust the stiffness and damping force of the suspensions according to road conditions, so as to maintain the balance state of the vehicle, avoid the tilting of the carriage, and reduce property losses and safety hazards caused by the tilting of the carriage.

[0003] However, using a high-pressure jet device or adjusting the stiffness and damping force of the active suspension according to road conditions has a high cost and a relatively complex structure. Therefore, the existing self-balancing carriage control technology needs to be improved. Summary of the Invention

[0004] Based on this, in order to solve the problems existing in the prior art, the present invention provides a self-balancing carriage control method and control system. The specific technical solutions are as follows:

[0005] A self-balancing carriage control method includes the following steps:

[0006] Collect pressure signals of the goods in the carriage on both sides of the carriage through pressure sensors installed on both sides of the carriage;

[0007] Process the pressure signals to obtain an oil pump pressure control command;

[0008] Respond to the oil pump pressure control command, adjust the output pressure of the oil pump to push the balance weight to move along the slide rail, and adjust the moments on both sides of the carriage, so as to maintain the balance of the carriage.

[0009] The self-balancing carriage control method collects the pressure signals of the goods in the carriage on both sides of the carriage through pressure sensors installed on both sides of the carriage, obtains an oil pump pressure control command based on the pressure signals, and then adjusts the output pressure of the oil pump to push the balance weight to move along the slide rail, and adjusts the moments on both sides of the carriage, so as to maintain the balance of the carriage. It does not need to install a high-pressure jet device, nor does it need to install an active suspension to adjust the stiffness and damping force. It not only reduces the cost, but also simplifies the overall structure, and can reduce the risk of roll while not increasing additional weight and cost.

[0010] Preferably, the specific method for processing the pressure signals to obtain an oil pump pressure control command includes the following steps:

[0011] Process the pressure signals to obtain the pressure values measured by the pressure sensors on the left and right sides of the carriage in real time;

[0012] Define an oil pump pressure regulation function based on the direct mapping relationship between the pressure value and the output pressure of the oil pump;

[0013] Obtain an oil pump pressure control command according to the oil pump pressure regulation function.

[0014] Preferably, the specific method for obtaining the oil pump pressure control command according to the oil pump pressure regulation function includes the following steps:

[0015] Obtain the vehicle driving speed, acceleration, and road surface inclination angle;

[0016] Construct a neural network model, and use the pressure value, driving speed, acceleration, and road surface inclination angle as the inputs of the neural network model to obtain a dynamic correction factor;

[0017] Obtain the initial output pressure of the oil pump according to the oil pump pressure regulation function, and correct the initial output pressure of the oil pump according to the dynamic correction factor to obtain the final output pressure of the oil pump;

[0018] Obtain an oil pump pressure control command according to the final output pressure of the oil pump.

[0019] Preferably, the oil pump pressure regulation function is expressed as

[0020] where K p , K i , K d respectively represent the proportional coefficient, integral coefficient, and differential coefficient of the PID controller, P1 and P2 respectively represent the pressure values measured in real time by the pressure sensors on the left and right sides of the carriage, and P represents the initial output pressure of the oil pump.

[0021] Preferably, the dynamic correction factor ΔP = f NN (P1, P2, v, a, θ);

[0022] where f NN represents the neural network model, and v, a, and θ respectively represent the driving speed, acceleration, and road surface inclination angle.

[0023] Preferably, the final output pressure of the oil pump P' = P · (1 + ΔP).

[0024] A self-balancing carriage control system for implementing the self-balancing carriage control method, which includes:

[0025] Pressure sensors, respectively installed on both sides of the carriage, for collecting the pressure signals of the goods in the carriage on both sides of the carriage;

[0026] A controller, configured to receive the pressure signal, process the pressure signal, and obtain an oil pump pressure control command;

[0027] An oil pump, installed on a vehicle, is used to respond to the oil pump pressure control instruction, adjust the output pressure to push a balance weight to move along a slide rail, and adjust the moments on both sides of the carriage, so as to maintain the balance of the carriage.

[0028] Wherein, the slide rail is installed at the bottom of the carriage along the width direction of the vehicle, and the balance weight is located at the bottom of the carriage and is slidably connected to the slide rail.

[0029] Preferably, the controller obtains the initial output pressure P of the oil pump according to the formula ;

[0030] Wherein, K p , K i , K d respectively represent the proportional coefficient, integral coefficient and differential coefficient of the PID controller, and P1 and P2 respectively represent the pressure values measured in real time by the pressure sensors on the left and right sides of the carriage.

