Apparatus and method for generating control input, and vehicle control system comprising same

Abstract

The present disclosure relates to an apparatus and a method for generating a control input, and a vehicle control system comprising same. Particularly, the apparatus for generating a control input, according to the disclosure, can generate a control input having robustness against a disturbance in integrated control for a vehicle.

Description

Control input generating device and method and vehicle control system including the same

[0001] The present embodiments relate to an apparatus and method for generating control inputs for controlling a plurality of actuators for following a target path of a vehicle.

[0002] In the field of automotive technology, various electronic control units are being applied to improve driving safety and ride comfort. In particular, with the advancement of autonomous vehicles and Advanced Driver Assistance Systems (ADAS), there is an increasing demand for technologies to precisely control vehicle behavior.

[0003] Generally, vehicle behavior is determined by the operation of multiple actuators, such as steering, braking, driving, and suspension systems. Conventionally, path-following technologies centered on single-actuator control, such as front-wheel steering, have been primarily applied. However, single-actuator control methods are known to have limitations in responding to disturbances or complex driving conditions that occur in various driving environments.

[0004] Accordingly, integrated control technology that considers multiple actuators simultaneously has been studied. While integrated control technology can improve vehicle stability by controlling actuators in conjunction, it is difficult to secure optimal performance due to differences in physical characteristics between actuators, response delays, and the influence of disturbances. Furthermore, external factors such as changes in road surface conditions and crosswinds during driving, as well as signal delays in electronic control systems, are pointed out as problems that can lead to a decrease in vehicle control stability.

[0005] Therefore, a new vehicle control technology is required that can effectively control multiple actuators while robustly responding to various disturbances and system delays.

[0006] Against this backdrop, the present disclosure aims to provide a control input generating device and method having robustness against disturbances in the integrated control of a vehicle.

[0007] In order to solve the aforementioned problem, in one aspect, the present disclosure may provide a control input generating device comprising a memory for storing at least one instruction and at least one processor for executing said at least one instruction, wherein the at least one processor, by executing said at least one instruction, obtains reference state information corresponding to a driving target path of a vehicle and obtains current state information regarding the current driving state of the vehicle, generates predicted state information based on said current state information, generates error information based on the difference between said predicted state information and said reference state information, and generates a control input for following said target path based on an objective function including said error information.

[0008] In another aspect, the present disclosure may provide a method for generating a control input, comprising the steps of: obtaining reference state information corresponding to a driving target path of a vehicle; obtaining current state information regarding the current driving state of the vehicle; generating predicted state information based on the current state information; generating error information based on the difference between the predicted state information and the reference state information; and generating a control input for following the target path based on an objective function including the error information.

[0009] In another aspect, the present disclosure may provide a vehicle control system comprising: a reference state generating device that generates reference state information corresponding to a target driving path of a vehicle; an observer that calculates current state information by estimating the actual driving state of a vehicle based on sensor data of the vehicle; a plurality of actuators for driving, braking, steering, and suspension control; and a control input generating device that generates control inputs for controlling the operation of the plurality of actuators. The control input generating device obtains the reference state information from the reference state generating device and obtains the current state information from the observer, generates predicted state information based on the current state information, generates error information based on the difference between the predicted state information and the reference state information, and generates control inputs for target path following based on an objective function including the error information.

[0010] According to the embodiments, a control input generating device and method robust to disturbances during integrated control of a plurality of actuators including steering, braking, driving, and suspension can be provided.

[0011] FIG. 1 is a block diagram of an exemplary computing system that can be used in the present disclosure.

[0012] Figure 2 is a diagram showing a typical vehicle integrated control system.

[0013] FIG. 3 is an exemplary flowchart illustrating a control input generation process according to one embodiment of the present disclosure.

[0014] FIG. 4 is an exemplary flowchart illustrating a process for generating predicted state information according to one embodiment of the present disclosure.

[0015] FIG. 5 is an exemplary flowchart illustrating a process for calculating a control input based on an objective function according to one embodiment of the present disclosure.

[0016] FIG. 6 is an exemplary flowchart illustrating the distribution process of an integrated control input according to an embodiment of the present disclosure.

[0017] FIG. 7 is an exemplary flowchart illustrating a control input update process according to one embodiment of the present disclosure.

[0018] FIG. 8 is an exemplary flowchart illustrating a method for generating a control input according to an embodiment of the present disclosure.

[0019] Hereinafter, some embodiments of the present disclosure will be described in detail with reference to the exemplary drawings. In assigning reference numerals to the components of each drawing, the same components may have the same reference numeral as much as possible, even if they are shown in different drawings. Furthermore, in describing the embodiments, if it is determined that a detailed description of related known components or functions may obscure the essence of the technical concept, such detailed description may be omitted. Where terms such as "comprising," "having," or "consisting of" are used in this specification, other parts may be added unless "only" is used. Where a component is expressed in the singular, it may include a plural unless otherwise specified.

[0020] Additionally, terms such as first, second, A, B, (a), (b), etc., may be used to describe the components of the present disclosure. These terms are used merely to distinguish the components from other components, and the nature, order, sequence, or number of the components are not limited by such terms.

[0021] In describing the positional relationship of components, where it is stated that two or more components are "connected," "combined," or "joined," it should be understood that while the two or more components may be directly "connected," "combined," or "joined," they may also be "connected," "combined," or "joined" with other components "intervened." Here, the other components may be included in one or more of the two or more components that are "connected," "combined," or "joined" with one another.

[0022] In describing the temporal flow relationship regarding components, methods of operation, or methods of production, for example, when the temporal or sequential relationship is described using "after," "following," "next," or "before," it may include cases where the relationship is not continuous unless "immediately" or "directly" is used.

[0023] Meanwhile, where numerical values ​​or corresponding information regarding a component (e.g., levels, etc.) are mentioned, even without separate explicit notation, the numerical values ​​or corresponding information may be interpreted as including a range of error that may occur due to various factors (e.g., process factors, internal or external shocks, noise, etc.).

[0024] FIG. 1 is a block diagram of an exemplary computing system that can be used in the present disclosure.

[0025] Referring to FIG. 1, the computing system may be configured to include one or more processors (110), memory (120), storage unit (130), user interface input unit (140), user interface output unit (150), bus (160), and network interface (170) in a form that reflects the hardware configuration of the control input generating device (100).

[0026] The processor (110) is a semiconductor device that performs a central processing unit (CPU) or other computational functions, and can execute instructions stored in memory (120) and / or storage unit (130) to obtain reference state information and current state information of the vehicle, generate predicted state information based on a prediction model, calculate error information between the predicted state information and the reference state information, and perform operations such as optimizing an objective function based on the error information to produce a control input.

[0027] The memory (120) may include volatile memory and non-volatile memory, and may include, for example, ROM (Read Only Memory) (124) and RAM (Random Access Memory) (125). The ROM (124) may store initial parameters of a prediction model, state matrix configuration values, objective function setting values, weight matrix setting values, etc., and the RAM (125) may be used to store temporary data such as the current state of the vehicle, reference state, error value, and objective function result value.

[0028] The storage unit (130) may include various non-volatile storage media such as HDD (Hard Disk Drive), SSD (Solid State Drive), and eMMC (embedded Multi Media Card), and may be used to store, for example, control history data, disturbance condition logs, actuator response characteristics, and integrated control input distribution results.

[0029] The user interface input section (140) is a component for system settings, diagnosis, and receiving user commands, and may include a touchscreen, button switch, external input terminal, etc., and the user interface output section (150) is a component for displaying the integrated control status, the result of applying the control input, the vehicle stability status, etc., and may be implemented as a display, LED, warning sound module, etc.

[0030] The bus (160) is an internal data path for transmitting and receiving data between the above components, allowing each component to be organically interconnected.

