Cruise control method and device, electronic device and storage medium

Abstract

The application provides a cruise control method and device, electronic equipment and a storage medium, and belongs to the technical field of cruise control. The current vehicle operating state data of a cruise vehicle team is acquired, the current vehicle operating state data is input into a target Bayesian regression model for disturbance prediction, environment disturbance data is obtained, the longitudinal relative distance and longitudinal relative speed between a preset merging vehicle and the cruise vehicle team are acquired, a safety barrier set is constructed according to the environment disturbance data, the longitudinal relative distance and the longitudinal relative speed, the safety barrier set is converted into a linear constraint, a control amount is generated according to the linear constraint, the vehicle spacing of the cruise vehicle team is adjusted according to the control amount, and the preset merging vehicle is controlled to join the cruise vehicle team according to the vehicle spacing. The cruise vehicle team can handle the vehicle merging behavior of the external surrounding environment, and the safety of the driving process is improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of cruise control, and particularly relates to a cruise control method and device, an electronic device and a storage medium. BACKGROUND

[0002] When an interconnected cruise vehicle fleet is running at high speed, the surrounding environment may have lane-changing and cutting behavior, which not only reduces the efficiency of the vehicle fleet running, but also causes safety hazards in the driving process. In order to improve the safety of the driving process, the cruise vehicle fleet needs to quickly and effectively respond to the interference of external cutting behavior to ensure that the cruise vehicle fleet can continue to run safely. SUMMARY

[0003] The main purpose of the embodiments of the present application is to provide a cruise control method and device, an electronic device and a storage medium, which aims to deal with the cutting behavior of vehicles in the surrounding environment and improve the safety of the driving process.

[0004] To achieve the above purpose, a cruise control method is provided in the first aspect of the embodiments of the present application, and the method comprises:

[0005] obtaining current vehicle running state data of a cruise vehicle fleet;

[0006] inputting the current vehicle running state data into a target Bayesian regression model for disturbance prediction to obtain environmental disturbance data;

[0007] obtaining a longitudinal relative distance and a longitudinal relative speed between a preset cutting vehicle and the cruise vehicle fleet;

[0008] constructing a safety barrier set according to the environmental disturbance data, the longitudinal relative distance and the longitudinal relative speed;

[0009] converting the safety barrier set into a linear constraint, and generating a control amount according to the linear constraint;

[0010] adjusting a vehicle spacing of the cruise vehicle fleet according to the control amount, and controlling the preset cutting vehicle to join the cruise vehicle fleet according to the vehicle spacing.

[0011] In some embodiments, before the current vehicle running state data is input into the target Bayesian regression model for disturbance prediction to obtain environmental disturbance data, the cruise control method further comprises:

[0012] calculating the similarity between the current vehicle running state data and the pre-acquired historical vehicle running state data;

[0013] performing data screening on the historical vehicle running state data according to the similarity to obtain candidate running state data;

[0014] determining target operating state data according to the current vehicle operating state data and the candidate operating state data;

[0015] The preset Bayesian regression model is updated according to the target operating state data to obtain the target Bayesian regression model.

[0016] In some embodiments, the preset Bayesian regression model includes a first initial kernel matrix and a second initial kernel matrix, where the second initial kernel matrix is ​​the inverse matrix of the first initial kernel matrix. Updating the preset Bayesian regression model according to the target operating state data to obtain the target Bayesian regression model includes:

[0017] determining a first kernel representation matrix according to the target operating state data and the historical vehicle operating state data;

[0018] performing a transposition process on the first kernel representation matrix to obtain a second kernel representation matrix;

[0019] Based on the Woodbury matrix identity, the first initial kernel matrix is ​​updated according to the first kernel representation matrix and the second kernel representation matrix to obtain a first target kernel matrix;

[0020] updating the second initial kernel matrix according to the first kernel representation matrix and the second kernel representation matrix to obtain a second target kernel matrix;

[0021] The target Bayesian regression model is determined according to the first target kernel matrix and the second target kernel matrix.

[0022] In some embodiments, updating the first initial kernel matrix according to the first kernel representation matrix and the second kernel representation matrix to obtain a first target kernel matrix includes:

[0023] Performing matrix multiplication on the first kernel representation matrix and the second kernel representation matrix to obtain a first intermediate matrix;

[0024] Perform matrix addition on the first initial kernel matrix and the first intermediate matrix to obtain the first target kernel matrix.

[0025] In some embodiments, updating the second initial kernel matrix according to the first kernel representation matrix and the second kernel representation matrix to obtain a second target kernel matrix includes:

[0026] Perform a first matrix calculation based on the first kernel representation matrix, the second kernel representation matrix, and the second initial kernel matrix to obtain a second intermediate matrix;

[0027] Perform a second matrix calculation based on the first kernel representation matrix, the second kernel representation matrix, and the second initial kernel matrix to obtain a third intermediate matrix;

[0028] The second initial kernel matrix is ​​updated according to the second intermediate matrix and the third intermediate matrix to obtain the second target kernel matrix.

[0029] In some embodiments, constructing a safety obstacle set based on the environmental disturbance data, the longitudinal relative distance, and the longitudinal relative speed includes:

[0030] Obtaining confidence of the environmental disturbance data;

[0031] If the confidence level is within a preset confidence interval, the safety obstacle set is constructed according to the environmental disturbance data, the longitudinal relative distance, and the longitudinal relative speed.

