Steering decision model optimization, steering decision method, device, equipment and medium

Abstract

The application discloses a kind of steering decision model optimization, steering decision method, device, equipment and medium.The optimization method of steering decision model includes: obtaining the vehicle steering decision model to be optimized and the standard verification sample set matched with vehicle steering decision model;PCA dimension reduction processing is carried out to standard verification sample set, and dimension reduction verification sample set is obtained;Parameter is set based on bat algorithm, and multiple rounds of iteration processing of bat colony is carried out using bat algorithm;When meeting bat iteration end condition, the final fitness value of steering decision model for each bat is calculated according to dimension reduction verification sample set, and global optimal bat is selected according to final fitness value;Global optimal bat is used to update vehicle steering decision model, and the optimized vehicle steering decision model is obtained.Through using the above technical scheme, the data processing time in the optimization process of vehicle steering decision model can be reduced, and the accuracy of the optimized vehicle steering decision model can be effectively improved.

Description

Technical Field

[0001] This invention relates to the field of autonomous driving technology, and in particular to a steering decision model optimization, steering decision method, device, equipment and medium. Background Technology

[0002] In the process of autonomous driving, in order to ensure the smoothness and safety of the vehicle's movement, it is necessary to determine whether the vehicle needs to be steered based on the current driving status of the vehicle and the surrounding environment.

[0003] In existing technologies, classifiers can be used for event decision-making in various scenarios, but existing classifiers cannot be directly used to obtain vehicle steering decisions based on vehicle-related information. Therefore, it is important to optimize existing classifiers to obtain optimal vehicle steering decisions. Summary of the Invention

[0004] This invention provides a steering decision model optimization, steering decision method, device, equipment, and medium, which can reduce data processing time during the vehicle steering decision model optimization process and effectively improve the accuracy of the optimized vehicle steering decision model.

[0005] According to one aspect of the present invention, an optimization method for a vehicle steering decision model is provided, comprising:

[0006] Obtain the vehicle steering decision model to be optimized and the standard validation sample set that matches the vehicle steering decision model. The standard validation sample includes multiple vehicle environment description feature values ​​and standard steering decision results.

[0007] The standard validation sample set is subjected to PCA (Principal Component Analysis) dimensionality reduction to obtain the dimensionality-reduced validation sample set;

[0008] Based on the parameter setting of the bat algorithm, the bat algorithm is used to perform multi-round iterative processing of bat swarms;

[0009] In each iteration, the vehicle steering decision model is updated using bats that match the current iteration. The fitness of inputting the dimensionality reduction validation sample set into each updated vehicle steering decision model is used as the evaluation index. Locally optimal bats are identified among the bats, and the bat swarm iterative optimization of the current iteration is performed using locally optimal bats. The bats include a penalty factor and a Gaussian kernel function.

[0010] When the bat iteration termination condition is met, the final fitness value of the turning decision model for each bat is calculated based on the dimensionality reduction verification sample set, and the globally optimal bat is selected based on the final fitness value.

[0011] The vehicle steering decision model is updated using the globally optimal bat, resulting in an optimized vehicle steering decision model.

[0012] According to another aspect of the present invention, a vehicle steering decision method is provided, executed by an in-vehicle infotainment system, comprising:

[0013] By collecting information from multiple sensors installed in the vehicle, multiple vehicle environment description feature values ​​corresponding to at least one associated vehicle related to the vehicle and the vehicle's steering decision are obtained.

[0014] The vehicle environment description features include the current vehicle speed and acceleration of the vehicle and related vehicles, the current left front coordinate of the vehicle and related vehicles, the current front longitudinal coordinate of the vehicle and related vehicles, the current vehicle width and length of the vehicle and related vehicles, the distance between the vehicle and at least one vehicle in front, and the distance between the vehicle and at least one vehicle behind.

[0015] The multiple vehicle environment description feature values ​​are input into the optimized vehicle steering decision model obtained by the method described in either Embodiment 1 or Embodiment 2 of the present invention to obtain the steering decision result.

[0016] Perform vehicle control operations that match the steering decision result.

[0017] According to another aspect of the present invention, an optimization apparatus for a vehicle steering decision model is provided, comprising:

[0018] The standard validation sample set acquisition module is used to acquire the vehicle steering decision model to be optimized and the standard validation sample set that matches the vehicle steering decision model. The standard validation sample includes multiple vehicle environment description feature values ​​and standard steering decision results.

[0019] The dimensionality reduction validation sample set acquisition module is used to perform PCA dimensionality reduction on the standard validation sample set to obtain the dimensionality reduction validation sample set.

[0020] The bat iteration module is used to set parameters based on the bat algorithm and perform multi-round iterative processing of bat swarms using the bat algorithm.

[0021] In each iteration, the vehicle steering decision model is updated using bats that match the current iteration. The fitness of inputting the dimensionality reduction validation sample set into each updated vehicle steering decision model is used as the evaluation index. Locally optimal bats are identified among the bats, and the bat swarm iterative optimization of the current iteration is performed using locally optimal bats. The bats include a penalty factor and a Gaussian kernel function.

[0022] The optimal bat selection module is used to calculate the final fitness value of each bat for the turning decision model based on the dimensionality reduction verification sample set when the bat iteration termination condition is met, and select the global optimal bat based on the final fitness value.

