A vehicle control method, electronic device, vehicle, medium, and program product

Abstract

The application relates to the technical field of assisted driving, and discloses a vehicle control method, an electronic device, a vehicle, a medium and a program product. The method comprises the following steps: acquiring surrounding environment information of a target vehicle, determining a target scene and a potential interactive vehicle based on the surrounding environment information; determining target interaction intensity information based on feature information of the potential interactive vehicle; determining target driving action information based on the target interaction intensity information and historical driving style information of a driving user in the target scene; and controlling the target vehicle to travel based on the target driving action information. The application quantifies the interaction intensity of the surrounding vehicles and makes personalized decisions in combination with the historical driving style of the driving user, so that vehicle control that is adapted to the preferences of the driver and ensures driving safety is realized, and the use experience and user trust of the driving assistance function are improved.

Description

Technical Field

[0001] This application relates to the field of driver assistance technology, and in particular to a vehicle control method, electronic device, vehicle, medium, and program product. Background Technology

[0002] Currently, navigation on autopilot (NOA) is becoming increasingly common in vehicles. This feature can automatically perform driving operations such as lane keeping, lane changing and overtaking, and ramp merging based on the navigation route, thereby reducing the driver's workload.

[0003] However, existing NOA (Noise, Adaptive, and Orientation) systems typically employ uniformly preset driving decision-making schemes. For example, NOA systems generally use rule-based decision-making schemes based on safe distance and speed. In lane-changing scenarios, fixed safe lane-changing clearance thresholds are pre-set, such as lateral distance thresholds to vehicles in front and behind. When the threshold conditions are met and navigation indicates a lane change is necessary, a lane change is initiated. For yielding scenarios, a preset deceleration and yielding strategy is used. This approach fails to meet the personalized driving needs of different drivers. Summary of the Invention

[0004] To address the problem that existing technologies have uniform preset driving decisions that cannot meet the personalized needs of different drivers, embodiments of this application provide a vehicle control method, electronic device, vehicle, medium, and program product.

[0005] In a first aspect, embodiments of this application provide a vehicle control method applied to an electronic device. The method includes: acquiring surrounding environment information of a target vehicle; determining a target scene corresponding to the target vehicle and at least one potential interactive vehicle corresponding to the target vehicle based on the surrounding environment information; determining target interaction intensity information corresponding to at least one potential interactive vehicle based on feature information of the at least one potential interactive vehicle; determining target driving action information based on the target interaction intensity information and the historical driving style information of the driver of the target vehicle in the target scene; and controlling the target vehicle to drive based on the target driving action information.

[0006] The vehicle control method provided in this application obtains information about the surrounding environment of the target vehicle, determines the target scene and potential interactive vehicles, and determines the interaction intensity information based on the feature information of the potential interactive vehicles. This allows for the quantification of the strength of the interaction intentions between the surrounding vehicles and the target vehicle. Simultaneously, it determines the driver's historical driving style information in the target scene, reflecting the driver's personalized driving preferences. The target driving action information is determined based on the interaction intensity information and historical driving style information, ensuring that the final driving decision considers both the interaction status of surrounding vehicles in the current traffic environment and the driver's personal driving style. This achieves personalized vehicle control, effectively avoiding problems such as mismatch between decisions and user preferences, and frequent user takeovers caused by all drivers sharing a unified decision-making logic in existing technologies. This improves the user experience and user trust in driver assistance functions.

[0007] In one possible implementation of the first aspect, the target scenario includes any of the following: adjacent lane vehicle insertion scenario, overtaking scenario, ramp merging scenario, and congestion interaction scenario.

[0008] Based on the above solution, by covering four typical strong interaction scenarios, the solution of this application can be applied to the most common high-frequency scenarios in NOA system, such as lane change interaction, congestion weaving, and ramp merging, and has good scenario versatility and scalability.

[0009] In one possible implementation of the first aspect, the feature information of the potential interactive vehicle includes at least one of the relative velocity of the potential interactive vehicle relative to the target vehicle, lateral acceleration, and collision time change rate; determining the target interaction intensity information corresponding to at least one potential interactive vehicle based on the feature information of at least one potential interactive vehicle includes: performing weighted fusion processing on the feature information of each potential interactive vehicle among the at least one potential interactive vehicle to determine the interaction intensity information of each potential interactive vehicle; and determining the target interaction intensity information based on the interaction intensity information of each potential interactive vehicle.

[0010] Based on the above scheme, by weighted fusion of three types of features—relative speed, lateral acceleration, and collision time change rate—the interaction intensity of surrounding vehicles can be comprehensively quantified from three dimensions: approach speed, lane change intention, and safe distance compression. This avoids the one-sidedness of judging by a single feature and improves the accuracy of interaction intensity perception.