[0032] In some embodiments, if the confidence level is within a preset confidence interval, constructing the safety obstacle set based on the environmental disturbance data, the longitudinal relative distance, and the longitudinal relative speed includes:

[0033] If the confidence level is within a preset confidence interval, determining a target relative distance based on the environmental disturbance data, the longitudinal relative distance, the longitudinal relative speed, and a preset fleet system response time;

[0034] The safety obstacle set is constructed according to the target relative distance and a preset safety distance threshold.

[0035] To achieve the above-mentioned objectives, a second aspect of an embodiment of the present application provides a cruise control device, the device comprising:

[0036] The first acquisition module is used to obtain the current vehicle operation status data of the cruising fleet;

[0037] A disturbance prediction module is used to input the current vehicle operating state data into a target Bayesian regression model to perform disturbance prediction and obtain environmental disturbance data;

[0038] A second acquisition module is used to acquire a longitudinal relative distance and a longitudinal relative speed between a preset lane-cutting vehicle and the cruising convoy;

[0039] A construction module, configured to construct a safety obstacle set according to the environmental disturbance data, the longitudinal relative distance, and the longitudinal relative speed;

[0040] a conversion module, configured to convert the safety obstacle set into a linear constraint and generate a control variable according to the linear constraint;

[0041] A cruise control module is configured to adjust a vehicle spacing of the cruise platoon according to the control quantity, and control the preset cut-in vehicle to join the cruise platoon according to the vehicle spacing.

[0042] To achieve the above object, a third aspect of the embodiments of the present application provides an electronic device, which comprises a memory and a processor, the memory stores a computer program, and the processor implements the cruise control method of the first aspect when executing the computer program.

[0043] To achieve the above object, a fourth aspect of the embodiments of the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the cruise control method of the first aspect.

[0044] The cruise control method, the cruise control device, the electronic device and the computer readable storage medium provided by the present application can obtain current vehicle running state data of a cruise platoon, input the current vehicle running state data into a target Bayesian regression model for disturbance prediction, learn model uncertainty caused by external environmental disturbance to the cruise platoon in real time, and obtain environmental disturbance data. The longitudinal relative distance and the longitudinal relative speed between a preset cut-in vehicle and the cruise platoon are obtained, a safety barrier set is constructed according to the environmental disturbance data, the longitudinal relative distance and the longitudinal relative speed, the safety barrier set is used to ensure that the cruise platoon asymptotically and quickly converges to a safe state from an unsafe state under the interference of the cut-in vehicle. The safety barrier set is converted into a linear constraint, the linear constraint is used to quickly limit the distance between the cut-in vehicle and the cruise platoon, a control quantity is generated according to the linear constraint, the vehicle spacing of the cruise platoon is adjusted according to the control quantity, and the cruise platoon and the cut-in vehicle maintain a safe cruise distance. The preset cut-in vehicle is controlled to join the cruise platoon according to the vehicle spacing, the vehicle cut-in behavior of the surrounding environment can be handled, and the safety of the driving process is improved. BRIEF DESCRIPTION OF DRAWINGS

[0045] Figure 1 is a flowchart of the cruise control method provided by the embodiments of the present application;

[0046] Figure 2 is another flowchart of the cruise control method provided by the embodiments of the present application;

[0047] Figure 3 is a flowchart of step S240 in Figure 2 ;

[0048] Figure 4 is a flowchart of step S330 in Figure 3 ;

[0049] Figure 5 is a flowchart of step S340 in Figure 3 ;

[0050] Figure 6 yes Figure 1 Flowchart of step S140 in FIG.

[0051] Figure 7 yes Figure 6 Flowchart of step S620 in FIG.

[0052] Figure 8 is a structural diagram of a cruise control device provided in an embodiment of the present application;

[0053] Figure 9 This is a schematic diagram of the hardware structure of the electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0054] In order to make the purpose, technical solutions and advantages of this application more clear, the following further describes this application in detail with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain this application and are not intended to limit this application.

[0055] It should be noted that although the device schematics illustrate functional module divisions and the flowcharts illustrate logical sequences, in certain circumstances, the steps shown or described may be performed in a sequence that differs from the module divisions in the device or the sequence in the flowcharts. The terms "first," "second," and so on, in the specification, claims, and drawings, are used to distinguish similar items and are not necessarily used to describe a specific sequence or precedence.

[0056] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application pertains. The terms used herein are for the purpose of describing the embodiments of this application only and are not intended to limit this application.

[0057] In the field of autonomous driving, connected cruise control can significantly improve transportation efficiency, especially in centralized freight transport and alleviating traffic congestion. When a connected cruise fleet operates at high speeds, surrounding vehicles may try to cut in, reducing the fleet's efficiency and posing a safety hazard. To improve driving safety, the fleet must quickly and effectively respond to these external lane-cutting interruptions to ensure continued safe operation.

[0058] Based on this, the embodiments of the present application provide a cruise control method, a cruise control device, an electronic device and a computer-readable storage medium, which are intended to handle the behavior of vehicles cutting in from the surrounding environment and improve the safety of the driving process.

[0059] The cruise control method, cruise control device, electronic device and computer-readable storage medium provided in the embodiments of the present application are specifically described through the following embodiments. First, the cruise control method in the embodiments of the present application is described.