[0031] Preferably, the controller obtains a dynamic correction factor according to the formula ΔP = f NN (P1, P2, v, a, θ), corrects the initial output pressure of the oil pump according to the dynamic correction factor to obtain the final output pressure of the oil pump, and obtains an oil pump pressure control instruction according to the final output pressure of the oil pump.

[0032] Wherein, f NN represents a neural network model, and v, a, and θ respectively represent the driving speed, acceleration, and road surface inclination angle.

[0033] Preferably, the controller obtains the final output pressure P' of the oil pump according to the formula P' = P·(1 + ΔP). BRIEF DESCRIPTION OF THE DRAWINGS

[0034] The present invention can be further understood from the following description in conjunction with the drawings. The components in the drawings are not necessarily drawn to scale, but the emphasis is placed on showing the principles of the embodiments. In different views, the same reference numerals designate corresponding parts.

[0035] Figure 1 is a schematic overall flow chart of a self-balancing carriage control method in an embodiment of the present invention;

[0036] Figure 2 is a schematic flow chart of a specific method for obtaining an oil pump pressure control instruction in an embodiment of the present invention Figure 1 ;

[0037] Figure 3 is a schematic flow chart of a specific method for obtaining an oil pump pressure control instruction in an embodiment of the present invention Figure 2 ;

[0038] Figure 4 It is a schematic diagram of the overall structure of a self-balancing carriage control system in an embodiment of the present invention;

[0039] Figure 5 It is a schematic diagram of a partial structure of a self-balancing carriage control system in an embodiment of the present invention.

[0040] Explanation of reference numerals: 1, carriage; 2, partition; 3, pressure sensor; 4, controller; 5, oil pump; 6, balancing weight. Specific embodiments

[0041] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below in conjunction with its embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and do not limit the protection scope of the present invention.

[0042] It should be noted that when an element is referred to as being "fixed to" another element, it can be directly on the other element or there can also be an intermediate element. When an element is considered to be "connected" to another element, it can be directly connected to the other element or there may be an intermediate element at the same time. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are only for the purpose of illustration and do not represent the only implementation.

[0043] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which the present invention belongs. The terms used herein in the description of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more of the related listed items.

[0044] The "first" and "second" in the present invention do not represent specific quantities and sequences, but are only used for name distinction.

[0045] Before elaborating on the embodiments of the present invention, a brief introduction to the prior art will be given first.

[0046] In the field of modern transportation, the stability and safety of vehicles have always been important issues that have received much attention. Especially for some special-purpose vehicles, such as vehicles for transporting precision instruments, liquid goods or vehicles traveling on complex road conditions, the balance control of the carriage is particularly crucial.

[0047] Currently, in order to maintain the balance of a vehicle, various technical solutions have been proposed and applied. One common technique is to use a high-pressure jet device. The working principle of this device is that when the vehicle tilts, it provides a downward force (i.e., an upward reaction force) to the tilted side. By the action of this external force, the tilting moment is offset, thereby maintaining the balanced state of the vehicle. To a certain extent, this method can quickly respond to the tilting of the vehicle and has a certain effect in some short-time working conditions with high balance requirements.

[0048] Meanwhile, some advanced vehicle systems also adopt active suspension technology. The active suspension can dynamically adjust the stiffness and damping of the suspension according to road conditions. When the vehicle is driving on a bumpy road, the active suspension can quickly increase the damping to reduce the vibration of the vehicle body; when driving on a flat road, it can reduce the stiffness to improve the riding comfort. This technology plays a positive role in improving the driving performance and riding experience of the vehicle.

[0049] However, the above existing balance control solutions have obvious limitations. First of all, whether it is the high-pressure jet device or the active suspension system, their costs are relatively high. The high-pressure jet device needs to be equipped with a complex high-pressure gas storage and injection system, including high-pressure gas cylinders, valves, pipelines, and precise control systems, etc. The research and development, production, and maintenance of these components all require high cost investment. The same is true for the active suspension system. It contains many precise sensors, actuators, and electronic control units, which greatly increase the manufacturing cost of the vehicle, causing a certain obstacle to large-scale popularization and application.

[0050] Secondly, the structures of these solutions are relatively complex. The complex structure not only increases the failure rate of the system but also makes the repair and maintenance work more difficult and cumbersome. Moreover, due to the complexity of its structure, it is difficult to effectively solve the problem of carriage tilt caused by cargo tilt or uneven force. For example, during transportation, the cargo may move due to vehicle acceleration, deceleration, turning, or road bumps, etc., resulting in an increase in the weight on one side of the carriage, thus causing tilt. The existing high-pressure jet devices and active suspension systems often cannot accurately sense and compensate for the tilt caused by the change in cargo position, and cannot provide a long-term stable solution, making it difficult to meet the strict requirements for carriage balance in actual transportation.