[0011] In one possible implementation of the first aspect, determining the target interaction intensity information based on the interaction intensity information of each potential interactive vehicle includes: when there are multiple potential interactive vehicles, determining the maximum value among the interaction intensity information corresponding to each potential interactive vehicle as the target interaction intensity information; and when there is only one potential interactive vehicle, determining the interaction intensity information corresponding to the potential interactive vehicle as the target interaction intensity information.

[0012] Based on the above scheme, by taking the maximum value of the interaction intensity information among multiple potential interactive vehicles as the target interaction intensity information, the decision-making basis can be taken as the vehicle with the strongest interaction intention, ensuring that the highest traffic participant is dealt with first in multi-vehicle interaction scenarios, thereby improving the safety of decision-making.

[0013] In one possible implementation of the first aspect, the historical driving style information of the driver of the target vehicle in the target scenario is determined based on the driver's historical driving behavior in the target scenario and in manual driving mode.

[0014] Based on the above scheme, by collecting the historical driving behavior of drivers in manual driving mode to determine historical driving style information, the learned driving style can truly reflect the driver's preferences and habits in actual driving, rather than the average style trained based on general data, thereby improving the accuracy of personalized adaptation.

[0015] In one possible implementation of the first aspect, the target driving action information is determined based on the target interaction intensity information and the historical driving style information of the driver of the target vehicle in the target scenario, including: inputting the surrounding environment information of the target vehicle into a first model to obtain the first driving tendency information output by the first model; determining driving action adjustment information based on the historical driving style information and the target interaction intensity information; adjusting the first driving tendency information based on the driving action adjustment information to obtain the second driving tendency information; and determining the target driving action information based on the second driving tendency information.

[0016] Based on the above scheme, the baseline decision tendency is output by the first model, and then the baseline output is adjusted in a personalized way based on the interaction intensity information and historical driving style information. This not only retains the end-to-end model's ability to generalize to complex scenarios, but also realizes interpretable and adjustable personalized offsets, so that the decision output is both in line with the commonality of the scenario and adapted to the user's personality.

[0017] In one possible implementation of the first aspect, the first driving tendency information includes the basic decision probabilities corresponding to multiple candidate driving actions, and the second driving tendency information includes the target decision probabilities corresponding to multiple candidate driving actions.

[0018] Based on the above scheme, by representing the first driving tendency information and the second driving tendency information in the form of decision probability distribution, it is possible to intuitively reflect the ranking of the merits of each candidate driving action, which is convenient for subsequent selection of the optimal action, and at the same time provides a unified numerical basis for probability offset correction.

[0019] In one possible implementation of the first aspect, determining the target driving action information based on the second driving tendency information includes: identifying the candidate driving action with the highest target decision probability among multiple candidate driving actions as the target driving action information.

[0020] Based on the above scheme, by selecting the candidate driving action with the highest probability of the target decision as the target driving action information, it can be ensured that the final action is the optimal action after personalized adjustment, thereby maximizing the decision benefit.

[0021] In one possible implementation of the first aspect, candidate driving actions include overtaking, yielding, and lane changing.

[0022] Based on the above scheme, by limiting candidate driving actions to three typical interactive actions such as overtaking, yielding, and changing lanes, the main decision-making choices of the NOA system in highly interactive scenarios can be covered, making the action space both complete and concise, and improving decision-making efficiency.

[0023] In one possible implementation of the first aspect, controlling the driving of the target vehicle based on the target driving action information includes: obtaining a risk assessment value corresponding to the proposed driving trajectory based on the proposed driving trajectory corresponding to the target driving action information; and controlling the driving of the target vehicle based on the target driving action information when the risk assessment value is less than or equal to a preset safety threshold.

[0024] Based on the above scheme, by obtaining a risk assessment value based on the proposed driving trajectory before performing a driving action, and only performing the action when the risk assessment value does not exceed the safety threshold, a safety baseline can be set for personalized decision-making, preventing unsafe behavior caused by personalized adjustments and ensuring the safety of driving decisions.

[0025] In one possible implementation of the first aspect, the method further includes: when the risk assessment value is greater than a preset safety threshold, determining the candidate driving action with the highest basic decision probability in the first driving tendency information, excluding the candidate driving action corresponding to the target driving action information, as the updated target driving action information.

[0026] Based on the above scheme, when the risk assessment value exceeds the safety threshold, the candidate driving action with the second highest probability of basic decision in the first driving tendency information is selected by downgrading. This can select the action that is as close to the baseline as possible while ensuring safety, and avoid directly downgrading to the default conservative action, which would lead to a decline in personalized experience.