[0060] The cruise control method provided in the embodiment of the present application relates to the field of cruise control technology. The cruise control method provided in the embodiment of the present application can be applied in a terminal, can be applied in a server side, or can be software running in a terminal or a server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or as a server cluster or distributed system composed of multiple physical servers, or as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the cruise control method, etc., but is not limited to the above forms.

[0061] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments in which tasks are performed by remote processing devices connected via a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0062] Figure 1 This is an optional flowchart of a cruise control method provided in an embodiment of the present application. Figure 1 The method may include but is not limited to steps S110 to S160.

[0063] Step S110, obtaining the current vehicle operation status data of the cruising fleet;

[0064] Step S120 , inputting the current vehicle operating state data into the target Bayesian regression model for disturbance prediction to obtain environmental disturbance data;

[0065] Step S130, obtaining a preset longitudinal relative distance and longitudinal relative speed between the vehicle cutting in and the cruising convoy;

[0066] Step S140, constructing a safety obstacle set based on the environmental disturbance data, the longitudinal relative distance, and the longitudinal relative speed;

[0067] Step S150, converting the safety obstacle set into linear constraints, and generating control variables according to the linear constraints;

[0068] Step S160: adjusting the distance between vehicles in the cruising convoy according to the control amount, and controlling the preset jam-in vehicles to join the cruising convoy according to the distance between vehicles.

[0069] In step S110 of some embodiments, the vehicle operating status of the cruising fleet at the current time is obtained through on-board sensors to obtain current vehicle operating status data. The cruising fleet includes multiple vehicles, and the vehicle operating status includes the vehicle position, vehicle speed, vehicle spacing, and vehicle position deviation of each vehicle. The vehicle position deviation is the deviation between the current vehicle position and the predicted vehicle position. The predicted vehicle position is the vehicle's reachable position at the current time estimated by the vehicle motion estimation model based on the vehicle operating status at historical time. By collecting the vehicle operating status in real time online, the Bayesian regression model can perform active learning, making the model highly flexible and capable of online learning, thereby improving the model's adaptability to the environment.

[0070] In step S120 of some embodiments, the surrounding environment, such as road gradient, friction, wind disturbance, vehicle cutting in, air resistance, etc., will cause disturbances to the cruising fleet, reducing the task execution efficiency of the cruising fleet. In order to enable the cruising fleet to quickly cope with these uncertain situations and correctly respond to surrounding vehicles cutting in, it is necessary to evaluate the model uncertainty brought to the cruising fleet by the external environmental disturbance, input the current vehicle operation state data into the target Bayesian regression model for disturbance prediction, and obtain the environmental disturbance data and the confidence of the environmental disturbance data. The environmental disturbance data is the disturbance vector caused by the external environment to the cruising fleet, such as the deviation between the predicted arrival position of the vehicle at the next time and the actual arrival position at the next time. Assuming that the disturbance vector has a certain smoothness, the cruising fleet model has local Lipschitz continuity, so that there is a unique solution for the vehicle operation state. The confidence is used to characterize the uncertainty of the prediction.

[0071] The target Bayesian regression model is represented as f(x*, D), where D represents the dataset and x* represents the current vehicle operating state data. The dataset includes multiple sample data, including the vehicle operating state at historical times and the expected distribution y under the vehicle operating state. The kernel function of the target Bayesian regression model is used to calculate the similarity between the current vehicle operating state data and the data points in the dataset. The predicted mean is determined based on the kernel matrix, noise variance, similarity, and expected distribution of the target Bayesian regression model, and this predicted mean is used as the environmental disturbance data. The predicted mean is calculated as follows:

[0072]

[0073] Where μ is the predicted mean; k is the kernel function; k(x*) represents the similarity; T represents the transposition operation; A is the kernel matrix; is the noise variance; I is the identity matrix; -1 indicates the matrix inverse operation.

[0074] The kernel function is used to calculate the kernel function value of the current vehicle operating status data. The prediction variance is determined based on the kernel function value, kernel matrix, noise variance, and similarity, and the prediction variance is used as the confidence level. The larger the prediction variance, the greater the uncertainty. The prediction variance is calculated as follows:

[0075]

[0076] Among them, σ 2 is the prediction variance; k(x*,xo) is the kernel function value.

[0077] Cruising fleets generate real-time vehicle status data. To ensure real-time disturbance prediction, the target Bayesian regression model must continuously adapt to the new vehicle status data. This embodiment of the present application employs an incremental active learning approach to train the target Bayesian regression model, improving the model's adaptability to the environment. The following describes the training method for the target Bayesian regression model in detail.

[0078] See also Figure 2 In some embodiments, before step S120, the cruise control method may further include but is not limited to steps S210 to S240:

[0079] Step S210, calculating the similarity between the current vehicle operating state data and the previously acquired historical vehicle operating state data;

[0080] Step S220 , screening the historical vehicle operating status data based on similarity to obtain candidate operating status data;

[0081] Step S230, determining target operating state data based on the current vehicle operating state data and the candidate operating state data;

[0082] Step S240 , updating the preset Bayesian regression model according to the target operating state data to obtain a target Bayesian regression model.

[0083] In step S210 of some embodiments, the historical vehicle operating status data is vehicle operating status data collected earlier than the current time. In order to improve the accuracy of the current time disturbance prediction, it is necessary to perform sample screening on the historical vehicle operating status data to obtain samples with information value. The vehicle operating status collected over a period of time is a continuous time series, and the data points of the continuous time series are correlated. In order to evaluate the correlation between the data points, the Gaussian kernel is used to calculate the similarity between the current vehicle operating status data and the previously acquired historical vehicle operating status data. The expression of the Gaussian kernel is defined as follows:

[0084]

[0085] Among them, x i and x j They represent two data points, namely the current vehicle operation status data and the historical vehicle operation status data; σ is the parameter of the Gaussian kernel.