[0023] The vehicle steering decision model optimization module is used to update the vehicle steering decision model using the global optimal bat, resulting in an optimized vehicle steering decision model.

[0024] According to another aspect of the present invention, a vehicle steering decision device is provided, executed by an in-vehicle system, comprising:

[0025] The vehicle environment description feature value acquisition module is used to acquire multiple vehicle environment description feature values ​​corresponding to at least one associated vehicle related to the vehicle and the vehicle steering decision by using the information collection results of multiple sensors installed in the vehicle.

[0026] The vehicle environment description features include the current vehicle speed and acceleration of the vehicle and related vehicles, the current left front coordinate of the vehicle and related vehicles, the current front longitudinal coordinate of the vehicle and related vehicles, the current vehicle width and length of the vehicle and related vehicles, the distance between the vehicle and at least one vehicle in front, and the distance between the vehicle and at least one vehicle behind.

[0027] The steering decision result acquisition module is used to input the multiple vehicle environment description feature values ​​into the optimized vehicle steering decision model obtained by the method described in either Embodiment 1 or Embodiment 2 of the present invention, and to obtain the steering decision result.

[0028] The vehicle control module is used to perform vehicle control operations that match the steering decision result.

[0029] According to another aspect of the present invention, an electronic device is provided, the electronic device comprising:

[0030] At least one processor; and

[0031] A memory communicatively connected to the at least one processor; wherein,

[0032] The memory stores a computer program that can be executed by the at least one processor. The computer program is executed by the at least one processor to enable the at least one processor to execute the optimization method of the vehicle steering decision model described in Embodiments 1 and 2 of the present invention, or to implement the vehicle steering decision method described in Embodiment 3 of the present invention.

[0033] According to another aspect of the present invention, a computer-readable storage medium is provided, the computer-readable storage medium storing computer instructions, the computer instructions being configured to cause a processor to execute and implement the optimization method of the vehicle steering decision model described in Embodiments 1 and 2 of the present invention, or to implement the vehicle steering decision method described in Embodiment 3 of the present invention.

[0034] The technical solution of this invention optimizes the vehicle steering decision model by reducing the dimensionality of the standard validation sample set and using the reduced dimensionality validation sample set to perform multiple iterations on the bat swarm based on the bat algorithm. This reduces the data processing time during the optimization process and effectively improves the accuracy of the optimized vehicle steering decision model.

[0035] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0036] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0037] Figure 1 This is a flowchart of an optimization method for a vehicle steering decision model according to Embodiment 1 of the present invention;

[0038] Figure 2 This is a flowchart of another optimization method for a vehicle steering decision model provided in Embodiment 2 of the present invention;

[0039] Figure 3a This is a flowchart of a vehicle steering decision method provided in Embodiment 3 of the present invention;

[0040] Figure 3b This is a vehicle position relationship diagram provided according to Embodiment 3 of the present invention;

[0041] Figure 4 This is a schematic diagram of the structure of an optimization device for a vehicle steering decision model according to Embodiment 4 of the present invention;

[0042] Figure 5 This is a schematic diagram of the structure of a vehicle steering decision device according to Embodiment 5 of the present invention;

[0043] Figure 6This is a schematic diagram of the structure of the electronic device that implements the vehicle steering decision model optimization method and the vehicle steering decision method according to embodiments of the present invention. Detailed Implementation

[0044] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0045] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0046] Example 1

[0047] Figure 1 This is a flowchart of an optimization method for a vehicle steering decision model provided in Embodiment 1 of the present invention. This embodiment is applicable to situations where PCA dimensionality reduction is performed on standard validation samples, and the vehicle steering decision model is optimized using the dimensionality-reduced validation sample set based on the Bat algorithm. This method can be executed by a vehicle steering decision model optimization device, which can be implemented in hardware and / or software, and can be configured in a computer with data processing capabilities. Figure 1 As shown, the method includes:

[0048] S110. Obtain the vehicle steering decision model to be optimized and a standard validation sample set that matches the vehicle steering decision model. The standard validation sample includes multiple vehicle environment description feature values ​​and standard steering decision results.

[0049] Preferably, in this embodiment of the invention, SVM (Support Vector Machine) is selected as the vehicle steering decision model. SVM is a type of support vector machine that can solve binary classification problems.

[0050] Among them, the vehicle environment description feature value is the input data of the vehicle steering decision model. After the vehicle environment description feature is input into the vehicle steering decision model, the vehicle steering decision model can output the corresponding decision result.

[0051] In the standard validation sample set, each standard validation sample may include multiple vehicle environment description features and standard steering decision results. The vehicle environment description features may include the current vehicle speed and acceleration of the current steering vehicle and all related vehicles, the current left-front coordinate of the current steering vehicle and all related vehicles, the current front ordinate of the current steering vehicle and all related vehicles, the current vehicle width and length of the current steering vehicle and all related vehicles, the distance between the current steering vehicle and at least one vehicle in front, and the distance between the current steering vehicle and at least one vehicle behind. The standard steering decision results may include lane change accuracy and lane keeping accuracy.

[0052] S120. Perform PCA dimensionality reduction on the standard validation sample set to obtain the dimensionality-reduced validation sample set.