[0051] With the rapid development of the modern logistics industry and the continuous improvement of the requirements for transportation safety and stability, there is an urgent need for a self-balancing carriage control method with relatively low cost, relatively simple structure, and capable of effectively solving the problem of carriage tilt caused by cargo tilt or uneven force, and providing long-term stable balance control.

[0052] For this reason, as Figure 1As shown, a self-balancing carriage control method in an embodiment of the present invention includes the following steps:

[0053] S1. Collect pressure signals of the goods in the carriage on both sides of the carriage through pressure sensors installed on both sides of the carriage.

[0054] Specifically, pressure sensors can be set on the left and right sides of the carriage. The sensors are signal-connected to the controller. When the pressure sensors sense the goods pressure, they transmit the pressure signals to the controller.

[0055] In order to better sense the goods pressure when the carriage is tilted through the pressure sensors, partitions can be respectively set on the left and right sides of the carriage. A movable compartment is formed between the partition and the carriage side plate, that is, the partition is configured to sense the goods pressure when the carriage is tilted and change the width of the compartment. The pressure sensors are installed on the partition and located in the compartment. Here, the partitions can be respectively fixedly installed on the left and right sides of the carriage, and are made of materials with certain elastic deformation, such as stainless steel plates or plastic plates. The partitions can also be respectively slidably connected to the carriage, that is, a small section of chute is set on the left and right sides of the carriage along the width direction of the carriage, and the partitions are installed in the chute.

[0056] S2. Process the pressure signals to obtain an oil pump pressure control instruction.

[0057] The controller processes the pressure signals to obtain the pressure values on the left and right sides of the carriage, analyzes and calculates the required oil pump pressure control instruction according to the pressure values on the left and right sides, and prompts the balance weight at the bottom of the carriage to move along the slide rail.

[0058] S3. Respond to the oil pump pressure control instruction, adjust the output pressure of the oil pump to push the balance weight to move along the slide rail, and adjust the torques on both sides of the carriage, so as to maintain the balance of the carriage.

[0059] Specifically, the output shaft of the oil pump is drivingly connected to the balance weight. The slide rail is arranged along the width direction of the carriage, and the balance weight is slidably connected to the slide rail. The oil pump can be a two-way hydraulic oil pump, which is configured to drive the piston rod to move in response to the oil pump pressure control instruction, so as to drive the balance weight to move along the slide rail, adjust the torques on the left and right sides of the carriage, and finally maintain the balance of the carriage.

[0060] Assume that P1 and P2 respectively represent the pressure values measured in real time by the pressure sensors on the left and right sides of the carriage. Then, the initial output pressure of the oil pump can be calculated according to the formula P = k×(P1 - P2). Wherein, k represents a proportionality coefficient, which is determined through experiments and optimization according to factors such as the structural parameters of the carriage and the mass of the balance weight. When P1 - P2 < 0, it means that the right side is more stressed, and the oil pump applies pressure to the left side to push the balance weight to move to the left; when P1 - P2 > 0, the oil pump applies pressure to the right side to push the balance weight to move to the right.

[0061] The oil pump outputs corresponding pressure according to the calculated initial output pressure value P, pushes the balance weight to move on the slide rail, changes the torque on both sides of the carriage, until the pressure on both sides of the carriage tends to be balanced, that is, |P1 - P2| ≤ ε, where ε is the set pressure balance threshold.

[0062] As an optimal technical solution, when the vehicle is an electric vehicle, the balance weight is a battery pack. In this way, it can not only effectively balance the carriage, but also avoid the increase of additional burden, so as to reduce the risk of roll while not increasing the additional mass and cost.

[0063] By using pressure sensors to monitor the pressure change of the carriage in real time, the system can respond quickly and make dynamic adjustments to ensure that the carriage always remains balanced under various conditions such as cargo shaking and turning, effectively improving the driving safety and stability of the vehicle. And by using a specific function formula P = k×(P1 - P2), a direct connection is established between the pressure difference on both sides of the carriage and the oil pump pressure regulation, realizing the quantification and precision of the control process, which is convenient for system debugging, optimization and maintenance.