[0086] In step S220 of some embodiments, if the similarity is greater than or equal to a preset similarity threshold, the historical vehicle operating state data is used as candidate operating state data. If the similarity is less than the preset similarity threshold, the comparison between the next historical vehicle operating state data and the current vehicle operating state data is continued.

[0087] In step S230 of some embodiments, the current vehicle operating state data and the candidate operating state data may be used as target operating state data, or only one of the current vehicle operating state data and the candidate operating state data may be used as the target operating state data.

[0088] In some embodiments, in step S240, a preset Bayesian regression model is updated based on the target operating state data to obtain a target Bayesian regression model, so that the target Bayesian regression model can adapt to real-time environmental changes. The preset Bayesian model is a model obtained by modeling the model uncertainty of the cruising fleet based on historical vehicle operating state data.

[0089] In the above steps S210 to S240, by screening the samples, historical samples with a high degree of correlation with the current sample can be obtained, so that the target Bayesian regression model capable of predicting disturbances in real time can be obtained by using the samples with a high degree of correlation.

[0090] See also Figure 3In some embodiments, the preset Bayesian regression model includes a first initial kernel matrix and a second initial kernel matrix, where the second initial kernel matrix is ​​the inverse matrix of the first initial kernel matrix. Step S240 may include, but is not limited to, steps S310 to S350:

[0091] Step S310, determining a first kernel representation matrix according to the target operating state data and the historical vehicle operating state data;

[0092] Step S320, performing transposition processing on the first kernel representation matrix to obtain a second kernel representation matrix;

[0093] Step S330 , based on the Woodbury matrix identity, the first initial kernel matrix is ​​updated according to the first kernel representation matrix and the second kernel representation matrix to obtain a first target kernel matrix;

[0094] Step S340, updating the second initial kernel matrix according to the first kernel representation matrix and the second kernel representation matrix to obtain a second target kernel matrix;

[0095] Step S350: determining a target Bayesian regression model according to the first target kernel matrix and the second target kernel matrix.

[0096] In step S310 of some embodiments, in Bayesian learning, the kernel matrix used to characterize the similarity between data needs to be updated with the arrival of new data to achieve incremental learning. At the same time, in order to ensure the real-time performance of disturbance prediction, the first initial kernel matrix and the second initial kernel matrix need to be calculated incrementally in real time. The kernel function of the preset Bayesian regression model can adopt the RBF kernel. Based on the kernel function, the similarity between the target operating state data (new data point) and the historical vehicle operating state data (existing data point) is calculated to obtain a first kernel representation matrix. Alternatively, the similarity between the current vehicle operating state data and the historical vehicle operating state data is calculated to obtain a first kernel representation matrix. Alternatively, the similarity between the current vehicle operating state data and the candidate operating state data is calculated to obtain a first kernel representation matrix.

[0097] In step S320 of some embodiments, a matrix transposition process is performed on the first kernel representation matrix to obtain a second kernel representation matrix. The second kernel representation matrix is ​​also used to represent the similarity between the new data point and the existing data point.

[0098] In step S330 of some embodiments, when the amount of data is large, directly calculating the updated second initial kernel matrix is ​​computationally intensive. In this embodiment, the Woodbury identity is used to reduce the computational burden. The Woodbury matrix identity is expressed as:

[0099] det(A+BC)=det(A)·det(C)·det(I+CA -1 B),

[0100] Among them, det represents the determinant; A is the original kernel matrix, that is, the first initial kernel matrix; A -1 is the second initial kernel matrix; B is the column vector consisting of the kernel function calculation results between the new data point and the existing data point, i.e., the first kernel representation matrix. This column vector can be the kernel representation of a new observation, i.e., the kernel function calculation results between the new observation and all data points in the existing dataset. Its dimension is 1×n, where n is the number of data samples in the existing dataset. C is the second kernel representation matrix, a vector or matrix matching the dimension of B. It is the kernel matrix consisting of the kernel function calculation results between the new data point and the existing data point; I is the identity matrix.

[0101] It should be noted that the Woodbury identity is related to confidence, reflecting the size of the covariance matrix update operation. Since the determinant of the covariance matrix can be viewed as the product of the uncertainties in all eigenvector directions, the larger the determinant, the greater the model's uncertainty about the data, and conversely, the smaller the determinant, the smaller the model's uncertainty about the data.

[0102] A high-speed cruising fleet generates new data every moment. Its data stream updates quickly and has high data dimensions. In large-scale data computing scenarios, to reduce the complexity of kernel matrix updates, the present embodiment introduces the Woodbury matrix identity. Based on the Woodbury matrix identity, the first initial kernel matrix is ​​updated according to the first kernel representation matrix and the second kernel representation matrix to obtain the first target kernel matrix. Through the Woodbury matrix identity, the kernel matrix and its inverse matrix can be incrementally updated in real time using only new data points, without the need to recalculate the entire new kernel matrix and its inverse, thus avoiding complex matrix calculations.