[0027] Secondly, embodiments of this application provide an electronic device, including one or more processors and one or more memories; the one or more memories store one or more programs, which, when executed by the processor, cause the electronic device to perform any of the methods described in the first aspect.

[0028] Thirdly, embodiments of this application provide a vehicle including an electronic device for performing the method of any of the first aspects.

[0029] Fourthly, embodiments of this application provide a computer-readable storage medium storing a program or instructions that, when executed by an electronic device, implement the method of any one of the first aspects.

[0030] Fifthly, embodiments of this application provide a computer program product including instructions that, when executed, cause the method of any one of the first aspects to be implemented.

[0031] In a sixth aspect, embodiments of this application provide a chip including a processor coupled to a memory for executing a program or instructions stored in the memory, wherein when the program or instructions are executed by the processor, the method of any one of the first aspects is implemented. Attached Figure Description

[0032] Figure 1 According to an embodiment of this application, a schematic diagram illustrating an application scenario of a vehicle control method is provided.

[0033] Figure 2 According to an embodiment of this application, a flowchart of a vehicle control method is illustrated.

[0034] Figure 3 According to an embodiment of this application, a flowchart of another vehicle control method is illustrated.

[0035] Figure 4 According to some embodiments of this application, a functional framework diagram of a vehicle is shown. Detailed Implementation

[0036] The illustrative embodiments of this application include, but are not limited to, a vehicle control method, electronic device, vehicle, medium, and program product.

[0037] Figure 1 This is a schematic diagram illustrating an application scenario of the vehicle control method provided in an embodiment of this application. For example... Figure 1 As shown, target vehicle 101 is driving on a multi-lane road. A slow-moving vehicle 102 is ahead of it, a rapidly approaching vehicle 103 is in the adjacent lane to its left, and a vehicle 104 is behind it in the adjacent lane to its right. This is an overtaking scenario. In this situation, the NOA (No Action Assist) function of target vehicle 101 is activated, requiring a driving decision, such as choosing to overtake, yield, or use the other lane to observe.

[0038] In the above scenarios, the driving decision-making schemes in some embodiments mainly include the following:

[0039] Firstly, there's the rule-based decision-making scheme based on safe distance and speed. This scheme pre-sets fixed safe lane-changing clearance thresholds, such as lateral distance thresholds to vehicles in front and behind. When the threshold conditions are met and the navigation requires a lane change, a lane change is initiated. For yielding scenarios, a preset deceleration and yielding strategy is used. The drawback of this scheme is that all drivers share the same decision-making logic, making it unsuitable for the individual driving styles of different drivers. For example, the scheme is too conservative for users who prefer aggressive play, while it appears harsh and aggressive for users who prefer a smooth and courteous driving style, easily leading to frequent driver intervention.

[0040] Secondly, there is the decision-making scheme based on the classic interaction model. This scheme calculates the lane change benefit based on the motion state of surrounding vehicles by minimizing the overall braking induced by lane changes (MOBIL) model or the Intelligent Driver Model (IDM) model, and then decides whether to change lanes or give way. However, this scheme also cannot meet the personalized needs of different drivers.

[0041] Third, there is the end-to-end imitation learning-based decision-making scheme. This scheme collects a large amount of human driving data to train a perception-decision mapping network, and outputs driving actions based on the perception-decision mapping network. However, its output behavior often reflects the performance of the "average driver" and cannot adapt to the differences in individual driver preferences.

[0042] In summary, none of the above solutions can meet the personalized needs of different drivers.

[0043] To address the aforementioned issues, this application provides a vehicle control method. In this method, the scene in which the target vehicle is located and potential interactive vehicles that may interact with the target vehicle can first be identified based on the surrounding environment information of the target vehicle. Then, the target driving action information is jointly determined based on the user's historical driving style information in the scene in which the target vehicle is located and the interaction intensity information of the potential interactive vehicles. The target vehicle is then controlled to drive based on the target driving action information.

[0044] Based on the above scheme, by determining the interaction intensity information of potential interacting vehicles, the strength of the interaction intention between the perceived surrounding vehicles and the target vehicle can be quantified. Simultaneously, determining the driver's historical driving style information in the target scenario reflects the driver's personalized driving preferences in similar traffic situations. By jointly determining the target driving action information based on interaction intensity information and historical driving style information, the final driving decision considers both the interaction status of surrounding vehicles in the current traffic environment and adapts to the driver's personal driving style. This achieves personalized vehicle control, effectively avoiding the problems of decision mismatch and frequent user takeover caused by all drivers sharing a unified decision-making logic in existing technologies. This improves the user experience and user trust in driver assistance functions.