[0031] Additionally, the computing system may include a network interface (170) for linkage with an in-vehicle CAN communication network or other vehicle networks. The network interface (170) is a communication module for the control input generating device (100) to exchange data with a plurality of electronic control units (ECUs) that perform steering, braking, driving, and suspension control, and can collect driving state information, sensor data, and actuator feedback signals from each ECU. Additionally, the network interface (170) can support the real-time reflection of the integrated control results based on predictive control in the entire vehicle control system by transmitting the integrated control input calculated by the processor (110) to each ECU. According to one embodiment of the present disclosure, the control input generating device (100) may be configured to calculate the integrated control input based on predictive control so as to respond robustly to various disturbances and actuator response delays that may occur in the chassis integrated control environment of the vehicle.

[0032] Generally, in a vehicle chassis integrated control environment, multiple actuators such as steering, braking, driving, and suspension operate simultaneously, and the control commands of each actuator influence each other. During this interaction process, a change in the response characteristics or a control delay of one actuator affects the control stability of other actuators, and as a result, unexpected disturbances may occur in the overall vehicle behavior.

[0033] For example, when steering control is performed simultaneously with braking control, changes in braking force cause fluctuations in road grip, which can lead to mutual interference such as variations in steering torque or yaw rate response. Furthermore, there is a possibility that the stability of the system may be compromised as errors accumulate between target and actual values ​​due to response delays or control history of each actuator.

[0034] In other words, if the interference and delay characteristics between multiple actuators are not considered, it may be difficult to secure robust control performance against disturbances using an individual control method. That is to say, if the control loop for each actuator is configured independently, it is difficult to perform real-time compensation for complex disturbances or delays caused by interference between actuators.

[0035] The control input generating device (100) of the present disclosure analyzes the dynamic characteristics of the entire vehicle in a single integrated control frame and simultaneously calculates control inputs corresponding to a plurality of actuators through a predictive control-based optimization process, thereby suppressing instability caused by interaction in advance.

[0036] Specifically, the control input generating device (100) can acquire reference state information corresponding to the driving goal of the vehicle and acquire current state information based on the driving state measured through a sensor. The control input generating device (100) uses the current state information as an input value to mathematically model the dynamic characteristics of the vehicle according to the predictive control principle to predict the future driving state, and calculates error information based on the difference between the predicted state information and the reference state information. The calculated error information is defined as the difference between the reference state information and the predicted state information and can represent the amount of correction required for the vehicle to follow the target path.

[0037] Additionally, the control input generating device (100) defines an objective function including calculated error information and optimizes the control input so that the objective function is minimized. The objective function includes an error term, a control input term, and a disturbance term, which can be set as terms to minimize the vehicle's tracking error, suppress excessive changes in the control input, and compensate for the influence of the disturbance, respectively. In particular, the control input generating device (100) calculates the control input to minimize the value of the objective function under conditions where the disturbance acts to the maximum extent, thereby ensuring stable and robust control performance even if there is a disturbance or a response delay of the actuator.

[0038] Additionally, the control input generating device (100) can simultaneously apply the calculated control input to a plurality of actuators corresponding to steering, braking, driving, and suspension control, and distribute the integrated control input by dynamically adjusting the control weight or weight of each actuator according to the vehicle's driving state and environmental changes.

[0039] At this time, the response characteristics or control limits of each actuator are taken into consideration, and the integrated control input can be configured to be optimized in a way that maintains the behavioral stability of the entire vehicle.

[0040] Additionally, the control input generating device (100) obtains the driving state after the calculated control input is applied in real time and repeatedly calculates the control input by updating the predicted state information and error information based on the driving state. Through this, a closed-loop control structure can be implemented in which real-time compensation is made for disturbances occurring during driving or delay characteristics of the actuator.

[0041] Through such a closed-loop control structure, the control input generating device (100) can stably follow the target path even when there is a sudden change in the driving environment or an unexpected disturbance occurs, and can improve long-term control reliability and driving safety.

[0042] A computing system according to the present disclosure may be connected to a server system. For example, a server system may be connected to the computing system according to the present disclosure via a network, either wired or wirelessly, to transmit and receive data and share computing resources.

[0043] For example, a server system can be built in the form of a cloud system. For instance, the server system may include a configuration that allows individual computing devices to connect to the server system via a network and shares computing resources with the connected devices. In this case, individual computing devices can access the cloud system from anywhere as long as they are connected to a network, such as the Internet.

[0044] For example, a cloud system can provide computing resources by flexibly scaling them up or down as needed, and can share these resources with other computing devices connected via a network. Furthermore, depending on the purpose or scope of use, a cloud system can be built based on various service models such as IaaS (Infrastructure as a Service), PaaS (Platform as a Service), and SaaS (Software as a Service).

[0045] For example, a cloud system may include at least one computing device, a storage device, and a network device. Each computing device included in the cloud system may include a processor and memory to enable processing various computing tasks, the storage device may include configurations related to data storage such as an HDD (Hard Disk Drive), SSD (Solid State Drive), NAS (Network Attached Storage), or SAN (Storage Area Network) for storing large amounts of data, and the network device may include configurations related to networking such as a switch, a router, a load balancer, and a firewall.

[0046] For example, when a computing system performs various calculations or data classification and generation tasks using an artificial intelligence model, it may utilize an artificial intelligence model stored in the computing system's memory.

[0047] Alternatively, in some cases, a computing system may share computing resources from a cloud system and utilize artificial intelligence models stored in the cloud system. In this case, the computing system can process related tasks by using the artificial intelligence models provided by the cloud system, even if it does not directly possess the configurations related to the artificial intelligence models itself.

[0048] The processes, systems, and methods described herein may be implemented by a computing system in response to a processor executing an array of instructions contained in main memory. These instructions may be read into main memory from other computer-readable media, such as storage devices. The execution of the array of instructions contained in main memory causes the computing system to perform the exemplary processes described herein. In a multiprocessing array, one or more processors may also be used to execute instructions contained in main memory. Hardwired circuits may be used in place of or in conjunction with hardware instructions with the systems and methods described herein. The systems and methods described herein are not limited to any specific combination of hardware circuits and software.

[0049] Although exemplary computing systems have been described above, the essence including the operations described herein may be implemented in other types of digital electronic circuits, or in computer software, firmware, or hardware including structures disclosed herein and structural equivalents thereof or combinations of one or more of these.

[0050] "Data processing system," "computing device," "module," "engine," "component," or "computing device" includes various devices, devices, and machines for processing data, including, for example, a programmable processor, a computer, a system on a chip, or a number of such things, or a combination thereof.

[0051] The processes and logic flows described herein may be performed by one or more programmable processors that execute one or more computer programs (e.g., components of a data processing system) to perform actions by operating input data and generating outputs. The processes and logic flows may also be performed by special-purpose logic circuits, e.g., FPGAs, ASICs, DSPs, DSPDs, or PLDs, and devices may also be implemented by special-purpose logic circuits. Devices suitable for storing computer program instructions and data include all forms of non-volatile memory, media, and memory devices, such as semiconductor memory devices like EPROM, EEPROM, and flash memory devices; magnetic disks, e.g., internal hard disks or removable disks; optomagnetic disks; and CD-ROM and DVD-ROM disks. Processors and memory may be complemented or integrated by special-purpose logic circuits.

[0052] Figure 2 is a drawing showing a general vehicle integrated control system (200).

[0053] Referring to FIG. 2, the vehicle integrated control system (200) may be configured to include an upper-level ADAS controller, a chassis integrated controller, and a plurality of chassis system controllers (e.g., a drive controller, a braking controller, a steering controller, a suspension controller). Here, the ADAS controller may be an upper-level control device that recognizes the vehicle driving environment and determines a driving goal.

[0054] For example, an ADAS controller can recognize the driving environment around the vehicle by fusing data input from external sensors such as cameras, radar, and lidar, as well as in-vehicle state sensors, and calculate driving goals related to the driving path, target speed, distance between vehicles, and lane keeping based on information such as lanes, vehicles, pedestrians, and road curvature.