[0053] Among them, PCA dimensionality reduction processing analyzes the important features in each vehicle environment description feature value in the standard verification sample, filters out redundant information, and realizes the transformation of vehicle environment description features from high dimension to low dimension.

[0054] The dimensionality reduction validation sample set can include dimensionality-reduced vehicle environment description feature values ​​and standard steering decision results.

[0055] The advantage of this setup is that by performing PCA dimensionality reduction on the standard validation sample set, the noise of the standard validation samples can be reduced, which can effectively reduce data processing time and improve the optimization efficiency of the vehicle steering decision model.

[0056] S130. Based on the bat algorithm, set parameters and use the bat algorithm to perform multi-round iterative processing of the bat swarm.

[0057] The parameters for the bat algorithm can include the number of bats in the colony, the maximum number of iterations, the initial position of the bats, the initial speed of the bats, the fitness function, the iteration elimination rate, and the position iteration function.

[0058] In each iteration, the vehicle steering decision model is updated using bats that match the current iteration. The fitness of the dimensionality-reduced validation sample set input to each updated vehicle steering decision model is used as the evaluation index. Locally optimal bats are identified among the bats, and the bat swarm optimization for the current iteration is performed using locally optimal bats. The bats include a penalty factor and a Gaussian kernel function.

[0059] It is understood that the bats proposed in this embodiment of the invention include a penalty factor and a Gaussian kernel function, meaning that the bat coordinates are composed of the penalty factor and the Gaussian kernel function. The process of iteratively applying the bat algorithm to a bat swarm can be understood as an optimization process of the penalty factor and the Gaussian kernel function. Each bat swarm can contain multiple bats, meaning each bat swarm can represent multiple sets of different penalty factors and Gaussian kernel functions.

[0060] In the traditional bat algorithm iteration process, the number of bats in the bat swarm, the initial position of each bat, the initial speed of each bat, and the objective function are first preset; then, the position and speed of each bat are iterated by learning from the best bat in each iteration; when the iteration termination condition is met, the iteration stops and the best bat is selected.

[0061] This invention addresses the issue that when optimizing a vehicle steering decision model using the traditional bat algorithm, the bat population tends to gravitate towards the individual bat with the highest fitness value to achieve a higher fitness score. However, this gravitation can lead to decreased population diversity and a tendency to fall into local optima. Therefore, this invention creatively proposes replacing the single individual bat with the highest fitness value with multiple individual bats when the bat population gravitates towards the high-fitness bat. Simultaneously, during each iteration of the bat population, a preset elimination rate is used to eliminate some bats with lower fitness values, and the same number of new bats are added to the population.

[0062] S140. When the bat iteration termination condition is met, calculate the final fitness value of the turning decision model for each bat based on the dimensionality reduction verification sample set, and select the globally optimal bat based on the final fitness value.

[0063] It should be noted that there is a correlation between the fitness value of the bat and the accuracy of the vehicle steering decision model; that is, the higher the fitness value of the bat, the higher the accuracy of the vehicle steering decision model.

[0064] In one specific implementation, the bat iteration termination condition can be considered met when the preset maximum number of iterations has been reached and there are bats in the swarm with fitness values ​​greater than the preset minimum fitness value. Optionally, random numbers can also be generated, and the relationship between the random numbers and the pulse emission rate can be used as one of the bat iteration termination conditions.

[0065] S150. The vehicle steering decision model is updated using the global optimal bat to obtain the optimized vehicle steering decision model.

[0066] Once the globally optimal bat is obtained, the optimal penalty factor and the optimal Gaussian kernel function can be obtained. By inputting the optimal penalty factor and the optimal Gaussian kernel function into the vehicle steering decision model, the optimized vehicle steering decision model can be obtained.

[0067] The technical solution of this invention optimizes the vehicle steering decision model by reducing the dimensionality of the standard validation sample set and using the reduced dimensionality validation sample set to perform multiple iterations on the bat swarm based on the bat algorithm. This reduces the data processing time during the optimization process and effectively improves the accuracy of the optimized vehicle steering decision model.

[0068] Example 2

[0069] Figure 2 This is a flowchart of another optimization method for a vehicle steering decision model provided in Embodiment 2 of the present invention. Based on the above embodiments, this embodiment specifically illustrates the optimization method for the vehicle steering decision model. Figure 2 As shown, the method includes:

[0070] S210. Obtain the vehicle steering decision model to be optimized and a standard validation sample set that matches the vehicle steering decision model. The standard validation sample set includes multiple vehicle environment description feature values ​​and standard steering decision results.

[0071] S220. After preprocessing the vehicle environment description feature values ​​in each standard verification sample, PCA is used to reduce the dimensionality of each vehicle environment description feature value to obtain a dimensionality-reduced verification sample set.

[0072] Data preprocessing can include data cleaning, data extraction, and data standardization.

[0073] The advantage of this setup is that by preprocessing the data, the vehicle environment description feature values ​​can be standardized, making the various vehicle environment description feature values ​​more balanced, which can effectively improve the decision accuracy of the vehicle steering decision model.

[0074] S230. Update the vehicle steering decision model to be optimized using the initial position of the bats, input the dimensionality reduction verification sample set into the updated vehicle steering decision model, obtain multiple locally optimal bats in the current iteration round through a preset fitness function, and save the relevant information of the locally optimal bats.