[0064] The self-balancing carriage control method collects the pressure signals of the cargo on both sides of the carriage by pressure sensors installed on both sides of the carriage, obtains the oil pump pressure control command based on the pressure signals, and then adjusts the oil pump output pressure to push the balance weight to move along the slide rail, adjusts the torque on both sides of the carriage, so as to maintain the balance of the carriage. It does not need to install a high-pressure jet device, nor does it need to install an active suspension to adjust the stiffness and damping force. It not only reduces the cost, but also simplifies the overall structure, and can reduce the risk of roll while not increasing the additional weight and cost. Generally speaking, the self-balancing carriage control method has low cost, relatively simple structure and can effectively solve the problem of carriage tilt caused by cargo tilt or uneven force.

[0065] Although the initial output pressure of the oil pump can be calculated according to the formula P = k×(P1 - P2) and a direct connection is established between the pressure difference on both sides of the carriage and the oil pump pressure regulation, it does not combine real-time response, eliminate cumulative error and suppress mutation, and it is difficult to achieve high-dynamic-precision balance control.

[0066] To solve the above problems, in one embodiment, as Figure 2 shown, the specific method for processing the pressure signal and obtaining the oil pump pressure control command includes the following steps:

[0067] S21, process the pressure signal to obtain the pressure values measured by the pressure sensors on the left and right sides of the carriage in real time.

[0068] S22, define an oil pump pressure regulation function based on the direct mapping relationship between the pressure value and the oil pump output pressure.

[0069] S23. Obtain the oil pump pressure control command according to the oil pump pressure regulation function.

[0070] Preferably, the oil pump pressure regulation function is expressed as where K p , K i , K d respectively represent the proportional coefficient, integral coefficient and differential coefficient of the PID controller, P1 and P2 respectively represent the pressure values measured in real time by the pressure sensors on the left and right sides of the carriage, and P represents the initial output pressure of the oil pump.

[0071] The proportional coefficient can be determined by system structure parameters (such as carriage stiffness, heavy object mass) through experiments or optimization algorithms. The role of the proportional term is to dynamically adjust the oil pump output pressure according to the current left-right pressure difference. The greater the pressure difference, the stronger the adjustment force, and quickly offset the instantaneous imbalance caused by uneven cargo distribution or external shocks.

[0072] The integral coefficient determines the cumulative response intensity of the system to the historical pressure difference. ∫(P1 - P2)dt represents the time integral of the pressure difference, reflecting the cumulative effect of the pressure difference. The role of the integral term includes two aspects: 1. Eliminate the steady-state error: when the proportional term cannot completely eliminate the small pressure difference, the integral term gradually corrects it by accumulating historical deviations to avoid the balance drift caused by long-term partial load; 2. Cope with slow-varying disturbances: compensate for low-frequency disturbances such as road surface slope changes and slow cargo movement.

[0073] Specifically, when the vehicle turns and rolls, and the centrifugal force causes a left-right pressure difference, the differential term anticipates the imbalance trend, the proportional term quickly corrects it, and the integral term compensates for the continuous partial load; when the cargo shakes, the integral term cancels the cumulative effect of the cargo movement, and the differential term can suppress the instantaneous shake.

[0074] The differential coefficient determines the sensitivity of the control to the change rate of the pressure difference. The main roles of the differential term include two aspects: 1. Predict the future trend: anticipate the imbalance direction in advance through the change rate of the pressure difference, apply a reverse control force to suppress oscillations, and improve dynamic stability; 2. Suppress high-frequency noise: filter sensor noise or sudden shocks (such as hard braking) to avoid overshoot.

[0075] Through the oil pump pressure regulation function, rapid response (the proportional term provides immediate correction, shortening the adjustment time), precise steady state (the integral term eliminates long-term errors, ensuring balance accuracy) and disturbance rejection stability (the differential term suppresses sudden disturbances, avoiding overshoot or oscillation) can be achieved.

[0076] Through this function, the system can achieve high-dynamic-precision balance control, taking into account both the response speed and stability, and is applicable to complex scenarios such as heavy trucks.

[0077] In one embodiment, such as Figure 3As shown, the specific method for obtaining the oil pump pressure control instruction according to the oil pump pressure adjustment function includes the following steps:

[0078] S231, obtain the vehicle driving speed, acceleration, and road surface inclination angle;

[0079] S232, construct a neural network model, and use the pressure value, driving speed, acceleration, and road surface inclination angle as the inputs of the neural network model to obtain a dynamic correction factor;

[0080] S233, obtain the initial output pressure of the oil pump according to the oil pump pressure adjustment function, and correct the initial output pressure of the oil pump according to the dynamic correction factor to obtain the final output pressure of the oil pump;

[0081] S234, obtain the oil pump pressure control instruction according to the final output pressure of the oil pump.

[0082] Preferably, the dynamic correction factor ΔP = f NN (P1, P2, v, a, θ), and the final output pressure P' of the oil pump = P·(1 + ΔP); where f NN represents the neural network model, and v, a, and θ represent the driving speed, acceleration, and road surface inclination angle respectively.