[0045] The vehicle control method provided in the embodiments of this application will be described in detail below. The executing entity of this vehicle control method can be an electronic device. Figure 2 A schematic flowchart of a vehicle control method provided in an embodiment of this application is shown, as follows: Figure 2 As shown, the method includes:

[0046] 201: Obtain information about the surrounding environment of the target vehicle.

[0047] In some embodiments, electronic devices can acquire information about the surrounding environment of the target vehicle through sensors configured on the target vehicle and a map. Sensors may include forward-facing cameras, surround-view cameras, corner radar, lidar, etc. Maps may include high-definition maps or navigation SD maps, etc.

[0048] In some embodiments, the surrounding environment information may include traffic participant information and road structure information. The traffic participant information may include characteristic information of surrounding vehicles, such as vehicle position, longitudinal velocity, lateral velocity, longitudinal acceleration, lateral acceleration, and vehicle type; the road structure information may include the number of lanes, lane types, and merging / emerging point locations. In some embodiments, the traffic participant information may also include interactive dynamic characteristics of surrounding vehicles, such as yaw rate, approach angle, time to collision (TTC), time headway (THW), and their rate of change.

[0049] For example, such as Figure 1 In the scenario shown, the target vehicle 101 detects the position and speed of the vehicle 102 in front using a forward-facing camera and corner radar, and detects the movement status of the vehicle 103 in the adjacent lane on the left and the vehicle 104 behind on the right using a surround-view camera. At the same time, it obtains information such as the number of lanes, lane type, and merging / exiting points of the current road through a high-precision map or a navigation SD map.

[0050] 202: Determine the target scene corresponding to the target vehicle and at least one potential interactive vehicle corresponding to the target vehicle based on the surrounding environment information.

[0051] After acquiring information about the surrounding environment, the electronic device identifies the target scenario in which the target vehicle is currently located. The target scenario can include any of the following: a vehicle cutting in from an adjacent lane, an overtaking scenario, a ramp merging scenario, and a congestion interaction scenario.

[0052] Specifically, the adjacent lane vehicle insertion scenario refers to a scenario where an adjacent lane vehicle is shifting towards the target vehicle's lane (i.e., there is an intention to insert); the overtaking scenario refers to a scenario where the vehicle in front is moving slowly and there is a usable gap in the adjacent lane, and the target vehicle needs to perform an overtaking operation; the ramp merging scenario refers to a scenario where the target vehicle is in the ramp merging area and needs to interact with vehicles on the main road; and the congestion interaction scenario refers to a scenario where multiple lanes are congested, the distance between vehicles is small, and there is no absolutely safe gap.

[0053] In some embodiments, the electronic device can filter out at least one potential interactive vehicle from the surrounding vehicles based on the position information of surrounding vehicles relative to the target vehicle in the surrounding environment information. Specifically, the electronic device delineates a preset spatial range centered on the target vehicle and identifies surrounding vehicles located within this preset spatial range as potential interactive vehicles. The preset spatial range can be set according to actual needs, for example, it can be an area centered on the target vehicle, extending a preset distance (e.g., 200 meters) longitudinally (i.e., in the direction of travel of the target vehicle) forward and backward, and laterally (i.e., in the direction perpendicular to the direction of travel) covering the lane where the target vehicle is located and the adjacent lanes.

[0054] For example, in Figure 1 In the overtaking scenario shown, the vehicle in front 102 and the vehicle on the left adjacent lane 103 are located within the preset space range of the target vehicle 101 and are therefore identified as potential interaction vehicles. However, the vehicle behind 104 on the right adjacent lane is not identified as a potential interaction vehicle because it is outside the preset range.

[0055] 203: Determine the target interaction intensity information corresponding to at least one potential interactive vehicle based on the feature information of at least one potential interactive vehicle.

[0056] In some embodiments, after identifying at least one potential interactive vehicle, the electronic device determines target interaction strength information based on the characteristic information of each potential interactive vehicle. The target interaction strength information is used to quantitatively characterize the strength of the interaction intention between surrounding vehicles and the target vehicle in the current traffic environment.

[0057] In some embodiments, the feature information of the potential interactive vehicles includes at least one of the relative velocity of the potential interactive vehicles relative to the target vehicle, lateral acceleration, and collision time change rate. The electronic device can perform weighted fusion processing on the feature information of each of the at least one potential interactive vehicles to determine the interaction intensity information of each potential interactive vehicle, and determine the target interaction intensity information based on the interaction intensity information of each potential interactive vehicle.

[0058] In some embodiments, determining target interaction intensity information based on the interaction intensity information of each potential interactive vehicle includes: when there are multiple potential interactive vehicles, determining the maximum value among the interaction intensity information corresponding to each potential interactive vehicle as the target interaction intensity information; when there is only one potential interactive vehicle, determining the interaction intensity information corresponding to that potential interactive vehicle as the target interaction intensity information.