[0075] Preferably, the number of locally optimal bats can be preset, and the bats with the highest fitness can be selected from the bat population in the current iteration based on the number of locally optimal bats.

[0076] The information related to the local optimal bat may include the location of the local optimal bat and its current iteration speed.

[0077] S240. Based on the preset iteration elimination rate and the fitness value of each bat in the current iteration round, eliminate at least one bat in the current bat group and generate new bats according to the preset generation rules, keeping the number of bats in the group unchanged, and iterate on each bat in the current bat group.

[0078] Specifically, based on a preset elimination rate and the fitness value of each bat in the current iteration round, at least one bat in the current bat population is eliminated, and new bats are generated according to preset generation rules, keeping the number of bats in the current population unchanged. Iteration is performed on each bat in the current population, which may specifically include:

[0079] The number of eliminated bats is calculated based on the preset iteration elimination rate, and the fitness value of each bat in the current iteration round is calculated.

[0080] In the current iteration, among all the bats, select at least one bat with the lowest fitness value as the eliminated bat according to the number of eliminated bats.

[0081] New bats are generated according to the preset generation rules to match the number of eliminated bats, and the new bats are added to the bat population. The number of bats in the current bat population remains unchanged, and the process is repeated for each bat in the current bat population.

[0082] The advantage of this setup is that it avoids the situation where the bat population learns from a single optimal bat in the next iteration, which would lead to a decrease in bat population diversity.

[0083] S250. Based on the local optimal bat information from the previous iteration and the preset position iteration function, calculate the optimal flight position for this iteration. Update the positions of each bat after the previous iteration based on the optimal flight position and the initial speed of the bat.

[0084] The optimal flight position for this iteration is calculated based on the local optimal bat information from the previous iteration and a preset position iteration function. This can specifically include:

[0085] Select two learning bats from the best bats in the previous iteration and obtain the current position of each learning bat;

[0086] The optimal flight position for this iteration is calculated based on the preset position iteration function;

[0087] The preset position iteration function is:

[0088]

[0089] Where f1 and f2 are [f min f max The random number f in ] max with f min The preset maximum and minimum fitness values ​​are... Let i be the position of the i-th bat in generation t. For each learning bat's location, X new1 This is the optimal flight position for this iteration.

[0090] If the traditional bat iterative algorithm is used to optimize the vehicle steering decision model, the optimal flight position in each round will tend to converge towards the position of the bat with the highest fitness. This can easily cause the bat population to get trapped in local optima after iteration. This invention creatively proposes to select two learning bats from multiple optimal bats in the bat population, and obtain the optimal flight position for the current iteration based on the position and fitness of the learning bats.

[0091] The advantage of this setup is that it overcomes the problem of the bat swarm tending to converge towards individual bats when optimizing vehicle steering decision models using the traditional bat algorithm, which leads to getting stuck in local extrema.

[0092] S260. Compare the fitness values ​​of each bat after the current iteration with the fitness values ​​of each bat after the previous iteration, save the positions of the bats with higher fitness values ​​in the two iterations, update the relevant parameters required for the next iteration, and proceed with the next iteration until the bat iteration stops when the stopping condition is met.

[0093] It should be noted that after the first iteration, it is necessary to check whether the bat iteration stopping condition is met. The step of determining whether the bat iteration stopping condition is met is mentioned here in this embodiment only for the purpose of understanding the entire bat iteration process, and does not restrict the location of the bat iteration stopping condition determination.

[0094] S270. When the bat iteration termination condition is met, calculate the final fitness value of the turning decision model for each bat based on the dimensionality reduction verification sample set, and select the globally optimal bat based on the final fitness value.

[0095] S280. The vehicle steering decision model is updated using the global optimal bat to obtain the optimized vehicle steering decision model.

[0096] The technical solution of this invention effectively avoids the problem of bat population diversity declining because the bat population learns only from a single optimal bat by eliminating bats with low fitness values ​​and generating new bats to join the bat population. By selecting two learning bats from multiple optimal bats and obtaining the optimal flight position for this iteration, the problem of bat populations getting trapped in local extrema can be overcome.

[0097] Example 3

[0098] Figure 3a This is a flowchart of a vehicle steering decision method provided in Embodiment 3 of the present invention. This embodiment is applicable to collecting multiple vehicle environment description feature values ​​related to vehicle steering decisions and controlling the vehicle based on the steering decision results output by the optimized vehicle steering decision model. This method can be executed by a vehicle steering decision device, which can be implemented in hardware and / or software. The vehicle steering decision device can be configured in a vehicle infotainment system with data processing capabilities. Figure 3a As shown, the method includes:

[0099] S310. By collecting information from multiple sensors installed in the vehicle, obtain multiple vehicle environment description feature values ​​corresponding to at least one associated vehicle related to the vehicle and the vehicle steering decision.

[0100] The vehicle environment description features include the current vehicle speed and acceleration of the vehicle and related vehicles, the current left front coordinate of the vehicle and related vehicles, the current front longitudinal coordinate of the vehicle and related vehicles, the current vehicle width and length of the vehicle and related vehicles, the distance between the vehicle and at least one vehicle in front, and the distance between the vehicle and at least one vehicle behind.