[0083] Specifically, P1 and P2 respectively represent the pressure values measured in real time by the pressure sensors on the left and right sides of the carriage. The real-time pressure value P1 on the left side of the carriage represents the degree of left inclination of the goods, and the real-time pressure value P2 on the left side of the carriage represents the degree of right inclination of the goods. The unit of both is kPa. The driving speed affects the inertial moment and dynamic response time, and can be obtained through the CAN bus / wheel speed sensor. The acceleration (including positive and negative signs) reflects the inertial offset trend of the goods caused by sudden acceleration and deceleration, and can be obtained through the IMU inertial measurement unit. The road surface inclination angle (pitch angle) increases the front axle load when going uphill and the rear axle load when going downhill, and can be obtained through a high-precision gyroscope.

[0084] The dynamic correction factor ΔP = f NN (P1, P2, v, a, θ) is a neural network-driven non-linear compensation module for dynamically optimizing the oil pump pressure control accuracy. This neural network can be an improved BP neural network. Its input layer includes 5 nodes (corresponding to 5 input variables), which are normalized to [-1, 1]; the hidden layer is a double hidden layer (8 + 6 nodes), and the ReLU activation function with a leakage factor is used The output layer is 1 node (corresponding to ΔP), and the Sigmoid function is used to constrain the output range to [-0.5, 0.5].

[0085] When training a neural network, it can be trained with multiple sets (such as 10,000 sets, 50,000 sets, or 100,000 sets) of historical operating condition data stored in a big data platform. The loss function of the neural network can be set as the Huber loss to balance the advantages of MAE and MSE.

[0086] Preferably, the loss function L = Σ(α·(P target - P') 2 + β·E). Where α represents the pressure tracking weight, which is used to amplify the penalty for pressure deviation to ensure the rapid response of the oil pump pressure to the target value. A larger α preferentially guarantees the pressure tracking accuracy, which may increase energy consumption, while a smaller α allows a certain pressure deviation to reduce energy consumption. β represents the energy consumption penalty term, which is used to suppress energy waste during system operation, such as unnecessary high-pressure output of the oil pump or frequent displacement of the battery pack. A larger β strictly limits energy consumption, which may sacrifice part of the pressure response speed, while a smaller β allows higher energy consumption to improve the flexibility of pressure control.

[0087] P target represents the target output pressure of the oil pump, which can be understood as the ideal pressure value generated based on historical operating condition data or a simulation model, reflecting the theoretical pressure demand required for carriage balance, and can be adjusted in real time according to vehicle speed and load. For example, the pressure needs to be increased during a turn to suppress roll. E represents the energy consumption, including the energy consumption of the oil pump operation (related to the pressure adjustment amplitude and duration) and the energy consumption of the balance weight movement (mechanical energy consumption caused by the displacement of the battery pack slide rail). The optimization goal of energy consumption is to update the control parameters (such as the proportional coefficient) through the gradient descent method to minimize the total energy consumption while meeting the pressure demand.

[0088] α and β can be adaptively adjusted under different operating conditions. For example, when driving on a slope, α is increased to ensure continuous pressure compensation for the influence of gravity, while when cruising on a flat road, β is increased to reduce energy consumption. When the pressure difference increases instantaneously, the neural network predicts the roll trend and suppresses oscillations through the differential term, while α ensures rapid pressure response.

[0089] For α and β, they can be obtained through training with historical data. For example, the big data platform replays similar operating conditions (such as load, road conditions, etc.), and the optimal combination of α and β is selected through cross-validation.

[0090] The dynamic correction factor ΔP = f NN (P1, P2, v, a, θ) The function can predict the pressure imbalance trend within the next 2 - 3 seconds by fusing the real-time pressure difference, vehicle motion parameters, and environmental variables, and generate a compensation amount in advance to solve the limitations of the traditional linear formula P = k×(P1 - P2) under complex operating conditions. It uses a neural network to fit the non-linear mapping between the pressure difference and the correction amount. At the same time, it integrates dynamic parameters (speed, acceleration) and road parameters (tilt angle) to solve the distortion problem of a single pressure signal in scenarios such as turning and slopes.