[0059] In some embodiments, the electronic device may perform weighted fusion processing on the feature information of each potential interactive vehicle to determine the interaction intensity information of each potential interactive vehicle in a manner that includes: normalizing the feature information of each potential interactive vehicle to obtain normalized components corresponding to each feature, and then performing weighted fusion on each normalized component to determine the interaction intensity information of each potential interactive vehicle.

[0060] For example, for any potential interactive vehicle j, the electronic device can obtain the relative velocity component, the lateral acceleration component, and the collision time rate of change component.

[0061] The relative velocity component (v_rel_norm) is calculated as follows:

[0062] v_rel_norm=clamp((Δv_rel+5.0) / 15.0,0,1).

[0063] Here, `clamp(x, min, max)` is a truncation function that restricts the input value `x` to the interval [min, max]: outputting `min` when `x` is less than `min`, outputting `max` when `x` is greater than `max`, and outputting `x` itself when `x` is within the interval [min, max]. In this formula, `min` is 0 and `max` is 1, meaning the result of the calculation `(Δv_rel+5.0) / 15.0` is truncated to between 0 and 1. `Δv_rel` is the relative speed of the potential interactive vehicle relative to the target vehicle, representing the speed difference between the two vehicles. A positive value indicates that the potential interactive vehicle is faster than the target vehicle and is approaching it. `Δv_rel=v_target-v_ego`, where `v_target` is the speed of the potential interactive vehicle and `v_ego` is the speed of the target vehicle.

[0064] The lateral acceleration component (lat_acc_norm) is calculated as follows:

[0065] lat_acc_norm=clamp(|a_y_target| / 2.0,0,1).

[0066] Where a_y_target is the lateral acceleration of the potential interacting vehicle, and |a_y_target| is the absolute value of the lateral acceleration. The lateral acceleration component reflects the intensity of the potential interacting vehicle's lane-changing intention; a larger lateral acceleration indicates a stronger lane-changing intention. The meaning of the clamp function is the same as described above and will not be repeated here.

[0067] The collision time TTC rate of change component (ttc_rate_norm) is calculated as follows:

[0068] ttc_rate_norm=max(0,-(d(TTC) / dt) / 1.0).

[0069] Where d(TTC) / dt is the rate of change of collision time. When d(TTC) / dt is negative, -(d(TTC) / dt) is positive, and the larger the value, the faster the TTC decreases, the more aggressive the driving behavior of the potential interacting vehicle, and the stronger the interaction intention. When d(TTC) / dt is positive, -(d(TTC) / dt) is negative, truncated to 0 by max(0,⋅), indicating that the potential interacting vehicle did not exert interaction pressure on the target vehicle. max(0,x) represents taking the larger value between 0 and x.

[0070] Then, the electronic device performs weighted fusion of the above three interaction feature components to obtain the interaction intensity information of the potential interactive vehicle j:

[0071] K_agg_j=w1×v_rel_norm+w2×lat_acc_norm+w3×ttc_rate_norm.

[0072] Wherein, K_agg_j represents the interaction intensity information of the potential interactive vehicle j, with a value range of [0,1]. A larger value indicates a stronger interaction intent and more aggressive driving behavior from the potential interactive vehicle. w1, w2, and w3 are preset weights, and w1+w2+w3=1. The values ​​of w1, w2, and w3 can be determined according to actual needs, for example, w1=0.4, w2=0.35, and w3=0.25. It is understood that the above weight values ​​are only one embodiment, and those skilled in the art can adjust the values ​​of each weight according to the actual application scenario.

[0073] When there are multiple potential interactive vehicles, the electronic device determines the target interactive intensity information K_agg from the maximum value of the interactive intensity information corresponding to each potential interactive vehicle, i.e., K_agg = max_j(K_agg_j). It can be understood that the vehicle with the strongest interactive intent often determines the decision-making tone in the current scenario, so taking the maximum value can make the target interactive intensity information accurately reflect the most influential interactive situation at present.

[0074] For example, in Figure 1 In the scenario shown, the potential interacting vehicles include adjacent vehicle 103 and vehicle 102 ahead. For adjacent vehicle 103, Δv_rel = +3m / s, a_y_target = 1.0m / s², and d(TTC) / dt = -0.6s / s. Therefore, v_rel_norm = (3+5) / 15 ≈ 0.533, lat_acc_norm = 0.5, ttc_rate_norm = 0.6, and K_agg = 0.4×0.533 + 0.35×0.5 + 0.25×0.6 ≈ 0.538. That is, the interaction intensity information of adjacent vehicle 103 is 0.538. If the interaction intensity information of vehicle 102 ahead is calculated based on the above method, it is 0.2. Then the electronic device will determine the maximum value of 0.538 as the target interaction intensity information.