[0101] To facilitate understanding of the positional relationship between this vehicle and related vehicles in practical applications Figure 3b This diagram shows the positional relationship between this vehicle and related vehicles. (For example...) Figure 3b As shown, vehicle M is the vehicle itself, vehicle MP is the vehicle in front of the vehicle, vehicle MB is the vehicle behind the vehicle, vehicle TP is the vehicle in front of the vehicle in the adjacent lane, and vehicle TB is the vehicle behind the vehicle in the adjacent lane.

[0102] In such Figure 3bBased on the vehicle position relationship diagram shown, the vehicle environment description features may include the speeds of vehicles M, MP, MB, TP, and TB; the accelerations of vehicles M, MP, MB, TP, and TB; the lengths of vehicles M, MP, MB, TP, and TB; the widths of vehicles M, MP, MB, TP, and TB; the current left front coordinates of vehicles M, TP, and TB; the current front y-coordinates of vehicles M, TP, and TB; the interval between vehicles M and MP; and the interval between vehicles M and MB.

[0103] S320. Input the multiple vehicle environment description feature values ​​into the optimized vehicle steering decision model to obtain the steering decision result.

[0104] In embodiments of the present invention, the steering decision result may include lane change accuracy and lane keeping accuracy, wherein the one with the highest accuracy is the vehicle control operation that the vehicle system needs to perform.

[0105] S330, Execute vehicle control operations that match the steering decision results.

[0106] The technical solution of this invention collects vehicle environment description feature values ​​by sensors, inputs these feature values ​​into an optimized vehicle steering decision model, and obtains steering decision results to control the vehicle. This method enables the vehicle system to make accurate automatic steering decisions based on the external vehicle environment, thereby improving the intelligence of the vehicle system.

[0107] Example 4

[0108] Figure 4 This is a schematic diagram of the structure of an optimization device for a vehicle steering decision model provided in Embodiment 4 of the present invention. Figure 4 As shown, the device includes: a standard verification sample set acquisition module 410, a dimensionality reduction verification sample set acquisition module 420, a bat iteration module 430, an optimal bat selection module 440, and a vehicle steering decision model optimization module 450.

[0109] The standard validation sample set acquisition module 410 is used to acquire the vehicle steering decision model to be optimized and the standard validation sample set that matches the vehicle steering decision model. The standard validation sample includes multiple vehicle environment description feature values ​​and standard steering decision results.

[0110] The dimension reduction verification sample set acquisition module 420 is used to perform PCA dimension reduction processing on the standard verification sample set to obtain the dimension reduction verification sample set.

[0111] The bat iteration module 430 is used to set parameters based on the bat algorithm and perform multi-round iterative processing of the bat swarm using the bat algorithm.

[0112] In each iteration, the vehicle steering decision model is updated using bats that match the current iteration. The fitness of inputting the dimensionality reduction validation sample set into each updated vehicle steering decision model is used as the evaluation index. Locally optimal bats are identified among the bats, and the bat swarm optimization for the current iteration is performed using locally optimal bats. The bats include a penalty factor and a Gaussian kernel function.

[0113] The optimal bat selection module 440 is used to calculate the final fitness value of each bat in the turning decision model based on the dimensionality reduction verification sample set when the bat iteration termination condition is met, and select the global optimal bat based on the final fitness value.

[0114] The vehicle steering decision model optimization module 450 is used to update the vehicle steering decision model using the global optimal bat to obtain the optimized vehicle steering decision model.

[0115] The technical solution of this invention optimizes the vehicle steering decision model by reducing the dimensionality of the standard validation sample set and using the reduced dimensionality validation sample set to perform multiple iterations on the bat swarm based on the bat algorithm. This reduces the data processing time during the optimization process and effectively improves the accuracy of the optimized vehicle steering decision model.

[0116] Based on the above embodiments, the dimension reduction verification sample set acquisition module 420 can be specifically used to: perform data preprocessing on the vehicle environment description feature values ​​in each standard verification sample, and then use PCA to perform data dimensionality reduction on the vehicle environment description feature values ​​to obtain the dimension reduction verification sample set.

[0117] Based on the above embodiments, the parameters for setting the bat algorithm may include the number of bats in the group, the maximum number of iterations, the initial position of the bats, the initial speed of the bats, the fitness function, the iteration elimination rate, and the position iteration function.

[0118] Based on the above embodiments, the bat iteration module 430 may include:

[0119] The local optimal bat acquisition unit is used to update the vehicle steering decision model to be optimized by using the initial position of the bats, input the dimensionality reduction verification sample set into the updated vehicle steering decision model, obtain multiple local optimal bats in the current iteration round through a preset fitness function, and save the relevant information of the local optimal bats.

[0120] The new bat generation unit is used to eliminate at least one bat in the current bat group according to the preset iteration elimination rate and the fitness value of each bat in the current iteration round, and generate new bats according to the preset generation rules, keeping the number of bats in the bat group unchanged, and iterating on each bat in the current bat group.

[0121] The optimal flight position acquisition unit is used to calculate the optimal flight position of the current iteration based on the local optimal bat information in the previous iteration and the preset position iteration function, and to update the positions of each bat after the previous iteration based on the optimal flight position and the initial speed of the bat.