[0055] In other words, the ADAS controller sets higher-level driving objectives, such as the driving path, target speed, and target yaw rate, and transmits corresponding reference state information to the chassis integrated controller. Based on the transmitted reference state information and current driving state information obtained from in-vehicle sensors, the chassis integrated controller calculates integrated control commands that consider the interactions between multiple control systems and distributes them to each system controller. Each system controller can perform functions such as vehicle acceleration, deceleration, direction change, and attitude stabilization by driving actuators corresponding to each domain (e.g., steering, braking, driving, suspension).

[0056] In a typical integrated control system, multiple actuators are controlled cooperatively to manage the vehicle's overall behavior in an integrated manner. However, in actual driving environments, disturbances such as changes in road surface friction, crosswinds, tire slip, and braking force deviations occur constantly, and the response delay characteristics of each actuator may vary. Consequently, in the case of integrated control that does not sufficiently consider these disturbances and delay characteristics, response imbalances or control deviations occurring in specific systems may affect the control stability of other systems, potentially destabilizing the dynamic characteristics of the entire system. Therefore, a control structure that considers the disturbances and delay characteristics of the actual vehicle environment is desirable.

[0057] The control input generating device (100) according to the present disclosure may be applied in a configuration corresponding to the computational unit of the chassis integrated controller shown in FIG. 2. The control input generating device (100) receives reference state information corresponding to a driving goal provided by a higher-level controller (e.g., ADAS controller) and sensor data of the vehicle, reflects actual driving conditions that may include disturbances and response delays into a prediction model, and calculates an integrated control input corresponding thereto.

[0058] According to one embodiment, a control input generating device (100) receives reference state information corresponding to a driving goal from a higher-level controller and collects current driving state information through a vehicle sensor. The control input generating device (100) inputs the collected current driving state information into a prediction model to predict the future driving state of the vehicle, and during the prediction process, evaluates the dynamic characteristics of the system by considering the influence of disturbances such as magnitude, direction of action, and timing of occurrence.

[0059] The control input generating device (100) configures an objective function to minimize the impact of disturbance and response delay characteristics on the stability of the vehicle, and optimizes the control input based on the objective function. That is, the control input generating device (100) does not simply calculate an input to correct the error at the current time, but rather adjusts the control input to maintain the stability of the vehicle by reflecting in advance the impact of disturbances on the vehicle's behavior through a prediction model.

[0060] To this end, the control input generating device (100) can define an objective function that includes error information between the predicted state information and the reference state information calculated from the prediction model, and calculate the control input by performing an optimization operation so that the value of the objective function is minimized.

[0061] Therefore, control inputs are automatically compensated so that the overall stability of the system can be maintained even if the magnitude or direction of disturbance changes, and robust control performance can be secured under various driving conditions.

[0062] Hereinafter, with reference to FIGS. 3 to 7, a specific operation procedure of the control input generation process performed by the processor of the control input generation device according to the present disclosure will be described in detail.

[0063]

[0064] FIG. 3 is an exemplary flowchart illustrating a control input generation process (300) according to one embodiment of the present disclosure.

[0065] Referring to FIG. 3, the processor can obtain reference state information corresponding to the vehicle's driving target path (S310).

[0066] The vehicle's driving target path may refer to a reference trajectory that includes information such as the target location the vehicle must reach while driving, the direction of travel, speed, and acceleration.

[0067] According to one embodiment, reference state information is a set of state variables that mathematically represent the driving target path of a vehicle, and may include reference values ​​regarding target position, speed, acceleration, yaw rate, lateral acceleration, lateral velocity, steering angle, and vehicle posture. For example, reference state information is generated according to a driving goal set by a higher-level controller (e.g., an ADAS controller) and can be used as data defining a reference state that the vehicle must achieve at a specific time or interval while driving.

[0068] In one embodiment, reference state information may be obtained through a reference state generating device. The reference state generating device is a device that calculates a target state that the vehicle must follow based on the vehicle's driving environment and path planning results, and may be, for example, an ADAS controller. As described in FIG. 2, the ADAS controller can recognize the vehicle's location and surrounding environment by utilizing external sensors (e.g., camera, lidar, radar, GPS, etc.) and map data, and determine a target trajectory that the vehicle must drive through a path planning algorithm.

[0069] In another embodiment, the processor may receive a driving target path from an external communication device, a navigation system, or a cloud server, and generate reference state information itself based thereon.

[0070] In this case, the processor can analyze information such as the target location, direction of travel, speed, and acceleration included in the received driving target path to define the target driving state of the vehicle and set the target driving state as reference state information.

[0071] For example, the processor can generate reference state information necessary for vehicle control without the intervention of an external control device by calculating the target yaw rate, lateral acceleration, and attitude angle, etc., by taking into account driving environment information such as the vehicle's driving scenario, road curvature, gradient, and traffic conditions.

[0072] In addition, in an embodiment, the processor can obtain current state information regarding the current driving state of the vehicle (S320).

[0073] In one embodiment, current state information is a dataset of physical state variables measured in real time while the vehicle is in motion, and may include the vehicle's position, speed, acceleration, yaw rate, lateral acceleration, steering angle, wheel speed, vehicle attitude, and road surface friction estimate. The current state information may be obtained through sensors mounted on the vehicle (e.g., IMU, steering angle sensor, wheel speed sensor, GPS sensor, etc.) or an ECU within the vehicle, and may be input into a prediction model to be used for predicting future driving states (i.e., predicted state information).

[0074] In addition, in an embodiment, the processor can generate predicted state information (S330) based on current state information.

[0075] According to one embodiment, predicted state information may refer to information that estimates the vehicle's behavior at a specific point in the future based on the vehicle's current driving state. For example, the predicted state information may be calculated by estimating how state variables, such as speed, acceleration, yaw rate, and attitude angle observed at the current point in time, will change over time.

[0076] In one embodiment, the processor can generate predicted state information corresponding to current state information by utilizing a prediction model.

[0077] According to one embodiment, the prediction model is a mathematical model that reflects the dynamic characteristics of a vehicle and can calculate changes in state variables over time based on the vehicle's state equation. For example, the prediction model may be defined in the form of a continuity equation or a discrete equation representing the dynamic characteristics of the vehicle, and can receive current state information (e.g., position, velocity, acceleration, yaw rate, steering angle, etc.) as input and calculate a predicted state after a certain period of time (e.g., predicted velocity, predicted attitude angle, predicted lateral acceleration, etc.).

[0078] The processor can generate predicted state information corresponding to the vehicle's future driving state by applying current state information as an input variable to the prediction model and calculating the state change at the next time point using parameters such as the state matrix and input matrix inherent in the model.

[0079] In various embodiments, the processor can calculate predicted state information by considering the vehicle's current driving state (i.e., current state information), the vehicle's control input, and disturbance factors together.

[0080] Here, the control input is a control command provided from a chassis integrated controller or each system controller (e.g., steering controller, braking controller, drive controller), and may include control signals such as steering angle, braking torque, and drive torque. The disturbance is an influence such as crosswind, changes in road surface friction, tire slip, and road surface irregularities occurring in the external environment or driving conditions of the vehicle, and may be detected or estimated through sensors or an environment estimation model.

[0081] More specifically, the prediction model can be configured to receive the vehicle's current state vector (x), control input (u), and disturbance term (d) as inputs, and to calculate the state at the next time point through the system matrix (A, B, W).

[0082] The processor can calculate linear or non-linear changes in the current state vector (x) through the state matrix (A) in the prediction model, reflect the influence of the control input (u) through the input matrix (B), and consider the influence of the disturbance term (d) on the system through the disturbance matrix (W).

[0083] Accordingly, the prediction model can estimate the future state step by step through a computational procedure of the form (state at the next time point) = (state matrix × current state) + (input matrix × control input) + (disturbance matrix × disturbance term).

[0084] For example, when a vehicle receives a constant steering angle input and a disturbance such as a crosswind acts upon it, the prediction model can simultaneously calculate the change in yaw rate caused by the steering input (u) and the change in lateral acceleration induced by the disturbance (d) to predict the vehicle attitude angle, yaw rate, speed, etc. at a future point in time.

[0085] In other words, by simultaneously considering the dynamic response to the vehicle's control input and changes in behavior caused by disturbances, the prediction model can quantitatively estimate possible state changes that may occur in actual driving situations.