[0075] 204: Determine the target driving action information based on the target interaction intensity information and the target vehicle's driver's historical driving style information in the target scenario.

[0076] In some embodiments, after determining the target interaction intensity information, the electronic device can combine the target interaction intensity information with the historical driving style information of the driver of the target vehicle in the target scenario to jointly determine the target driving action information. The historical driving style information is used to characterize the driver's personalized driving preferences in similar traffic scenarios and can be learned based on the driver's historical driving behavior in manual driving mode.

[0077] In some embodiments, the electronic device can input surrounding environment information into a pre-trained first model to obtain baseline driving tendency information, and then make personalized adjustments to the baseline driving tendency information based on target interaction intensity information and historical driving style information to obtain target driving action information.

[0078] The following sections will provide a detailed explanation of the learning process for historical driving style information and the determination process for target driving action information.

[0079] In some embodiments, the historical driving style information of the driver of the target vehicle in the target scenario is determined based on the driver's historical driving behavior in the target scenario and in manual driving mode.

[0080] In some embodiments, during manual driving mode, the electronic device continuously collects the driver's driving operations. Whenever the device detects that the driver has completed a preset driving action (e.g., overtaking, yielding, or changing lanes), it records the scenario type corresponding to that action and the decision action chosen by the driver, storing this as a historical driving decision event. Based on the stored historical driving decision events, the electronic device determines the driver's historical driving style information in different scenarios.

[0081] In some embodiments, after acquiring a driving decision event in manual driving mode, the electronic device can determine the instantaneous driving style information corresponding to the event based on the action type of the decision action and the driving characteristic information during the execution of the decision action.

[0082] Instantaneous driving style information is used to characterize the degree of a driver's driving tendency in a single driving decision event. The calculation method for instantaneous driving style information can be as follows:

[0083] κ_game=g(a)×(1+β×(1-d_margin))×(1+γ×a_y_max / a_y_max_safe).

[0084] Wherein, κ_game represents instantaneous driving style information, used to characterize the degree of driving tendency of the driver in a single driving decision event. g(a) is the basic tendency value of action, which is related to the type of decision action selected by the driver, for example, g(overtaking) = 0.9, g(borrowing lanes) = 0.5, g(yielding) = 0.1. d_margin is the minimum safe distance margin, that is, the ratio of the actual distance to the safe distance, which can range from [0,2]. a_y_max is the maximum lateral acceleration during the execution of the driving decision event, and a_y_max_safe is the preset safe lateral acceleration threshold (e.g., 1.5 m / s²). β and γ are preset coefficients, for example, β = 0.3, γ = 0.2. The calculated κ_game is truncated to the interval [0,1], the closer to 1, the more aggressive the driving behavior, and the closer to 0, the more conservative. It is understood that the above values ​​of basic tendency value of action, preset coefficients and safety threshold are only one embodiment, and those skilled in the art can adjust them according to actual needs.

[0085] Then, based on the instantaneous driving style information and the existing historical driving style information under the corresponding scenario type, the electronic device updates the historical driving style information by means of a moving average:

[0086] K_style_new[s]=clamp(α×κ_game+(1-α)×K_style_old[s],0,1).

[0087] Where s represents the scene type, K_style_old[s] represents the existing historical driving style information under this scene type, K_style_new[s] represents the updated historical driving style information, and α represents the learning rate. The learning rate can be adaptively adjusted: for example, when |κ_game-K_style_old|>0.3, the representation deviation is large, and the value of α can be larger, for example, setting α=0.3 to quickly adapt to changes in driving style; otherwise, α=0.1 is used for smooth updates. It is understood that the above values ​​of the learning rate and the threshold for adaptive adjustment are only one embodiment, and those skilled in the art can adjust them according to factors such as the desired style learning speed and smoothness. The meaning of the clamp function is the same as described above, and will not be repeated here.

[0088] For example, if a driver's existing historical driving style information K_style_old=0.55 in an overtaking scenario, and the driver completes an overtaking operation, the instantaneous driving style information κ_game=0.9 with a deviation of 0.35, and we take α=0.3, then the updated historical driving style information K_style_new=0.3×0.9+0.7×0.55=0.655.

[0089] After obtaining the updated historical driving style information, the electronic device associates and stores the updated historical driving style information with the corresponding scenario type. In other words, the electronic device independently maintains the corresponding historical driving style information for each scenario type. When the NOA function is activated, the electronic device, based on the current target scenario type, retrieves the historical driving style information corresponding to the current target scenario from the pre-stored historical driving style information corresponding to each of the multiple scenario types.