[0122] The iterative parameter update unit is used to compare the fitness values ​​of each bat after the current iteration with the fitness values ​​of each bat after the previous iteration, save the positions of the bats with higher fitness values ​​in the two iterations, and update the relevant parameters required for the next iteration. The next iteration is carried out until the bat iteration stops when the stopping condition is met.

[0123] Based on the above embodiments, the new bat generation unit can be specifically used for:

[0124] The number of eliminated bats is calculated based on the preset iteration elimination rate, and the fitness value of each bat in the current iteration round is calculated.

[0125] In the current iteration, among all the bats, select at least one bat with the lowest fitness value as the eliminated bat according to the number of eliminated bats.

[0126] New bats are generated according to the preset generation rules to match the number of eliminated bats, and the new bats are added to the bat population. The number of bats in the current bat population remains unchanged, and the process is repeated for each bat in the current bat population.

[0127] Based on the above embodiments, the optimal flight position acquisition unit can be specifically used for:

[0128] Select two learning bats from the best bats in the previous iteration and obtain the current position of each learning bat;

[0129] The optimal flight position for this iteration is calculated based on the preset position iteration function;

[0130] The preset position iteration function is:

[0131]

[0132] Where f1 and f2 are [f min f max The random number f in ] max with f minThe preset maximum and minimum fitness values ​​are... Let i be the position of the i-th bat in generation t. For each learning bat's location, X new1 This is the optimal flight position for this iteration.

[0133] The vehicle steering decision model optimization device provided in this embodiment of the invention can execute the vehicle steering decision model optimization method provided in Embodiment 1 and Embodiment 2 of the invention, and has the corresponding functional modules and beneficial effects of the execution method.

[0134] Example 5

[0135] Figure 5 This is a schematic diagram of a vehicle steering decision device provided in Embodiment 5 of the present invention. Figure 5 As shown, the device includes: a vehicle environment description feature value acquisition module 510, a steering decision result acquisition module 520, and a vehicle control module 530.

[0136] The vehicle environment description feature value acquisition module 510 is used to acquire multiple vehicle environment description feature values ​​corresponding to at least one associated vehicle related to the vehicle and the vehicle steering decision by using the information acquisition results of multiple sensors installed in the vehicle.

[0137] The vehicle environment description features include the current vehicle speed and acceleration of the vehicle and related vehicles, the current left front coordinate of the vehicle and related vehicles, the current front longitudinal coordinate of the vehicle and related vehicles, the current vehicle width and length of the vehicle and related vehicles, the distance between the vehicle and at least one vehicle in front, and the distance between the vehicle and at least one vehicle behind.

[0138] The steering decision result acquisition module 520 is used to input the plurality of vehicle environment description feature values ​​into the optimized vehicle steering decision model obtained by the method described in any one of Embodiment 1 and Embodiment 2 of the present invention, and to obtain the steering decision result.

[0139] The vehicle control module 530 is used to perform vehicle control operations that match the steering decision results.

[0140] The technical solution of this invention collects vehicle environment description feature values ​​by sensors, inputs these feature values ​​into an optimized vehicle steering decision model, and obtains steering decision results to control the vehicle. This method enables the vehicle system to make accurate automatic steering decisions based on the external vehicle environment, thereby improving the intelligence of the vehicle system.

[0141] The vehicle steering decision device provided in this embodiment of the invention can execute the vehicle steering decision method provided in the three embodiments of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0142] Example 6

[0143] Figure 6 A schematic diagram of an electronic device 60 that can be used to implement embodiments of the present invention is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices (e.g., helmets, glasses, watches, etc.), and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the invention described and / or claimed herein.

[0144] like Figure 6 As shown, the electronic device 60 includes at least one processor 61 and a memory, such as a read-only memory (ROM) 62 and a random access memory (RAM) 63, communicatively connected to the at least one processor 61. The memory stores computer programs executable by the at least one processor. The processor 61 can perform various appropriate actions and processes based on the computer program stored in the ROM 62 or loaded into the RAM 63 from storage unit 68. The RAM 63 may also store various programs and data required for the operation of the electronic device 60. The processor 61, ROM 62, and RAM 63 are interconnected via a bus 64. An input / output (I / O) interface 65 is also connected to the bus 64.

[0145] Multiple components in electronic device 60 are connected to I / O interface 65, including: input unit 66, such as keyboard, mouse, etc.; output unit 67, such as various types of monitors, speakers, etc.; storage unit 68, such as disk, optical disk, etc.; and communication unit 69, such as network card, modem, wireless transceiver, etc. Communication unit 69 allows electronic device 60 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0146] Processor 61 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of processor 61 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. Processor 61 performs the various methods and processes described above, such as the optimization method for a vehicle steering decision model and the vehicle steering decision method as described in embodiments of the present invention.

[0147] In some embodiments, the optimization method for the vehicle steering decision model and the vehicle steering decision method may be implemented as a computer program tangibly contained in a computer-readable storage medium, such as storage unit 68. In some embodiments, part or all of the computer program may be loaded and / or installed on electronic device 60 via ROM 62 and / or communication unit 69. When the computer program is loaded into RAM 63 and executed by processor 61, one or more steps of the optimization method for the vehicle steering decision model and the vehicle steering decision method described above may be performed. Alternatively, in other embodiments, processor 61 may be configured to execute the optimization method for the vehicle steering decision model and the vehicle steering decision method by any other suitable means (e.g., by means of firmware).