[0086] According to one embodiment, the prediction model can be extended to include derived items such as the rate of change of speed, the rate of change of attitude angle, and the rate of change of lateral acceleration based on the kinematic relationship of the vehicle. For example, the rate of change of the vehicle's yaw rate can be predicted through a combination of variables such as vehicle speed, steering angle, and lateral acceleration, and the rate of change of the vehicle's attitude angle can be calculated as the result of yaw rate integration.

[0087] In addition, the prediction model is sampled at constant discrete time units (Δt), and the processor can generate a future state sequence for a continuous driving section by iteratively predicting the next state based on the current state vector, control input, and disturbance term for each period.

[0088] According to one embodiment of the present disclosure, a processor can reflect the actual dynamic characteristics of a vehicle in a model by determining or correcting the state matrix of a prediction model based on vehicle-specific data.

[0089] Here, vehicle-specific information may include physical parameters that differ from vehicle to vehicle, such as vehicle mass, wheelbase, tread, tire cornering stiffness, yaw moment inertia, suspension coefficient, and road surface friction coefficient.

[0090] The processor can obtain vehicle-specific information from a database learned or estimated during vehicle manufacturing or driving, and set or update the values ​​of the state matrix (A), input matrix (B), and disturbance matrix (W) within the prediction model.

[0091] In other words, the processor is configured so that the prediction model accurately reflects the actual behavioral characteristics of each vehicle through a state matrix that incorporates vehicle-specific information, thereby enabling stable state prediction even when disturbance conditions or changes in the driving environment occur.

[0092] FIG. 4 is an exemplary flowchart illustrating a process (400) for generating predicted state information according to one embodiment of the present disclosure.

[0093] Referring to FIG. 4, the processor can obtain current state information regarding the current driving state of the vehicle (S410).

[0094] Current status information is data collected from multiple sensors mounted on the vehicle, and may include measured values ​​such as the vehicle's speed, acceleration, yaw rate, attitude angle, steering angle, lateral acceleration, and lateral slip rate.

[0095] In addition, the processor can apply a state matrix determined based on vehicle unique information to the current state information to reflect vehicle dynamic characteristics in the prediction model (S420).

[0096] According to one embodiment, the processor can preprocess the acquired vehicle unique information to match the input format of the prediction model.

[0097] The preprocessing process may include unit conversion, scaling, normalization, and parameter mapping. For example, parameters with different unit measurement methods, such as tire stiffness values ​​or suspension coefficients, can be converted to fit the dimensions of the state matrix used by the model, and items with large differences in value magnitude, such as mass or inertia coefficients, can be adjusted using learned normalization factors to maintain their relative weight.

[0098] The processor can configure the prediction model to reflect the actual physical characteristics of the vehicle by using preprocessed vehicle-specific information to set or update the coefficients of the state matrix, input matrix, and disturbance matrix within the prediction model.

[0099] For example, even when the same steering angle input is given, if the vehicle's mass is large, the actual yaw rate response appears small and the response delay is longer.

[0100] The processor can increase the inertia term of the state matrix to reflect mass differences between vehicles and correct the damping coefficient of the input matrix for vehicles with sensitive braking systems, thereby adjusting the model response to match the actual vehicle response.

[0101] In addition, for vehicles with low tire stiffness or road surface environments with low friction, the processor can reduce the friction-related coefficients of the disturbance matrix to enhance the prediction sensitivity to disturbances.

[0102] The matrix correction procedure described above enables the prediction model to reflect differences in dynamic characteristics by vehicle entity and environment, thereby significantly reducing the difference between the vehicle's actual response and the model's predicted response.

[0103] Additionally, the processor can generate predicted state information corresponding to the current state information (S430) by utilizing a prediction model that includes a corrected state matrix.

[0104] The processor applies current state information, control inputs, and disturbance terms as inputs to the prediction model, and can calculate the driving state at the next time point based on the relationship between state variables within the prediction model.

[0105] According to one embodiment, in the case of a general prediction model that does not reflect vehicle-specific information, the same state matrix is ​​applied to all vehicles, so a problem may occur in which an error accumulates between the actual response and the predicted response in vehicles with different mass, stiffness, and friction characteristics.

[0106] Meanwhile, since the processor according to one embodiment of the present disclosure corrects the model by reflecting vehicle-specific parameters in the prediction model, it can produce a prediction result that corresponds to the actual response characteristics of the vehicle even under the same control input conditions.

[0107] For example, we can assume a case where one of two vehicles given the same steering input is a sports sedan (high tire stiffness) and the other is an SUV (low tire stiffness).

[0108] While a fixed parameter-based prediction model that does not reflect vehicle-specific information predicts the yaw rate response of two vehicles identically, a processor according to one embodiment of the present disclosure corrects the model by reflecting vehicle-specific information in the prediction model, thereby accurately reflecting the actual yaw rate response difference of each vehicle.

[0109] In other words, by generating predicted state information aligned with the dynamic characteristics of each vehicle, the processor minimizes prediction uncertainty caused by vehicle modeling errors and enables the generation of stable and highly reliable control inputs even under various driving conditions.

[0110] Referring again to FIG. 3, the processor can generate error information (S340) based on the difference between the predicted state information and the reference state information.

[0111] Here, error information may refer to an indicator that quantitatively represents how much the vehicle's predicted driving behavior deviates from the target driving path or reference state.

[0112] For example, the processor can compare values ​​such as the vehicle's position, speed, yaw rate, and attitude angle included in the predicted state information with the target position, target speed, target yaw rate, and target attitude angle included in the reference state information, respectively, calculate the difference, and produce it in the form of an error vector.

[0113] In one embodiment, the processor may classify the calculated error vector into components such as lateral position error, longitudinal velocity error, yaw angle error, and attitude angle error, and assign a weighting coefficient according to the degree of influence each item has on driving stability and path following performance.

[0114] For example, the processor can apply a relatively high weight to the lateral position error by reflecting the curvature information of the reference path when the vehicle travels through a curved section, and can apply a high weight to the speed error or yaw rate error in a straight section.

[0115] In other words, the processor can support more precise correction when calculating control inputs by comprehensively reflecting not only simple positional deviations but also the vehicle's attitude stability when configuring error information.

[0116] In one embodiment, the processor may calculate a control input to correct the state deviation of the vehicle based on the generated error information, or reflect it as an evaluation item of the objective function in a subsequent optimization operation.

[0117] Additionally, according to an embodiment, the processor can generate a control input for target path following (S350) based on an objective function containing error information.

[0118] In one embodiment, the objective function is defined as a mathematical indicator for quantitatively evaluating the driving performance of a vehicle, in order to simultaneously ensure control stability and energy efficiency while minimizing the deviation between the predicted state and the reference state. The processor may define the objective function to derive control inputs that enable the overall system operation to be performed most stably and efficiently, by considering various state changes and disturbances that occur during the driving of the vehicle.

[0119] According to one embodiment, the objective function may include an error term corresponding to error information, a control input term corresponding to a control input, and a disturbance term corresponding to a disturbance.

[0120] For a specific example, the processor can calculate an error term indicating the degree to which state variables, such as the vehicle's position, speed, yaw rate, and attitude angle included in the predicted state information, deviate from the target value included in the reference state information, and adjust the control input so that the magnitude of the error term is minimized.

[0121] In addition, the processor can suppress sudden changes in steering torque or braking torque and ensure the response stability of the system by applying control constraints through the control input term so that the magnitude or rate of change of the control input is not excessive.

[0122] Meanwhile, the disturbance term is a term that reflects the influence of external environmental factors or changes in driving conditions on vehicle behavior. The processor can apply conditions in which disturbances such as crosswinds, uneven road surfaces, and tire slip act to the prediction model, and calculate control inputs such that the value of the objective function is minimized even when the disturbance acts most unfavorably.

[0123] For example, the processor can adjust the control input to prioritize vehicle attitude stability by increasing the weight of the disturbance term when strong crosswinds are predicted in curved sections, and can simultaneously secure path following accuracy and energy efficiency by increasing the weight of the error term and the control input term in straight sections.