[0090] After determining the target interaction intensity information and the historical driving style information in the target scenario, the electronic device determines the target driving action information based on the target interaction intensity information and the historical driving style information. In some embodiments, the target interaction intensity information can be an intensity parameter, and the historical driving style information can be a style parameter; both are quantifiable values ​​between 0 and 1.

[0091] In some embodiments, the electronic device inputs the surrounding environment information of the target vehicle into the first model, obtains the first driving tendency information output by the first model, and can adjust the first driving tendency information based on the target interaction intensity information and historical driving style information to obtain the second driving tendency information, and determine the target driving action information based on the second driving tendency information.

[0092] In some embodiments, the first model is an end-to-end decision network pre-trained based on historical driving data. The first driving tendency information includes the basic decision probabilities corresponding to multiple candidate driving actions. Candidate driving actions include overtaking, yielding, and lane-changing actions.

[0093] In one implementation, the first model outputs the baseline value Z_base=[z_o,z_y,z_l] corresponding to each candidate driving action, where z_o corresponds to the basic decision probability of overtaking, z_y corresponds to the basic decision probability of yielding, and z_l corresponds to the basic decision probability of lane-changing.

[0094] In some embodiments, the electronic device determines driving action adjustment information based on historical driving style information and target interaction intensity information. The first driving tendency information is adjusted based on the driving action adjustment information to obtain second driving tendency information. The second driving tendency information includes the target decision probabilities corresponding to multiple candidate driving actions.

[0095] In one implementation, the driving action adjustment information can be calculated using the following formula:

[0096] M=(K_style-0.5)×(2×K_agg-1).

[0097] Where M represents driving action adjustment information (or modulation factor), with a theoretical range of [-0.5, 0.5]. M>0 indicates a tendency towards more aggressive actions (such as overtaking or accelerating into lane), M<0 indicates a tendency towards more conservative actions (such as yielding or slowing down and waiting), and M≈0 indicates maintaining the baseline decision. K_style represents historical driving style information, and K_agg represents target interaction intensity information.

[0098] To avoid unnecessary intervention, a dead zone mechanism can be set: when |K_style-0.5|<0.1 and / or |2×K_agg-1|<0.2, the driving action adjustment information is determined to be zero, that is, no adjustment is made to the primary driving tendency information.

[0099] It's understandable that |K_style-0.5|<0.1 indicates that the historical driving style information K_style is between 0.4 and 0.6, suggesting that the driver doesn't have a clear tendency towards being aggressive or conservative. |2×K_agg-1|<0.2 indicates that the target interaction intensity information K_agg is between 0.4 and 0.6, suggesting that the interaction intent of surrounding vehicles is unclear and the interaction intensity is neither strong nor weak. The information is insufficient to determine the other party's intention. In both cases, the primary driving tendency information is not adjusted to avoid making incorrect corrections when information is ambiguous.

[0100] In some embodiments, the first driving tendency information is adjusted based on driving action adjustment information, and the second driving tendency information is obtained by the following method: the electronic device obtains the original output value (i.e., the basic decision probability) corresponding to each candidate driving action in the first driving tendency information; based on the driving action adjustment information and the action indication information corresponding to each candidate driving action, the original output value corresponding to each candidate driving action is offset and corrected to obtain the corrected output value corresponding to the candidate driving action; the corrected output value is normalized to obtain the target decision probability of each candidate driving action, i.e., the second driving tendency information.

[0101] In some embodiments, the corrected output value can be obtained using the following formula:

[0102] Z_final(action)=Z_base(action)+λ×M×I(action).

[0103] Where Z_final(action) is the output value corresponding to the corrected candidate driving action, Z_base(action) is the original output value corresponding to the candidate driving action output by the first model, λ is the personalization intensity coefficient (e.g., λ=2.0), M is the driving action adjustment information, and I(action) is the action indication information: the action indication information corresponding to the overtaking action is a positive indication value (e.g., I_o=+1), the action indication information corresponding to the yielding action is a negative indication value (e.g., I_y=-1), and the action indication information corresponding to the lane-changing action is zero (e.g., I_l=0).

[0104] After obtaining the corrected output value Z_final, the electronic device normalizes the corrected output value Z_final to obtain the target decision probability of each candidate driving action (the value is between [0,1] and the sum of the probabilities of each action is 1), which is the second driving tendency information.

[0105] In some embodiments, the method of determining the target driving action information based on the second driving tendency information may include: determining the candidate driving action with the highest target decision probability among multiple candidate driving actions as the target driving action information.