[0091] The loss function \(L = \sum(\alpha\cdot(P target - P')) 2 + \beta\cdot E)\) realizes the following functions: 1. Minimize the deviation between the final output pressure \(P'\) and the target output pressure \(P\), target ensuring that the system response meets expectations; 2. Introduce an energy consumption penalty term \(\beta\cdot E\) to suppress unnecessary energy consumption (such as high-frequency adjustment of the oil pump or frequent movement of the battery pack); 3. By adjusting the weight coefficients \(\alpha\) and \(\beta\), find the optimal solution between pressure control accuracy and energy efficiency,

[0092] avoiding system performance imbalance caused by a single target. That is to say, this loss function realizes the dual goals of adaptive balance control and low-energy consumption operation,

[0093] and is applicable to scenarios with high requirements for energy efficiency and stability such as heavy trucks. Traditional PID control relies on linear formulas and is difficult to handle the complex non-linear relationship between the pressure difference and the vehicle motion state. In this embodiment, the neural network constructs a multi-layer non-linear mapping through activation functions (such as ReLU, Sigmoid) to capture hidden correlations,

[0094] such as the superposition effect of centrifugal force on the pressure difference during high-speed turning. In basic PID control, the integral term may cause adjustment lag, and the differential term is vulnerable to noise interference. The predicted dynamic correction factor \(\Delta P\) of the neural network anticipates the future pressure change trend,

[0095] compensates the adjustment amount in advance, and reduces overshoot and oscillation. In addition, the neural network predicts the pressure difference change trend based on real-time input parameters (left and right pressure differences, vehicle speed, acceleration, road surface inclination angle),

[0096] and outputs the dynamic correction factor \(\Delta P\). Through this mechanism, the system can perceive complex working conditions such as turning, slopes, and bumps in advance, Figure 4 and adjust the oil pump pressure to offset inertial or gravitational disturbances. An embodiment of the present invention also provides a self-balancing carriage control system for implementing the self-balancing carriage control method as Figure 5 shown, which includes a pressure sensor 3, a controller 4, an oil pump 5, and a balancing weight 6. The pressure sensors are respectively installed on both sides of the carriage 1 and are used to collect the pressure signals of the goods in the carriage on both sides of the carriage;

[0097] the controller is used to receive the pressure signals, process the pressure signals, and obtain an oil pump pressure control instruction; the oil pump is installed on the vehicle and is used to respond to the oil pump pressure control instruction, adjust the output pressure to push the balancing weight to move along the slide rail,

[0098] Wherein, the slide rail is installed at the bottom of the carriage along the vehicle width direction, and the balance weight is located at the bottom of the carriage and is slidably connected to the slide rail.

[0099] In order to better sense the cargo pressure when the carriage is tilted through the pressure sensor, partitions 2 can be respectively arranged on the left and right sides of the carriage. A movable compartment is formed between the partition 2 and the side plate of the carriage 1, that is, the partition is configured to sense the cargo pressure when the carriage is tilted so that the width of the compartment changes. The pressure sensor is installed on the partition and is located in the compartment. Here, the partitions can be respectively fixedly installed on the left and right sides of the carriage and are made of a material with a certain elastic deformation, such as a stainless steel plate or a plastic plate. The partitions can also be respectively slidably connected to the carriage, that is, a small section of chute is arranged on the left and right sides of the carriage along the vehicle width direction, and the partitions are installed in the chute.

[0100] When the vehicle is an electric vehicle, the balance weight is a battery pack. In this way, the carriage can be effectively balanced and the increase of additional burden can be avoided, so that while reducing the risk of roll, no additional mass and cost are added.

[0101] A moment balance equation M1 - M2 = m·g·(x' - x) combining the slide rail displacement of the balance weight and the carriage moment is provided in the controller. Wherein, M1 and M2 respectively represent the moments on the left and right sides of the carriage. When M1 > M2, the left side of the carriage is subjected to a greater moment and the balance weight needs to be moved to the right to balance. When M1 < M2, the right side of the carriage is subjected to a greater moment and the balance weight needs to be moved to the left.

[0102] m represents the mass of the balance weight. The greater the mass, the stronger the moment adjustment ability of the balance weight. g represents the acceleration due to gravity. x' represents the theoretical balance point position of the balance weight, which is used as the displacement reference benchmark and is preferably the midpoint position of the slide rail here. When x = x', the moment difference is zero and the carriage is in a balanced state. x represents the real-time displacement position of the balance weight on the slide rail. By adjusting the value of x, the position of the balance weight relative to x' is changed, thereby adjusting the moment difference.

[0103] When x < x', the balance weight moves to the left, the left moment increases, resulting in M1 - M2 > 0; when x > x', the balance weight moves to the right, the right moment increases, resulting in M1 - M2 < 0.

[0104] By driving the balance weight to move on the slide rail to adjust x, finally making |M1 - M2| < ε', where ε' represents the preset moment tolerance threshold.