[0103] After incrementally updating the kernel matrix (the first initial kernel matrix) and its inverse (the second initial kernel matrix), the target Bayesian regression model can use the new kernel matrix to make perturbation predictions and provide confidence in the perturbation predictions. Because the kernel matrix and its inverse directly affect the calculation of the predictive distribution, quickly and efficiently updating these matrices using the Woodbury matrix identity ensures that the model can still provide reliable predictions and confidence assessments after receiving new data.

[0104] In step S340 of some embodiments, based on the Woodbury matrix identity, the second initial kernel matrix is ​​updated according to the first kernel representation matrix and the second kernel representation matrix to obtain a second target kernel matrix.

[0105] In step S350 of some embodiments, the first target kernel matrix is ​​used as the kernel matrix of the preset Bayesian regression model, and the second target kernel matrix is ​​used as the inverse of the kernel matrix of the preset Bayesian regression model to obtain the target Bayesian regression model.

[0106] In the above steps S310 to S350, in order to make real-time predictions of environmental disturbances, the Bayesian regression model needs to continuously adapt to new sensor input data to update its prediction of the environmental state in real time. The Woodbury matrix identity enables small incremental processing of data streams and dynamic updating of the Bayesian regression model, rather than processing the entire data set at one time, thereby improving computational efficiency. Through this incremental update mechanism, the Bayesian regression model can effectively process large-scale data sets while maintaining high efficiency and high accuracy, which is crucial for vehicle systems that require rapid response. In the presence of uncertain interference from aerodynamics and road friction in the environment, incremental and active Bayesian learning is used to reduce the algorithmic complexity of Bayesian learning, thereby ensuring real-time learning of uncertainty.

[0107] See also Figure 4 In some embodiments, step S330 may include but is not limited to steps S410 to S440:

[0108] Step S410: performing matrix multiplication on the first kernel representation matrix and the second kernel representation matrix to obtain a first intermediate matrix;

[0109] Step S420: performing matrix addition on the first initial kernel matrix and the first intermediate matrix to obtain a first target kernel matrix.

[0110] In step S410 of some embodiments, a first core representation matrix and a second core representation matrix are matrix multiplied to obtain a first intermediate matrix. If the first core representation matrix is ​​represented as B and the second core representation matrix is ​​represented as C, then the first intermediate matrix is ​​represented as B·C, where · represents a matrix multiplication operation.

[0111] In step S420 of some embodiments, the first initial kernel matrix A and the first intermediate matrix B·C are added to obtain a first target kernel matrix.

[0112] Through the above steps S410 to S420, the original kernel matrix can be updated in an incremental learning manner, so that the target Bayesian regression model can estimate the environmental disturbance in real time based on the original kernel matrix.

[0113] See also Figure 5 In some embodiments, step S340 may include but is not limited to steps S510 to S530:

[0114] Step S510, performing a first matrix calculation based on the first kernel representation matrix, the second kernel representation matrix, and the second initial kernel matrix to obtain a second intermediate matrix;

[0115] Step S520, performing a second matrix calculation according to the first kernel representation matrix, the second kernel representation matrix and the second initial kernel matrix to obtain a third intermediate matrix;

[0116] Step S530, updating the second initial kernel matrix according to the second intermediate matrix and the third intermediate matrix to obtain a second target kernel matrix.

[0117] In step S510 of some embodiments, the second initial kernel matrix, the first kernel representation matrix, the second kernel representation matrix and the second initial kernel matrix are multiplied to obtain a second intermediate matrix. The second intermediate matrix is represented as: A -1 ·B·C·A -1 wherein A -1 is the second initial kernel matrix, -1 represents matrix inversion, B represents the first kernel representation matrix, and C represents the second kernel representation matrix.

[0118] In step S520 of some embodiments, the second kernel representation matrix, the second initial kernel matrix and the first kernel representation matrix are multiplied, and a unit matrix is added to the result of the multiplication to obtain a third intermediate matrix. The third intermediate matrix is represented as: I+C·A -1 ·B, I is a unit matrix.

[0119] In step S530 of some embodiments, for each element in the matrix, an element in the second intermediate matrix is divided by an element at a corresponding position in the third intermediate matrix to obtain a fourth intermediate matrix. The second initial kernel matrix is subtracted from the fourth intermediate matrix to obtain the second target kernel matrix.

[0120] Through the above steps S510 to S530, the inverse of the original kernel matrix can be updated in real time in an incremental learning manner, the complexity of matrix calculation is reduced, and the time cost required for model learning is reduced, so that the target Bayesian regression model can estimate the environmental disturbance based on the inverse of the original kernel matrix in real time.

[0121] In step S130 of some embodiments, in order to ensure the normal operation of the cruise platoon and avoid the collision between the cruise platoon and the surrounding cut-in vehicle, the operating state parameters of the surrounding environment vehicle are obtained, and the operating state parameters include vehicle lateral position, vehicle longitudinal position, vehicle longitudinal speed, vehicle lateral speed, etc. The vehicle lateral position is the horizontal position of the vehicle relative to the center line of the road, the vehicle longitudinal position is the front and rear position of the vehicle relative to the cruise platoon, the vehicle longitudinal speed is the speed along the forward direction of the vehicle, and the vehicle lateral speed is the speed perpendicular to the forward direction of the vehicle. The longitudinal relative distance between the preset cut-in vehicle and the cruise platoon is calculated according to the vehicle longitudinal position, and the longitudinal relative speed between the preset cut-in vehicle and the cruise platoon is calculated according to the vehicle longitudinal speed. The longitudinal relative speed is the difference between the vehicle longitudinal speeds of the preset cut-in vehicle and the cruise platoon in the longitudinal direction of the road.