[0124] In other words, the processor can calculate an optimal control input that balances driving stability, responsiveness, and control efficiency by integrally considering the driving path, disturbance conditions, and control constraints to minimize the value of the objective function.

[0125] According to one embodiment, the processor may perform an operation to calculate a control input that minimizes the value of the objective function under the assumption that the disturbance occurrence is at its maximum. This configuration may be intended to maintain the stability of the entire system by considering in advance the impact of unpredictable disturbances on vehicle behavior during driving.

[0126] More specifically, the processor can set variables such as the magnitude, direction, and duration of the disturbance through a prediction model, and calculate an objective function by assuming the condition in which the disturbance acts most adversely on the vehicle's stability.

[0127] Subsequently, the processor can control the vehicle so that trajectory deviation or attitude instability is minimized even if disturbances occur during actual driving by adjusting the control input so that the value of the objective function is minimized under assumed conditions.

[0128] For example, if control inputs are calculated by assuming a situation where a crosswind acts continuously in a certain direction, changes in the vehicle's attitude angle or yaw rate deviations can be corrected immediately even if similar disturbances occur during actual driving.

[0129] In other words, the processor can ensure robustness by optimizing the control input by preemptively considering the case where the disturbance acts in the most unfavorable direction, thereby enabling stable trajectory tracking without separate correction control after the disturbance occurs.

[0130] Unlike a configuration that simply performs correction control after a disturbance occurs, the above-described configuration has the advantage of minimizing control response delay and improving driving stability by allowing the control input to be calculated under conditions where the disturbance is predicted in advance.

[0131] According to one embodiment of the present disclosure, the objective function may include a plurality of weight matrices corresponding to each of the error term, the control input term, and the disturbance term.

[0132] Multiple weight matrices serve as sets of parameters for independently adjusting driving performance elements (e.g., accuracy, stability, robustness) that the control system must consider, and they play a role in numerically reflecting the importance of each item during the optimization process of control inputs.

[0133] In one embodiment, a plurality of weighting matrices may include a first weighting matrix regarding error information reduction, a second weighting matrix regarding control input limiting, and a third weighting matrix regarding a maximum disturbance occurrence condition.

[0134] According to one embodiment, the first weighting matrix determines the weight of the error term so that the vehicle follows the target path more accurately, and the second weighting matrix may be set to ensure control stability so that the control input does not fluctuate excessively or exceed the actuator limit. Additionally, the third weighting matrix is ​​a weight corresponding to the disturbance term and may be utilized to compensate for the impact of external vehicle factors, such as crosswinds, reduced road surface friction, and tire slip, on system stability.

[0135] As multiple weight matrices are included in the objective function, the processor can maintain a balance between disturbance compensation performance and driving responsiveness by adjusting the relative weights between each item, even in situations where disturbance acts as a maximum condition.

[0136] For example, when it is predicted that a disturbance will act significantly in the lateral direction upon entering a curved section, the processor can increase the value of the third weighting matrix to strengthen the influence of the disturbance compensation term, and at the same time lower the first weighting matrix to mitigate sensitivity to path-following errors, thereby prioritizing the maintenance of a stable attitude of the vehicle. Conversely, when a condition is detected where the magnitude of the disturbance has decreased in a straight section, the weight of the first weighting matrix is ​​increased again to strengthen the responsiveness of returning to the center of the trajectory, and the third weighting matrix is ​​decreased to reduce unnecessary control torque.

[0137] In addition, in wet or icy sections where the road surface friction coefficient is rapidly reduced, the processor can prevent tire slip or wheel spin by increasing the second weighting matrix to limit the rate of change of the control input.

[0138] In a specific embodiment, the processor can adjust a plurality of weight matrices based on the vehicle's driving state, disturbance magnitude, and driving environment.

[0139] More specifically, the processor can recognize the driving state of the vehicle (e.g., speed, yaw rate, steering angle, acceleration), disturbance magnitude (e.g., crosswind intensity, road surface friction coefficient, change in gradient), and driving environment (e.g., road curvature, lane shape, road surface condition, weather conditions) by analyzing data obtained from sensor modules mounted on the vehicle or an external perception system.

[0140] The processor evaluates the vehicle's dynamic stability indicators (e.g., body lateral acceleration, yaw angle deviation, steering response delay, etc.) based on the recognized information and can classify whether the current driving state corresponds to normal driving, a disturbance influence section, or a low-friction risk section. Based on the classification result, the processor can update multiple weight matrices reflected in the objective function on a situational basis.

[0141] For example, if the processor detects that the vehicle is entering a curved section through curvature sensor and lane recognition data, and at the same time detects a disturbance exceeding a reference value from the crosswind sensor, it can increase the value of the third weighting matrix to increase the weight of the disturbance compensation term and relatively decrease the first weighting matrix to prioritize attitude stability over trajectory center return.

[0142] On the other hand, if the processor determines through the speed sensor and road surface friction estimation module that the disturbance effect is minimal in the straight section and the road surface friction coefficient is at a normal level, it can increase the weight of the first weighting matrix to enable precise trajectory tracking control.

[0143] In addition, the processor can suppress wheel spin by analyzing braking input and wheel speed data and, when the road surface slip rate increases above a reference level, by adjusting the value of the second weighting matrix upward to limit the rate of change of the control input.

[0144] As described above, the processor can achieve optimal control adaptable to changes in driving conditions by integrally analyzing driving status, disturbance magnitude, and driving environment information, and dynamically adjusting multiple weight matrices according to the control objectives (accuracy, stability, robustness) required for each driving condition.

[0145] In other words, the processor can minimize the degradation of control performance even under environmental changes or unpredictable disturbances, and can generate control inputs with improved driving stability and responsiveness.

[0146] FIG. 5 is an exemplary flowchart illustrating a control input calculation process (500) based on an objective function according to one embodiment of the present disclosure.

[0147] Referring to FIG. 5, a processor according to one embodiment can generate error information based on the difference between predicted state information and reference state information (S510) and input the generated error information into the error term of the objective function (S520).

[0148] Here, error information is an indicator that quantitatively represents how much the vehicle's predicted behavior deviates from the target driving path, and may consist of the difference between the predicted position, speed, yaw rate, attitude angle, etc., and the target value corresponding to the reference state.

[0149] For example, if the predicted yaw rate is calculated to be lower than the reference yaw rate while the vehicle is driving through a curved section, the processor may determine this as an 'understeer tendency' and record the deviation value as error information. Conversely, if the predicted lateral position deviates from the reference path in a straight section, it may be reflected as a trajectory tracking error.

[0150] Each error item may be assigned a weight based on its impact on driving stability, tracking performance, or control responsiveness, and the processor may construct an error term within the objective function by applying the weight information together. Through this, the objective function takes on a smaller value as the predicted state of the vehicle approaches the reference state, and can be used as an evaluation criterion for optimization operations performed in subsequent steps (S531~S540).

[0151] The processor can adjust the first weighting matrix (S531) to reduce error information. More specifically, the first weighting matrix is ​​a parameter for adjusting the weight of the error term so that the vehicle accurately follows a reference driving path, and the error sensitivity can be set differently depending on the driving situation.

[0152] For example, when a vehicle enters a curved section at high speed, a small error can have a significant impact on driving stability, so the processor can adjust the values ​​of the first weighting matrix to increase sensitivity to lateral position error and yaw angle error.

[0153] On the other hand, in driving environments where stability is sufficiently ensured, such as low-speed congested sections, the processor can adjust the values ​​of the first weighting matrix to prioritize the smoothness of the control input and energy efficiency. Through this adjustment, the processor can dynamically change the weight of the error term according to driving conditions, thereby maintaining a balance between stability and responsiveness.

[0154] Additionally, the processor may adjust the second weighting matrix (S532) to limit the size of the control input. The second weighting matrix is ​​used to control the size or rate of change of control commands transmitted to vehicle actuators, such as steering, driving, and braking, and is utilized to prevent excessive fluctuations in the control input.