[0148] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0149] Computer programs used to implement the methods of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when executed by the processor, the computer programs cause the functions / operations specified in the flowcharts and / or block diagrams to be performed. The computer programs may be executed entirely on a machine, partially on a machine, or as a standalone software package, partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0150] In the context of this invention, a computer-readable storage medium can be a tangible medium that may contain or store a computer program for use by or in conjunction with an instruction execution system, apparatus, or device. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination thereof. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fibers, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0151] To provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the electronic device. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0152] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or computing systems that include middleware components (e.g., application servers), or computing systems that include frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include local area networks (LANs), wide area networks (WANs), blockchain networks, and the Internet.

[0153] A computing system can include clients and servers. Clients and servers are generally located far apart and typically interact through communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or cloud host, which is a hosting product within the cloud computing service system to address the shortcomings of traditional physical hosts and VPS services, such as high management difficulty and weak business scalability.

[0154] It should be understood that the various forms of processes shown above can be used, with steps reordered, added, or deleted. For example, the steps described in this invention can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution of this invention can be achieved, and this is not limited herein.

[0155] The specific embodiments described above do not constitute a limitation on the scope of protection of this invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this invention should be included within the scope of protection of this invention.

Claims

1. An optimization method for a vehicle steering decision model, characterized in that, include: Obtain the vehicle steering decision model to be optimized and the standard validation sample set that matches the vehicle steering decision model. The standard validation sample includes multiple vehicle environment description feature values ​​and standard steering decision results. Principal component analysis (PCA) was performed on the standard validation sample set to reduce its dimensionality, resulting in a dimensionality-reduced validation sample set. Based on the parameter setting of the bat algorithm, the bat algorithm is used to perform multi-round iterative processing of bat swarms; In each iteration, the vehicle steering decision model is updated using bats that match the current iteration. The fitness of inputting the dimensionality reduction validation sample set into each updated vehicle steering decision model is used as the evaluation index. Locally optimal bats are identified among the bats, and the bat swarm iterative optimization of the current iteration is performed using locally optimal bats. The bats include a penalty factor and a Gaussian kernel function. When the bat iteration termination condition is met, the final fitness value of the turning decision model for each bat is calculated based on the dimensionality reduction verification sample set, and the globally optimal bat is selected based on the final fitness value. The vehicle steering decision model is updated using the global optimal bat, resulting in an optimized vehicle steering decision model. The parameters set for the bat algorithm include the number of bats in the colony, the maximum number of iterations, the initial position of the bats, the initial speed of the bats, the fitness function, the iteration elimination rate, and the position iteration function. Based on the parameter setting of the bat algorithm, the bat algorithm is used to perform multi-round iterative processing of bat swarms, including: The vehicle steering decision model to be optimized is updated using the initial position of the bats. The dimensionality reduction verification sample set is input into the updated vehicle steering decision model. Multiple locally optimal bats in the current iteration are obtained through a preset fitness function, and the relevant information of the locally optimal bats is saved. Based on the preset iteration elimination rate and the fitness value of each bat in the current iteration round, at least one bat in the current bat group is eliminated, and a new bat is generated according to the preset generation rules, keeping the number of bats in the current bat group unchanged, and iterating on each bat in the current bat group. Based on the local optimal bat information from the previous iteration and the preset position iteration function, the optimal flight position for this iteration is calculated. The positions of each bat after the previous iteration are iteratively updated based on the optimal flight position and the initial speed of the bat. The fitness values ​​of each bat after the current iteration are compared with those after the previous iteration. The positions of the bats with higher fitness values ​​in the two iterations are saved. At the same time, the relevant parameters required for the next iteration are updated, and the next iteration is carried out until the bat iteration stopping condition is met, at which point the bat iteration stops.

2. The method according to claim 1, characterized in that, PCA dimensionality reduction was performed on the standard validation sample set to obtain a dimensionality-reduced validation sample set, including: After preprocessing the vehicle environmental description feature values ​​in each standard verification sample, PCA is used to reduce the dimensionality of each vehicle environmental description feature value to obtain a dimensionality-reduced verification sample set.

3. The method according to claim 1, characterized in that, The process of eliminating at least one bat from the current bat population based on a preset elimination rate and the fitness value of each bat in the current iteration round, and generating new bats according to a preset generation rule while keeping the number of bats in the current population unchanged, and iterating over each bat in the current bat population, includes: The number of eliminated bats is calculated based on the preset iteration elimination rate, and the fitness value of each bat in the current iteration round is calculated. In the current iteration, among all the bats, select at least one bat with the lowest fitness value as the eliminated bat according to the number of eliminated bats. New bats are generated according to the preset generation rules to match the number of eliminated bats, and the new bats are added to the bat population. The number of bats in the current bat population remains unchanged, and the process is repeated for each bat in the current bat population.