[0122] Referring to Figure 6 In some embodiments, step S140 can include but is not limited to steps S610 to S620:

[0123] In step S610, the confidence of the environment disturbance data output by the target Bayesian regression model is obtained.

[0124] In step S620, if the confidence is in the preset confidence interval, a safety barrier set is constructed according to the environment disturbance data, the longitudinal relative distance and the longitudinal relative speed.

[0125] In step S610 of some embodiments, the confidence of the environment disturbance data output by the target Bayesian regression model is obtained.

[0126] In step S620 of some embodiments, the preset confidence interval is a 3σ interval calculated according to the prediction mean and the prediction variance. If the confidence is in the preset confidence interval, a random safety barrier set is constructed according to the environment disturbance data, the longitudinal relative distance and the longitudinal relative speed, so as to ensure that the cruise platoon converges from an unsafe state to a safe state asymptotically and stably under the condition that the safety distance is not violated and no collision occurs, and a safe inter-vehicle distance is maintained with the cut-in vehicle.

[0127] Through steps S610 to S620 described above, the safety barrier set can be obtained, and the cruise platoon is controlled based on the safety barrier set to cope with the cut-in behavior of the surrounding vehicle, thereby ensuring the safety of the driving process.

[0128] Referring to Figure 7 In some embodiments, step S620 can include but is not limited to steps S710 to S720:

[0129] Step S710: If the confidence level is within the preset confidence interval, the target relative distance is determined based on the environmental disturbance data, the longitudinal relative distance, the longitudinal relative speed, and the preset fleet system response time;

[0130] Step S720: construct a safety obstacle set based on the target relative distance and a preset safety distance threshold.

[0131] In step S710 of some embodiments, the preset platoon system response time is the response time of the cruising platoon to a lane-cutting behavior. The safety obstacle set is constructed based on a control obstacle function. If the confidence level is within a preset confidence interval, the control obstacle function is used to calculate the environmental disturbance data, the longitudinal relative distance, the longitudinal relative speed, and the preset platoon system response time to obtain a target relative distance. The target relative distance is the longitudinal relative distance between the preset lane-cutting vehicle and the cruising platoon under the influence of the surrounding environmental disturbance. The control obstacle function is expressed as:

[0132] h=δp i -δv i ·t r +δd,

[0133] Where h is the relative distance to the target; δp i is the longitudinal relative distance; δv i is the longitudinal relative velocity; t r represents the preset fleet system response time; δd is the environmental disturbance data, which represents the distance disturbance caused by disturbances such as road gradient, friction, and wind disturbance in the longitudinal relative distance.

[0134] In step S720 of some embodiments, a constraint condition is constructed based on the target relative distance and the preset safety distance threshold, and the constraint condition is used as a safety obstacle set. The safety obstacle set is defined as: S = {x|h(x)>0}, where x represents the current vehicle operating state data, h(x) = δp i -δv i ·t r +δd-d safe , d safe The preset safety distance threshold.

[0135] When surrounding vehicles maneuver to cut in, the cut-in vehicle may enter the safety obstacle set S, such that h < 0. To effectively prevent potential collisions, it is necessary to ensure that the state of the cut-in vehicle smoothly converges to and remains within the safety obstacle set S. The safety obstacle set explicitly addresses the problem of safe recovery after a cut-in.

[0136] Through the above steps S710 to S720, a safety obstacle set can be obtained, so that the cruising fleet can deal with the vehicle cutting in based on the safety obstacle set.

[0137] In step S150 of some embodiments, the safety obstacle set is transformed to obtain linear constraints for processing the restoration of the safe distance for lane cutting, and the control quantity is generated according to the linear constraints. The linear constraints are used as constraints for the quadratic optimization problem:

[0138] δh(x):=h(x k+1 )>(1-α)h(x k ),

[0139] Among them, k is the current time, x k Indicates the current vehicle operating status; x k+1 represents the vehicle's operating state at the next moment; α∈(0,1) is the gradient factor that adjusts the recovery speed, which can be used to adjust the speed at which the function h changes from negative to positive values. A smaller α indicates a slower recovery speed and a smoother speed adjustment for the vehicle that cuts in; a larger α indicates a faster recovery speed.

[0140] Linear constraints leverage the uncertain perturbations output by the Bayesian regression model to enable an uncertain nonlinear vehicle system to reach a desired state, such as a desired cruising speed. Integrating linear constraints into a real-time, high-frequency, adaptive connected cruise control system, operating at a frequency of 50 Hz or higher, enables the cruise fleet to rapidly respond to sudden lane-cutting maneuvers by surrounding vehicles. This allows the fleet to adapt to external environmental disturbances in real time, ensuring that the fleet can quickly and stably maintain a safe distance from surrounding vehicles that change lanes and cut in. This ensures not only driving safety but also overall fleet efficiency.

[0141] It should be noted that the connected cruise control system repairs the predicted trajectory of the connected cruise control system in a minimally intervening manner to cope with the surrounding vehicle's cutting-in behavior. The objective function of the quadratic optimization problem can be the deviation between the predicted trajectory and the original trajectory, that is, the predicted position of the vehicle at the next moment (x k+1 The deviation between the vehicle position contained in and the original arrival position at the next moment is quantified using the bi-norm, ensuring the platoon's adaptability to the environment while minimizing the impact on driving efficiency. The predicted vehicle position is the platoon's vehicle position predicted based on linear constraints, while the original arrival position is the position the platoon would have arrived at had no lane-cutting occurred.