[0155] For example, since the vehicle may become unstable when the control input changes abruptly on a road surface with low tire grip (e.g., wet road, icy section), the processor can increase the value of the second weighting matrix to strengthen the weight of the control input term.

[0156] Accordingly, control commands are suppressed from becoming excessively large during the optimization process, thereby ensuring both ride comfort and control stability simultaneously. Conversely, in sports driving mode where control responsiveness is prioritized, the second weighting matrix can be relatively lowered to improve steering responsiveness.

[0157] Additionally, the processor can adjust the third weighting matrix (S533) to assume the maximum condition for disturbance occurrence.

[0158] The third weighting matrix is ​​a weight corresponding to the disturbance term, and is set to compensate for the effect of external environmental factors (e.g., crosswind, uneven road surface, tire slip) on vehicle behavior.

[0159] For example, based on crosswind sensor or vehicle body attitude sensor data, if crosswind is detected above a certain threshold or the road surface friction coefficient decreases rapidly, the processor can set the third weighting matrix high to strengthen the weight of the disturbance term.

[0160] When the third weight matrix is ​​adjusted, the compensation term for disturbances is relatively strengthened during the operation of the objective function, thereby minimizing trajectory deviation or attitude instability caused by disturbances. Conversely, in normal driving situations where disturbances are rarely detected, the third weight matrix can be lowered to reduce unnecessary control energy consumption and increase efficiency.

[0161] According to the above process, the processor can dynamically adjust a plurality of weight matrices based on the driving state, disturbance magnitude, and driving environment conditions, and construct an objective function that reflects the result.

[0162] In addition, in an embodiment, the processor may perform an operation to calculate a control input (S540) that minimizes the value of the objective function.

[0163] More specifically, the processor can set possible ranges for the magnitude, direction, and duration of disturbance terms and search for the disturbance condition that is most adverse to vehicle behavior.

[0164] The processor can assume the detected disturbance condition as the 'disturbance maximum condition' and calculate the control input to minimize the value of the objective function under the assumed condition.

[0165] In other words, the processor optimizes control inputs such as steering, driving, and braking to minimize the error between the predicted state and the reference state, while assuming in advance the worst-case disturbance conditions that the vehicle may face (e.g., strong crosswinds, sudden drop in friction, uneven road surface, etc.). This operation has a preemptive correction effect to prevent the vehicle's behavior from becoming unstable when the disturbance actually occurs, and consequently improves the vehicle's driving stability and trajectory tracking performance simultaneously.

[0166] For example, if it is predicted that a strong crosswind will persist in a certain direction in a curved section, the processor can set the crosswind in that direction as the maximum disturbance condition and calculate the steering input and braking torque so that the value of the objective function is minimized under that condition.

[0167] The calculated control input operates to immediately correct changes in the vehicle's attitude angle or yaw rate deviations even when the same crosswind occurs during actual driving, thereby minimizing trajectory deviation or vehicle shaking.

[0168] In summary, the processor considers the condition where disturbance is maximum (max) in advance and calculates control inputs to minimize system error and control energy (min) under that condition, thereby enabling the maintenance of stable vehicle behavior without additional correction processes even if unexpected disturbances occur during driving. Unlike control methods that perform feedback correction after a disturbance occurs, this configuration reduces control delay and improves robustness against disturbances by reflecting the impact of disturbances in advance during the prediction phase.

[0169] In other words, the control input produced by the processor is designed to have inherent robustness against disturbance conditions, so that the vehicle's attitude stability and trajectory following performance can be continuously maintained even if the magnitude or direction of the disturbance changes.

[0170] Accordingly, the control input generating device of the present disclosure can realize a stable control response with minimal intervention of additional correction control, even if disturbances actually occur.

[0171] FIG. 6 is an exemplary flowchart illustrating the distribution process (600) of an integrated control input according to an embodiment of the present disclosure.

[0172] In the present disclosure, the control input may be an integrated control input for controlling a plurality of actuators. An integrated control input refers to a signal that computationally integrates control requests generated from a plurality of chassis system controllers, such as those for steering, braking, driving, and suspension of a vehicle, within a single frame to generate an optimal control command while considering the dynamic characteristics of the entire vehicle.

[0173] In other words, unlike conventional methods where each actuator operates individually, the integrated control input may be a higher-level control command designed to simultaneously control interacting physical elements such as steering force, braking force, driving force, and suspension damping force by modeling the entire behavior of the vehicle as a single system.

[0174] For example, when a vehicle enters a curve at high speed, the processor can collectively generate control commands to adjust steering torque to the steering controller, braking pressure to the braking controller, and damping force to the suspension controller, respectively, in order to stably maintain lateral acceleration and yaw rate. Such integrated control inputs allow multiple control systems to operate complementarily, thereby minimizing response imbalances or control interference that may occur during single actuator control and ensuring the overall driving stability of the vehicle.

[0175] Referring to FIG. 6, a processor according to one embodiment can generate an integrated control input (S610) based on an objective function. In one embodiment, the integrated control input is a higher-level control command for controlling the overall behavior of a vehicle, and refers to an integrated control signal that can be commonly applied to actuators of steering, braking, driving, and suspension systems.

[0176] The processor can calculate the optimal integrated control input considering the interaction of each control system by reflecting the previously calculated error information and disturbance conditions in the objective function.

[0177] For example, to simultaneously stabilize changes in the vehicle's lateral acceleration, yaw rate, and attitude angle, the processor can analyze the correlation between steering control inputs (e.g., wheel steering angle), braking control inputs (e.g., braking pressure), and driving control inputs (e.g., driving torque), and perform an integrated computation so that the combination of these inputs satisfies the minimum conditions of the objective function.

[0178] In other words, the processor does not simply merge independent control commands for each system, but can perform integrated optimization calculations based on the vehicle's overall dynamic model so that all control variables operate interdependently. In this process, the processor dynamically corrects the mutual influences between control variables (e.g., reduced braking stability due to increased steering force, changes in driving force due to braking force distribution, etc.) to calculate an integrated control input that maximizes the vehicle's overall stability.

[0179] Accordingly, the generated integrated control input is in a form in which cooperative control between multiple control systems is realized, thereby enabling stable tracking of the vehicle's target trajectory while securing control performance that is robust against disturbances and response delays.

[0180] In addition, in an embodiment, the processor can acquire driving state information and driving environment information (S620).

[0181] In one embodiment, the driving state information is data measured through internal vehicle sensors (e.g., wheel speed sensor, steering angle sensor, yaw rate sensor, IMU, suspension stroke sensor, etc.) and may include the vehicle's speed, acceleration, steering angle, yaw rate, lateral acceleration, suspension displacement, etc.

[0182] In addition, driving environment information can be obtained from external perception sensors (e.g., cameras, lidar, radar) and road surface friction estimation sensors, and may include information such as road surface condition, curvature, slope, friction coefficient, crosswind strength and direction.

[0183] The processor can integrate and analyze driving state information and driving environment information to determine the current dynamic behavior state of the vehicle and generate basic data to determine the relative control contribution that each control system must perform.

[0184] In addition, in an embodiment, the processor can dynamically adjust (S630) the weights applied to each actuator based on driving state information and driving environment information.

[0185] According to one embodiment, the processor determines the driving conditions of the vehicle (e.g., straight sections, curved sections, slippery road surfaces, high-speed driving, etc.) in real time and can dynamically change the weight of steering, braking, driving, and suspension control accordingly.

[0186] For example, the processor can prioritize propulsion and stability by increasing the weight of the drive system in high-speed straight sections, and enhance the vehicle's rotational responsiveness and attitude stability by increasing the weight of the steering and braking systems in sharp curve sections.

[0187] As another example, on wet or icy roads with a low coefficient of friction, the weighting of the braking system can be lowered and the weighting of the suspension system increased to prioritize maintaining traction.

[0188] In this way, the processor can ensure the flexibility of the control system so that the integrated control input is optimized for each situation by updating weights in real time that reflect the influence of each control system according to driving conditions.