[0105] As a preferred technical solution, the controller obtains the initial output pressure P of the oil pump according to the formula ; wherein, where K p 、K i 、K drespectively represent the proportional coefficient, integral coefficient, and derivative coefficient of the PID controller, and P1 and P2 respectively represent the pressure values measured in real time by the pressure sensors on the left and right sides of the carriage.

[0106] Preferably, the proportional coefficient K p = K0·(1 + Sigmoid(ΔP)). Wherein, K0 represents the basic proportional coefficient, which is obtained by experimental calibration from the carriage structure parameters (such as the center of gravity position, slide rail length) and the mass of the balance weight, and reflects the control response intensity of the system under static or conventional working conditions. For example, when the carriage load is large, the value of K0 will increase to match the higher torque requirement.

[0107] is used to map the input to the (0, 1) interval to avoid system oscillation caused by sudden changes in the proportional coefficient, such as preventing high-frequency fluctuations in the oil pump pressure under bumpy roads; and ensuring that the corrected proportional coefficient K p is within the range of (K0, 2K0), taking into account both sensitivity and stability.

[0108] The proportional coefficient K p = K0·(1 + Sigmoid(ΔP))'s dynamic correction mechanism includes positive correction and negative correction.

[0109] In positive correction, when ΔP > 0 (such as predicting a sharp turn or cargo shift), the Sigmoid output approaches 1, and K p ≈ 2K0, the system response speed increases, quickly offsetting the pressure difference. Example: When a heavy truck starts on a slope, the neural network predicts the backward trend based on the acceleration and road surface inclination angle, increasing the proportional coefficient K p to enhance the oil pump adjustment force.

[0110] In negative correction, when ΔP < 0 (such as predicting a flat road surface or uniform driving), the Sigmoid output approaches 0, and K p ≈ K0, the system returns to the basic control mode to reduce energy consumption. Example: When the vehicle is driving straight, the neural network recognizes the steady-state working condition and reduces K p to avoid unnecessary displacement of the balance weight on the slide rail.

[0111] This proportional coefficient K p = K0·(1 + Sigmoid(ΔP)) function combines the steady-state characteristics of traditional PID control with the dynamic prediction advantages of neural networks, solving the limitations of fixed proportional coefficients under complex working conditions; suppressing the interference of high-frequency noise (such as sensor errors or road surface vibrations) on the proportional coefficient through the Sigmoid function, automatically reducing the control intensity under low-risk working conditions, and reducing the energy consumption of the oil pump.

[0112] Through the oil pump pressure regulation function It can achieve fast response (the proportional term provides immediate correction, shortening the adjustment time), precise steady state (the integral term eliminates long-term errors, ensuring balance accuracy), and disturbance rejection stability (the derivative term suppresses sudden disturbances, avoiding overshoot or oscillation). Through this oil pump pressure regulation function, the system can achieve high-dynamic-precision balance control, taking into account both the response speed and stability, and is applicable to complex scenarios such as heavy trucks.

[0113] The controller obtains a dynamic correction factor according to the formula ΔP = f NN (P1, P2, v, a, θ), corrects the initial output pressure of the oil pump according to the dynamic correction factor to obtain the final output pressure of the oil pump, and obtains an oil pump pressure control instruction according to the final output pressure of the oil pump; where f NN represents a neural network model, and v, a, and θ respectively represent the driving speed, acceleration, and road surface inclination angle.

[0114] The controller obtains the final output pressure P' of the oil pump according to the formula P' = P·(1 + ΔP).

[0115] Traditional PID control relies on linear formulas and is difficult to handle the complex non-linear relationship between the pressure difference and the vehicle motion state. In this embodiment, the neural network constructs a multi-layer non-linear mapping through activation functions (such as ReLU, Sigmoid) to capture hidden correlations, such as the superposition effect of centrifugal force on the pressure difference during high-speed turning.

[0116] In basic PID control, the integral term may cause adjustment lag, and the derivative term is vulnerable to noise interference. The predicted dynamic correction factor of the neural network anticipates the future pressure change trend, compensates the adjustment amount in advance, and reduces overshoot and oscillation.

[0117] In addition, the neural network predicts the pressure difference change trend based on real-time input parameters (left and right pressure difference, vehicle speed, acceleration, road surface inclination angle) and outputs a dynamic correction factor. Through this mechanism, the system can perceive complex working conditions such as turning, slopes, and bumps in advance, and adjust the oil pump pressure to offset inertial or gravitational disturbances.

[0118] In summary, the control system makes the balance control system adaptive by integrating real-time data and prediction capabilities, and significantly improves the stability and energy efficiency ratio under complex working conditions.