[0142] Control variables are variables that influence the vehicle's operating state, such as throttle position and steering angle. After applying these control variables to adjust vehicle spacing, sensors can continuously collect data, interact with the environment in real time, maintain a safe distance from vehicles cutting in, and update the Bayesian regression model, further enhancing the incremental learning process.

[0143] In step S160 of some embodiments, the vehicle motion trajectory is adjusted according to the control amount, thereby adjusting the vehicle spacing of the cruising fleet. The above steps S110 to S160 are repeated for each vehicle in the cruising fleet until the entire connected cruising fleet is completed. The preset vehicle that cuts in is added to the cruising fleet according to the vehicle spacing control.

[0144] Although deep reinforcement learning methods can improve the adaptability of autonomous vehicles in complex traffic scenarios, they still have shortcomings in solving safety issues in the case of model disturbances and cutting in, and the deep reinforcement learning process is usually not explainable. In complex traffic environments, understanding the logic and reasons behind decisions is crucial to ensuring system safety. In contrast, deep reinforcement learning is usually considered a black box model, and its decision-making process is difficult to explain. The cruise control method of the embodiment of the present application not only improves the safety and task execution efficiency of the interconnected cruise system in the case of model disturbances and surrounding vehicles cutting in, but also focuses on explainability. The use of an efficient and online Bayesian learning method based on incremental learning and active learning can effectively utilize the real-time interaction data between the vehicle and the environment to quantify the uncertainty of environmental disturbances, so as to learn the model disturbances of the external environment on the system in real time, making the system's decision-making and control processes more transparent and understandable, which helps to improve the system's explainability and make decisions and controls more credible. At the same time, it ensures that the system can respond safely and efficiently in complex traffic environments, allowing the vehicle to safely and quickly return to a safe state and maintain an appropriate distance from surrounding vehicles, effectively avoiding the risk of collision and improving the overall safety of autonomous driving. At the same time, it enables the cruising fleet to adapt to changes in the surrounding environment more quickly, maintaining the high efficiency of fleet transportation, and is particularly suitable for areas such as centralized freight transportation and alleviating traffic congestion.

[0145] The present application also provides a cruise control device that can implement the above cruise control method. The cruise control device includes:

[0146] A first acquisition module 810 is used to acquire current vehicle operation status data of the cruising fleet;

[0147] The disturbance prediction module 820 is used to input the current vehicle operating state data into the target Bayesian regression model to perform disturbance prediction and obtain environmental disturbance data;

[0148] The second acquisition module 830 is used to obtain the longitudinal relative distance and longitudinal relative speed between the preset lane-cutting vehicle and the cruising convoy;

[0149] A construction module 840 is used to construct a safety obstacle set based on environmental disturbance data, longitudinal relative distance, and longitudinal relative speed;

[0150] a conversion module 850 for converting the safety obstacle set into linear constraints and generating control quantities according to the linear constraints;

[0151] The cruise control module 860 is used to adjust the distance between vehicles in the cruise convoy according to the control amount, and to control the preset vehicles that cut in to join the cruise convoy according to the distance between vehicles.

[0152] The specific implementation of the cruise control device is substantially the same as the specific embodiment of the cruise control method described above, and will not be described in detail herein.

[0153] The present application also provides an electronic device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the above-mentioned cruise control method when executing the computer program. The electronic device can be any smart terminal including a tablet computer, an in-vehicle computer, or the like.

[0154] See also Figure 9 , Figure 9 The hardware structure of an electronic device according to another embodiment is shown. The electronic device includes:

[0155] The processor 910 may be implemented as a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application.

[0156] The memory 920 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 920 can store an operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program codes are stored in the memory 920 and are called by the processor 910 to execute the cruise control method of the embodiments of this application.

[0157] Input / output interface 930, used to implement information input and output;

[0158] Communication interface 940, used to implement communication interaction between this device and other devices, which can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WiFi, Bluetooth, etc.);

[0159] bus 950 , which transmits information between various components of the device (e.g., processor 910 , memory 920 , input / output interface 930 , and communication interface 940 );

[0160] The processor 910 , the memory 920 , the input / output interface 930 , and the communication interface 940 are connected to each other in communication within the device via a bus 950 .

[0161] An embodiment of the present application further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned cruise control method is implemented.

[0162] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely located relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0163] The embodiments described in the embodiments of this application are intended to more clearly illustrate the technical solutions of the embodiments of this application and do not constitute a limitation on the technical solutions provided by the embodiments of this application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0164] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0165] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, i.e., they may be located in one place or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of this embodiment.

[0166] Those skilled in the art will appreciate that all or some of the steps in the methods, systems, and functional modules / units in the devices disclosed above may be implemented as software, firmware, hardware, or appropriate combinations thereof.

[0167] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0168] It should be understood that in this application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the previous and next associated objects are in an "or" relationship. "At least one of the following items" or similar expressions refers to any combination of these items, including any combination of single items or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0169] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the above-mentioned units is only a logical function division. In actual implementation, there may be other division methods, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. The mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0170] The units described above as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0171] In addition, the functional units in the various embodiments of the present application may be integrated into a single processing unit, or each unit may exist physically separately, or two or more units may be integrated into a single unit. The aforementioned integrated units may be implemented in the form of hardware or software functional units.