4. The method according to claim 1, characterized in that, The step of calculating the optimal flight position for the current iteration based on the local optimal bat information from the previous iteration and a preset position iteration function includes: Select two learning bats from the best bats in the previous iteration and obtain the current position of each learning bat; The optimal flight position for this iteration is calculated based on the preset position iteration function; The preset position iteration function is: ； in, , for The random number in f max with f min The preset maximum and minimum fitness values ​​are... Let i be the position of the i-th bat in generation t. For each learning bat's location, X new1 This is the optimal flight position for this iteration.

5. A vehicle steering decision method, executed by an in-vehicle infotainment system, characterized in that, include: By collecting information from multiple sensors installed in the vehicle, multiple vehicle environment description feature values ​​corresponding to at least one associated vehicle related to the vehicle and the vehicle's steering decision are obtained. The vehicle environment description features include the current vehicle speed and acceleration of the vehicle and related vehicles, the current left front coordinate of the vehicle and related vehicles, the current front longitudinal coordinate of the vehicle and related vehicles, the current vehicle width and length of the vehicle and related vehicles, the distance between the vehicle and at least one vehicle in front, and the distance between the vehicle and at least one vehicle behind. The plurality of vehicle environment description feature values ​​are input into the optimized vehicle steering decision model obtained by the method described in any one of claims 1-4, and the steering decision result is obtained. Perform vehicle control operations that match the steering decision result.

6. An optimization device for a vehicle steering decision model, characterized in that, include: The standard validation sample set acquisition module is used to acquire the vehicle steering decision model to be optimized and the standard validation sample set that matches the vehicle steering decision model. The standard validation sample includes multiple vehicle environment description feature values ​​and standard steering decision results. The dimensionality reduction validation sample set acquisition module is used to perform principal component analysis (PCA) dimensionality reduction on the standard validation sample set to obtain the dimensionality reduction validation sample set. The bat iteration module is used to set parameters based on the bat algorithm and perform multi-round iterative processing of bat swarms using the bat algorithm. In each iteration, the vehicle steering decision model is updated using bats that match the current iteration. The fitness of inputting the dimensionality reduction validation sample set into each updated vehicle steering decision model is used as the evaluation index. Locally optimal bats are identified among the bats, and the bat swarm iterative optimization of the current iteration is performed using locally optimal bats. The bats include a penalty factor and a Gaussian kernel function. The optimal bat selection module is used to calculate the final fitness value of each bat for the turning decision model based on the dimensionality reduction verification sample set when the bat iteration termination condition is met, and select the global optimal bat based on the final fitness value. The vehicle steering decision model optimization module is used to update the vehicle steering decision model using the global optimal bat to obtain the optimized vehicle steering decision model. The parameters set for the bat algorithm include the number of bats in the colony, the maximum number of iterations, the initial position of the bats, the initial speed of the bats, the fitness function, the iteration elimination rate, and the position iteration function. The bat iteration module includes: The local optimal bat acquisition unit is used to update the vehicle steering decision model to be optimized by using the initial position of the bats, input the dimensionality reduction verification sample set into the updated vehicle steering decision model, obtain multiple local optimal bats in the current iteration round through a preset fitness function, and save the relevant information of the local optimal bats. The new bat generation unit is used to eliminate at least one bat in the current bat group according to the preset iteration elimination rate and the fitness value of each bat in the current iteration round, and generate new bats according to the preset generation rules, keeping the number of bats in the bat group unchanged, and iterating on each bat in the current bat group. The optimal flight position acquisition unit is used to calculate the optimal flight position of the current iteration based on the local optimal bat information in the previous iteration and the preset position iteration function, and to iteratively update the positions of each bat after the previous iteration based on the optimal flight position and the initial speed of the bat. The iterative parameter update unit is used to compare the fitness values ​​of each bat after the current iteration with the fitness values ​​of each bat after the previous iteration, save the positions of the bats with higher fitness values ​​in the two iterations, and update the relevant parameters required for the next iteration. The next iteration is carried out until the bat iteration stops when the stopping condition is met.

7. A vehicle steering decision device, executed by a vehicle infotainment system, characterized in that, include: The vehicle environment description feature value acquisition module is used to acquire multiple vehicle environment description feature values ​​corresponding to at least one associated vehicle related to the vehicle and the vehicle steering decision by using the information collection results of multiple sensors installed in the vehicle. The vehicle environment description features include the current vehicle speed and acceleration of the vehicle and related vehicles, the current left front coordinate of the vehicle and related vehicles, the current front longitudinal coordinate of the vehicle and related vehicles, the current vehicle width and length of the vehicle and related vehicles, the distance between the vehicle and at least one vehicle in front, and the distance between the vehicle and at least one vehicle behind. The steering decision result acquisition module is used to input the plurality of vehicle environment description feature values ​​into the optimized vehicle steering decision model obtained by the method described in any one of claims 1-4, and to acquire the steering decision result; The vehicle control module is used to perform vehicle control operations that match the steering decision result.

8. An electronic device, characterized in that, The electronic device includes: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor, the computer program being executed by the at least one processor to enable the at least one processor to perform the optimization method of the vehicle steering decision model according to any one of claims 1-4, or to perform the vehicle steering decision method according to claim 5.

9. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions that, when executed by a processor, implement the optimization method of the vehicle steering decision model according to any one of claims 1-4, or execute the vehicle steering decision method according to claim 5.

Images