[0119] The technical features of the above-described embodiments can be combined arbitrarily. For the sake of brevity of description, not all possible combinations of the technical features in the above embodiments are described. However, as long as these technical feature combinations do not conflict, they should be considered as the scope recorded in this specification.

[0120] The above-described embodiments merely represent several implementation manners of the present invention. The description is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all fall within the protection scope of the present invention. Therefore, the protection scope of the present invention patent shall be subject to the appended claims.

Claims

1. A self-balancing carriage control method, characterized in that, The self-balancing carriage control method includes the following steps: Collect the pressure signals of the goods in the carriage on both sides of the carriage through the pressure sensors installed on both sides of the carriage; Process the pressure signals to obtain the oil pump pressure control command; Respond to the oil pump pressure control command, adjust the output pressure of the oil pump to push the balance weight to move along the slide rail, and adjust the moments on both sides of the carriage, so as to maintain the balance of the carriage.

2. The self-balancing carriage control method according to claim 1, characterized in that, The specific method of processing the pressure signals to obtain the oil pump pressure control command includes the following steps: Process the pressure signals to obtain the pressure values measured in real time by the pressure sensors on the left and right sides of the carriage; Define an oil pump pressure adjustment function based on the direct mapping relationship between the pressure values and the output pressure of the oil pump; Obtain the oil pump pressure control command according to the oil pump pressure adjustment function.

3. The self - balancing carriage control method according to claim 2, wherein, The specific method of obtaining the oil pump pressure control command according to the oil pump pressure adjustment function includes the following steps: Obtain the vehicle driving speed, acceleration and road surface inclination angle; Construct a neural network model, and use the pressure values, driving speed, acceleration and road surface inclination angle as the inputs of the neural network model to obtain a dynamic correction factor; Obtain the initial output pressure of the oil pump according to the oil pump pressure adjustment function, and correct the initial output pressure of the oil pump according to the dynamic correction factor to obtain the final output pressure of the oil pump; Obtain the oil pump pressure control command according to the final output pressure of the oil pump.

4. The self-balancing carriage control method according to claim 3, wherein The oil pump pressure regulation function is expressed as Among them, K p , K i , K d respectively represent the proportional coefficient, integral coefficient, and differential coefficient of the PID controller, P1 and P2 respectively represent the pressure values measured in real time by the pressure sensors on the left and right sides of the carriage, and P represents the initial output pressure of the oil pump.

5. The self-balancing carriage control method according to claim 4, characterized in that, The dynamic correction factor ΔP = f NN (P1, P2, v, a, θ); where f NN represents the neural network model, and v, a, and θ represent the driving speed, acceleration, and road surface inclination angle, respectively.

6. The self-balancing carriage control method according to claim 5, wherein The final output pressure P' of the oil pump = P·(1 + ΔP).

7. A self-balancing carriage control system for implementing the self-balancing carriage control method according to any one of claims 1-6, characterized in that, It includes: Pressure sensors, which are respectively installed on both sides of the carriage and are used to collect the pressure signals of the goods in the carriage on both sides of the carriage; A controller, which is used to receive the pressure signals, process the pressure signals, and obtain the oil pump pressure control command; An oil pump, which is installed on the vehicle and is used to respond to the oil pump pressure control command, adjust the output pressure to push the balance weight to move along the slide rail, and adjust the moments on both sides of the carriage, so as to maintain the balance of the carriage; Wherein, the slide rail is installed at the bottom of the carriage along the vehicle width direction, and the balance weight is located at the bottom of the carriage and is slidably connected to the slide rail.

8. The self-balancing carriage control system according to claim 7, wherein, The controller obtains the initial output pressure P of the oil pump according to the formula ​ Among them, K p , K i , K d respectively represent the proportional coefficient, integral coefficient, and differential coefficient of the PID controller, and P1 and P2 respectively represent the pressure values measured in real time by the pressure sensors on the left and right sides of the carriage.

9. The self-balancing carriage control system according to claim 8, wherein, The controller obtains a dynamic correction factor according to the formula ΔP = f NN (P1, P2, v, a, θ), corrects the initial output pressure of the oil pump according to the dynamic correction factor to obtain the final output pressure of the oil pump, and obtains an oil pump pressure control command according to the final output pressure of the oil pump; Among them, f NN represents a neural network model, and v, a, and θ represent the driving speed, acceleration, and road surface inclination angle respectively.

10. A self-balancing carriage control system according to claim 9, characterized in that, The controller obtains the final output pressure P' of the oil pump according to the formula P' = P·(1 + ΔP).

Images