[0189] Additionally, in the embodiment, the processor can distribute the integrated control input (S640).

[0190] According to one embodiment, the processor can divide the integrated control input into control commands corresponding to each actuator according to previously adjusted weights.

[0191] For example, if the integrated control input is defined as the total control torque for vehicle attitude stabilization, the processor can appropriately distribute this control torque to the steering system, braking system, drive system, and suspension system so that each actuator operates cooperatively.

[0192] Specifically, the steering controller can output a steering angular velocity control command, the braking controller can output a braking pressure control command for each left and right wheel, the drive controller can output a torque distribution control command, and the suspension controller can output a damping force control command.

[0193] Through this distribution structure, the processor can achieve the vehicle's overall control objectives (e.g., following a target trajectory, attitude stabilization, and ride comfort) while preventing control imbalances caused by overload or response delays in specific systems. In other words, the distribution process of integrated control inputs functions as a step that improves the control efficiency and robustness of the entire system by optimizing the role of each actuator according to driving conditions.

[0194] FIG. 7 is an exemplary flowchart illustrating a control input update process (700) according to one embodiment of the present disclosure.

[0195] Referring to FIG. 7, a processor according to one embodiment can apply a control input generated by utilizing an objective function to a plurality of actuators (S710). In this case, the control input is an integrated control command calculated through the integrated optimization operation described above, and is distributed to a steering controller, a braking controller, a driving controller, and a suspension controller, respectively, to control the actual behavior of the vehicle.

[0196] For example, the processor can achieve longitudinal and lateral stabilization of the vehicle by simultaneously outputting steering torque, left and right braking pressure, driving torque distribution ratio, and damping force control signals for trajectory tracking in curved driving sections.

[0197] In addition, in an embodiment, the processor can obtain subsequent vehicle state information (S720) corresponding to a time after the control input is applied to a plurality of actuators.

[0198] Subsequent vehicle state information is data that reflects the actual impact of the control input on the vehicle's behavior, and may include actual feedback data such as vehicle speed, yaw rate, lateral acceleration, tire slip ratio, body attitude angle, and crosswind sensor values. Subsequent vehicle state information can be used as a verification input to analyze the deviation between the prediction model and the actual response.

[0199] In addition, in an embodiment, the processor can calculate subsequent predicted state information and subsequent error information based on subsequent vehicle state information (S730).

[0200] Specifically, the processor can update error information by reapplying the acquired subsequent vehicle state information as input to the prediction model and comparing the deviation between the predicted state information and the actual state.

[0201] For example, if the actual yaw rate of the vehicle is observed to be lower than predicted after the application of a control input, the processor determines that an understeer tendency has occurred in that section and can reflect the deviation as subsequent error information.

[0202] In this way, subsequent error information serves as real-time learning feedback to correct disturbances or system nonlinearities that were not reflected in the prediction phase.

[0203] In addition, in an embodiment, the processor can update the control input (S740) based on subsequent error information.

[0204] When subsequent error information increases above a certain threshold or deviations continuously accumulate in a specific control axis (e.g., yaw rate, lateral position error, etc.), the processor can readjust the weight matrix within the objective function to enhance the control sensitivity of the axis and update the control input by performing a new optimization operation.

[0205] For example, if a disturbance repeatedly acts in a specific direction (e.g., left crosswind direction), the processor can modify the control input by increasing the weight of the steering system to strengthen the steering torque weight and lowering the weight of the braking system to maintain overall energy efficiency.

[0206] According to this configuration, the control system of the present disclosure implements an adaptive optimization structure combining predictive control and real-time feedback control.

[0207] In other words, since the control input is not calculated and fixed as a one-time event but is continuously re-evaluated and corrected based on the actual vehicle response, control stability can be maintained even if disturbance environments or road surface conditions change.

[0208] For example, the configuration of Fig. 7 can improve control delay and disturbance adaptability issues through a self-updating control loop of the control input, and can secure significant robustness compared to a single feedforward or simple feedback control method.

[0209] FIG. 8 is an exemplary flowchart illustrating a method (800) for generating a control input according to one embodiment of the present disclosure. The steps illustrated in FIG. 8 may be changed in order as necessary, and at least one step may be omitted or added. The steps of FIG. 8 are merely one embodiment of the present disclosure, and the scope of the present disclosure is not limited thereto. Regarding features illustrated in FIG. 8 that overlap with features previously described in relation to FIG. 1 to 7, reference is made to FIG. 1 to 7, and the description thereof is omitted here.

[0210] According to one embodiment of the present disclosure, a control input generation method may include the step (S810) of obtaining reference state information corresponding to a driving target path of a vehicle.

[0211] According to one embodiment of the present disclosure, a control input generation method may include a step (S820) of obtaining current state information regarding the current driving state of a vehicle.

[0212] According to one embodiment of the present disclosure, a control input generation method may include the step (S830) of generating predicted state information based on current state information.

[0213] In one embodiment, the step of generating predicted state information based on current state information is,

[0214] It may include a step of generating predicted state information corresponding to current state information by utilizing a prediction model.

[0215] In addition, in the embodiments, the prediction model may be a model that predicts future driving conditions based on a plurality of state variables reflecting the dynamic characteristics of the vehicle.

[0216] In addition, in an embodiment, the step of generating predicted state information may include the step of applying a state matrix determined based on vehicle unique information to the current state information to reflect the vehicle's dynamic characteristics in the prediction model.

[0217] According to one embodiment of the present disclosure, a control input generation method may include a step (S840) of generating error information based on the difference between predicted state information and reference state information.

[0218] According to one embodiment of the present disclosure, a control input generation method may include the step of generating a control input for target path following (S850) based on an objective function including error information.

[0219] In one embodiment, the objective function may include an error term corresponding to error information, a control input term corresponding to a control input, and a disturbance term corresponding to a disturbance.

[0220] In addition, in an embodiment, the step of generating a control input may include the step of performing an operation to calculate a control input such that the value of the objective function is minimized under the assumption that the disturbance occurrence is maximum.

[0221] In addition, in an embodiment, the objective function may include a plurality of weighting matrices corresponding to each of the error term, the control input term, and the disturbance term, and the plurality of weighting matrices may include a first weighting matrix regarding error information reduction, a second weighting matrix regarding control input limitation, and a third weighting matrix regarding a maximum disturbance occurrence condition.

[0222] In addition, in the embodiment, a plurality of weighting matrices can be adjusted based on the vehicle's driving state, disturbance magnitude, and driving environment.

[0223] In addition, in an embodiment, the control input may be an integrated control input for controlling a plurality of actuators.

[0224] According to one embodiment of the present disclosure, the control input generation method may further include the step of dynamically adjusting and distributing weights applied to each actuator based on driving state information and driving environment information for the integrated control input.

[0225] According to one embodiment of the present disclosure, a method for generating a control input may further include the step of obtaining subsequent vehicle state information corresponding to a time after the control input is applied to a plurality of actuators.

[0226] According to one embodiment of the present disclosure, a control input generation method may further include the step of calculating subsequent predicted state information and subsequent error information based on subsequent vehicle state information.

[0227] According to one embodiment of the present disclosure, a method for generating a control input may further include the step of updating the control input based on subsequent error information.

[0228]

[0229] A vehicle control system according to another embodiment of the present disclosure is disclosed.

[0230] A vehicle control system according to another embodiment of the present disclosure may include a reference state generating device that generates reference state information corresponding to a target driving path of a vehicle, an observer that calculates current state information by estimating the actual driving state of a vehicle based on sensor data of the vehicle, a plurality of actuators for driving, braking, steering, and suspension control, and a control input generating device that generates a control input for controlling the operation of the plurality of actuators.

[0231] The components included in the aforementioned vehicle control system are merely examples for the sake of understanding, and the present disclosure is not limited thereto.

[0232] In one embodiment, a control input generating device of a vehicle control system may obtain reference state information from a reference state generating device, obtain current state information from an observer, generate predicted state information based on the current state information, generate error information based on the difference between the predicted state information and the reference state information, and generate a control input for target path following based on an objective function including the error information.