[0172] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, which is stored in a storage medium and includes multiple instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: various media that can store programs, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.

[0173] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but are not intended to limit the scope of the present invention. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and essence of the present invention should be within the scope of the present invention.

Claims

1. A cruise control method, characterized in that: The method comprises: Obtain the current vehicle operation status data of the cruise fleet; Inputting the current vehicle operating state data into a target Bayesian regression model for disturbance prediction to obtain environmental disturbance data; Obtaining a longitudinal relative distance and a longitudinal relative speed between a preset lane-cutting vehicle and the cruising convoy; constructing a safety obstacle set according to the environmental disturbance data, the longitudinal relative distance, and the longitudinal relative speed; Converting the safety obstacle set into linear constraints and generating control quantities according to the linear constraints; The vehicle spacing of the cruising convoy is adjusted according to the control amount, and the preset jam-in vehicle is controlled to join the cruising convoy according to the vehicle spacing.

2. The cruise control method according to claim 1, characterized in that: Before inputting the current vehicle operating state data into the target Bayesian regression model for disturbance prediction to obtain environmental disturbance data, the cruise control method further includes: Calculating the similarity between the current vehicle operating state data and pre-acquired historical vehicle operating state data; Screening the historical vehicle operating status data according to the similarity to obtain candidate operating status data; determining target operating state data according to the current vehicle operating state data and the candidate operating state data; The preset Bayesian regression model is updated according to the target operating state data to obtain the target Bayesian regression model.

3. The cruise control method according to claim 2, characterized in that: The preset Bayesian regression model includes a first initial kernel matrix and a second initial kernel matrix, where the second initial kernel matrix is ​​the inverse matrix of the first initial kernel matrix. The preset Bayesian regression model is updated according to the target operating state data to obtain the target Bayesian regression model, including: determining a first kernel representation matrix according to the target operating state data and the historical vehicle operating state data; performing a transposition process on the first kernel representation matrix to obtain a second kernel representation matrix; Based on the Woodbury matrix identity, the first initial kernel matrix is ​​updated according to the first kernel representation matrix and the second kernel representation matrix to obtain a first target kernel matrix; updating the second initial kernel matrix according to the first kernel representation matrix and the second kernel representation matrix to obtain a second target kernel matrix; The target Bayesian regression model is determined according to the first target kernel matrix and the second target kernel matrix.

4. The cruise control method according to claim 3, characterized in that: The updating of the first initial kernel matrix according to the first kernel representation matrix and the second kernel representation matrix to obtain a first target kernel matrix includes: Performing matrix multiplication on the first kernel representation matrix and the second kernel representation matrix to obtain a first intermediate matrix; Perform matrix addition on the first initial kernel matrix and the first intermediate matrix to obtain the first target kernel matrix.

5. The cruise control method according to claim 3, characterized in that: The updating of the second initial kernel matrix according to the first kernel representation matrix and the second kernel representation matrix to obtain a second target kernel matrix includes: Perform a first matrix calculation based on the first kernel representation matrix, the second kernel representation matrix, and the second initial kernel matrix to obtain a second intermediate matrix; Perform a second matrix calculation based on the first kernel representation matrix, the second kernel representation matrix, and the second initial kernel matrix to obtain a third intermediate matrix; The second initial kernel matrix is ​​updated according to the second intermediate matrix and the third intermediate matrix to obtain the second target kernel matrix.

6. The cruise control method according to any one of claims 1 to 5, characterized in that: The constructing of a safety obstacle set according to the environmental disturbance data, the longitudinal relative distance, and the longitudinal relative speed includes: Obtaining confidence of the environmental disturbance data; If the confidence level is within a preset confidence interval, the safety obstacle set is constructed according to the environmental disturbance data, the longitudinal relative distance, and the longitudinal relative speed.

7. The cruise control method according to claim 6, characterized in that: If the confidence level is within a preset confidence interval, constructing the safety obstacle set according to the environmental disturbance data, the longitudinal relative distance, and the longitudinal relative speed, including: If the confidence level is within a preset confidence interval, determining a target relative distance based on the environmental disturbance data, the longitudinal relative distance, the longitudinal relative speed, and a preset fleet system response time; The safety obstacle set is constructed according to the target relative distance and a preset safety distance threshold.

8. A cruise control device, characterized in that: The device comprises: The first acquisition module is used to obtain the current vehicle operation status data of the cruising fleet; A disturbance prediction module is used to input the current vehicle operating state data into a target Bayesian regression model to perform disturbance prediction and obtain environmental disturbance data; A second acquisition module is used to acquire a longitudinal relative distance and a longitudinal relative speed between a preset lane-cutting vehicle and the cruising convoy; A construction module, configured to construct a safety obstacle set according to the environmental disturbance data, the longitudinal relative distance, and the longitudinal relative speed; a conversion module, configured to convert the safety obstacle set into a linear constraint and generate a control variable according to the linear constraint; A cruise control module is configured to adjust the vehicle spacing of the cruising fleet according to the control amount, and control the preset jam-in vehicle to join the cruising fleet according to the vehicle spacing.

9. An electronic device, characterized in that: The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the cruise control method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the cruise control method according to any one of claims 1 to 7 is implemented.

Images