[0233] In one embodiment, the objective function may include an error term corresponding to error information, a control input term corresponding to a control input, and a disturbance term corresponding to a disturbance.

[0234] In one embodiment, a control input generating device of a vehicle control system can perform an operation to calculate a control input that minimizes the value of an objective function under the assumption that the disturbance occurrence is at its maximum.

[0235] Accordingly, the vehicle control system according to the present disclosure organically links prediction-based integrated control and actuator distributed control, thereby enabling the simultaneous securing of path-following stability and driving safety even under various driving situations and disturbance conditions.

[0236] The devices, methods, configurations, glyphs, and operations described herein may be implemented in digital electronic circuits, or computer software, firmware, or hardware comprising structures disclosed herein and structural equivalents, or combinations of one or more of these. The glyphs described herein may be implemented as one or more computer programs, for example, as one or more modules of computer program instructions encoded on a computer storage medium to control execution by a data processing device or operation by a data processing device. Program instructions may be encoded in artificially generated propagated signals, for example, mechanically generated electrical, optical, or electromagnetic signals generated to encode information for transmission to a suitable receiver device for execution by a data processing device. The computer storage medium may be or may include a computer-readable storage device, a computer-readable storage substrate, a random or serial access memory array or device, or a combination of one or more of these. Although the computer storage medium is not a propagated signal, the computer storage medium may be a source or destination of computer program instructions encoded in an artificially generated propagated signal. Additionally, a computer storage medium may be one or more individual physical components or media (e.g., multiple CDs, disks, or other storage devices) or may include. The operations described herein may be implemented as operations performed by a data processing device on data stored in one or more computer-readable storage devices or data received from other sources.

[0237] The foregoing description is merely an illustrative explanation of the technical concept of the present disclosure, and those skilled in the art to which the present disclosure pertains may make various modifications and variations within the scope of the essential characteristics of the technical concept. Furthermore, since these embodiments are intended to explain rather than limit the technical concept of the present disclosure, the scope of the technical concept is not limited by these embodiments.

[0238]

[0239] CROSS-REFERENCE TO RELATED APPLICATION

[0240] This patent application claims priority pursuant to Section 119(a) of the U.S. Patent Act (35 USC § 119(a)) to Patent Application No. 10-2025-0010208 filed in Korea on January 23, 2025 and Patent Application No. 10-2025-0200204 filed in Korea on December 16, 2025, all of which are incorporated by reference into this patent application. Furthermore, this patent application claims priority in countries other than the United States for the same reasons as above, all of which are incorporated by reference into this patent application.

Claims

1. In a control input generating device, Memory for storing at least one instruction; and It includes at least one processor that executes the above at least one instruction, and The above at least one processor, by executing the above at least one instruction, Acquire reference state information corresponding to the vehicle's driving target path, and Acquires current status information regarding the current driving state of the above vehicle, and Based on the above current state information, predicted state information is generated, and Error information is generated based on the difference between the above predicted state information and the above reference state information, and Generating a control input for following the target path based on an objective function including the above error information, Control input generating device.

2. In Paragraph 1, The above processor is, Generate the predicted state information corresponding to the current state information using a prediction model, and The above prediction model is, A control input generating device, which is a model for predicting future driving states based on multiple state variables reflecting the dynamic characteristics of the vehicle.

3. In Paragraph 2, The above processor is, A control input generating device that applies a state matrix determined based on vehicle unique information to the current state information to reflect the dynamic characteristics of the vehicle in the prediction model.

4. In Paragraph 1, The above objective function is, A control input generating device comprising an error term corresponding to the above error information, a control input term corresponding to the above control input, and a disturbance term corresponding to the disturbance.

5. In Paragraph 4, The above processor is, A control input generating device that performs an operation to calculate a control input such that the value of the objective function is minimized under the assumption that the above disturbance occurrence is maximum.

6. In Paragraph 4, The above objective function is, It includes a plurality of weighting matrices corresponding to each of the above error term, the above control input term and the above disturbance term, and The above plurality of weight matrices are, A control input generating device comprising a first weighting matrix for reducing error information, a second weighting matrix for limiting control input, and a third weighting matrix for maximum disturbance occurrence conditions.

7. In Paragraph 6, The above processor is, A control input generating device that adjusts the plurality of weighting matrices based on the driving state, disturbance magnitude, and driving environment of the vehicle.

8. In Paragraph 1, The above control input is, It is an integrated control input for controlling multiple actuators, and The above processor is, A control input generating device that distributes the integrated control input by dynamically adjusting the weights applied to each actuator based on driving state information and driving environment information.

9. In Paragraph 8, The above processor is, Obtain subsequent vehicle state information corresponding to a point in time after the above control input is applied to the plurality of actuators, and Based on the above subsequent vehicle status information, subsequent predicted status information and subsequent error information are calculated, and Updating the control input based on the above subsequent error information, Control input generating device.

10. A step of obtaining reference state information corresponding to the vehicle's driving target path; A step of obtaining current state information regarding the current driving state of the above vehicle; A step of generating predicted state information based on the above current state information; A step of generating error information based on the difference between the predicted state information and the reference state information; and A method comprising the step of generating a control input for following the target path based on an objective function including the error information above. Method for generating control inputs.

11. In Paragraph 10, The step of generating predicted state information based on the above current state information is: The method includes the step of generating the predicted state information corresponding to the current state information using a prediction model, and The above prediction model is, A control input generation method, which is a model for predicting future driving states based on multiple state variables reflecting the dynamic characteristics of the vehicle.

12. In Paragraph 11, The step of generating the above-mentioned predicted state information is, A method for generating control inputs, comprising the step of applying a state matrix determined based on vehicle unique information to the current state information to reflect the dynamic characteristics of the vehicle in the prediction model.

13. In Paragraph 10, The above objective function is, A method for generating a control input, comprising an error term corresponding to the above error information, a control input term corresponding to the above control input, and a disturbance term corresponding to the disturbance.

14. In Paragraph 13, The step of generating the above control input is, A method for generating a control input, comprising the step of performing an operation to calculate a control input such that the value of the objective function is minimized under the assumption that the above disturbance occurrence is maximum.

15. In Paragraph 13, The above objective function is, It includes a plurality of weighting matrices corresponding to each of the above error term, the above control input term and the above disturbance term, and The above plurality of weight matrices are, A method for generating a control input, comprising a first weighting matrix for reducing error information, a second weighting matrix for limiting the control input, and a third weighting matrix for the maximum condition for disturbance occurrence.

16. In Paragraph 15, The above plurality of weight matrices are, A method for generating control inputs that is adjusted based on the driving state, disturbance magnitude, and driving environment of the vehicle.

17. In Paragraph 10, The above control input is, It is an integrated control input for controlling multiple actuators, and The above integrated control input is, A control input generation method in which weights applied to each actuator are dynamically adjusted and distributed based on driving state information and driving environment information.

18. In Paragraph 17, The above method for generating control inputs is, A step of obtaining subsequent vehicle state information corresponding to a point in time after the above control input is applied to the plurality of actuators; A step of calculating subsequent predicted state information and subsequent error information based on the above subsequent vehicle state information; and A method for generating a control input, further comprising the step of updating the control input based on the above subsequent error information.

19. In a vehicle control system, Reference state generating device that generates reference state information corresponding to the vehicle's driving target path; An observer that estimates the actual driving state of a vehicle based on sensor data and calculates current state information; A plurality of actuators for driving, braking, steering and suspension control; and It includes a control input generating device for generating a control input to control the operation of the plurality of actuators above, and The control input generating device acquires reference state information from the reference state generating device, acquires current state information from the observer, generates predicted state information based on the current state information, generates error information based on the difference between the predicted state information and the reference state information, and generates a control input for the target path following based on an objective function including the error information. Vehicle control system.

20. In Paragraph 19, The above objective function is, A vehicle control system comprising an error term corresponding to the above error information, a control input term corresponding to the above control input, and a disturbance term corresponding to the disturbance.

Images