[0106] For example, the baseline value Z_base of the first model output is [1.2 (overtaking), 0.5 (yielding), 0.3 (borrowing lanes)], corresponding to a basic decision probability of approximately [0.55, 0.27, 0.18]. Assuming the historical driving style information K_style = 0.8 (biased towards aggressive) and the target interaction intensity information K_agg = 0.6, then the driving action adjustment information M = (0.8 - 0.5) × (2 × 0.6 - 1) = 0.3 × 0.2 = 0.06, λ × M = 0.12. After correction, Z_final = [1.32, 0.38, 0.30], and the normalized target decision probability is approximately [0.62, 0.23, 0.15]. The overtaking action has the highest target decision probability, and the electronic equipment identifies the overtaking action as the target driving action information.

[0107] 205: Control the driving of the target vehicle based on the target driving action information.

[0108] In some embodiments, after determining the target driving action information, the electronic device controls the target vehicle to drive based on the target driving action information.

[0109] In some embodiments, to prevent unsafe behavior caused by personalized adjustments, the electronic device can also perform safety verification before controlling vehicle driving. That is, it calculates a risk assessment value based on the proposed driving trajectory corresponding to the target driving action information, and decides whether to execute the driving action based on the comparison result of the risk assessment value and a preset safety threshold. This will be explained in detail below.

[0110] In some embodiments, the electronic device obtains a risk assessment value corresponding to the proposed driving trajectory based on the target driving action information.

[0111] In some embodiments, to assess the safety of the proposed driving trajectory, the electronic device constructs a potential energy field (i.e., a risk field) centered on the target vehicle. The potential energy field characterizes the risk level of the target vehicle at various locations along the proposed driving trajectory; the greater the total potential energy U(x,y) at location (x,y), the higher the risk at that location. The total potential energy U(x,y) is obtained by superimposing the obstacle potential energy U_obs(x,y), the lane line potential energy U_lane(x,y), and the velocity potential energy U_vel(x,y), i.e.:

[0112] U(x,y)=U_obs(x,y)+U_lane(x,y)+U_vel(x,y).

[0113] Wherein, U_obs(x,y) is the obstacle potential energy, used to characterize the risk posed by surrounding obstacles (such as other vehicles) to the target vehicle; U_lane(x,y) is the lane line potential energy, used to characterize the risk of the vehicle deviating from the lane boundary; and U_vel(x,y) is the velocity potential energy, used to characterize the risk posed by obstacles approaching the target vehicle along the intended driving trajectory.

[0114] The obstacle potential energy is calculated as follows: U_obs = Σ_j A_obs_j × exp(-((x-x_j)² / (2σ_x²)+(y-y_j)² / (2σ_y²))), where (x_j, y_j) is the position of the j-th obstacle, σ_x is the standard deviation of the Gaussian distribution in the longitudinal direction (e.g., σ_x = 3.0m), σ_y is the standard deviation of the Gaussian distribution in the lateral direction (e.g., σ_y = 1.0m), and A_obs_j is the potential energy amplitude of the j-th obstacle. A_obs_j is determined as follows: A_obs_j = A_obs × (1 + K_agg_j), where A_obs is the basic potential energy amplitude (e.g., A_obs = 5.0), and K_agg_j is the interaction strength information of the j-th obstacle. For obstacles with zero or small interaction intensity information, K_agg_j≈0 and A_obs_j≈A_obs, meaning they have the same potential energy amplitude as the baseline. For obstacles with large interaction intensity information, K_agg_j is larger, and A_obs_j increases accordingly; that is, the greater the interaction intensity information of the obstacle, the greater the corresponding potential energy amplitude. The method for determining the interaction intensity information of obstacles is the same as the method for determining the interaction intensity information of potential interactive vehicles described earlier.

[0115] The lane line potential energy is calculated as follows: U_lane = A_lane / (dist_to_lane_boundary + ε), where A_lane is the lane line potential energy amplitude (e.g., A_lane = 2.0), dist_to_lane_boundary is the distance from the vehicle to the lane boundary line, and ε is a small constant to prevent division by zero (e.g., ε = 0.1).

[0116] The velocity potential energy is calculated as follows: U_vel = A_vel × max(0, v_rel_along), where A_vel is the velocity potential energy coefficient (e.g., A_vel = 0.5), v_rel_along is the relative velocity component of the obstacle relative to the target vehicle along the intended driving trajectory, and max(0, v_rel_along) means that velocity potential energy is generated only when the obstacle approaches the target vehicle along the intended driving trajectory.

[0117] For a proposed driving trajectory τ:(x(t),y(t)), t∈[0,T], U(x(t),y(t)) represents the total potential energy at position (x(t),y(t)) of the proposed driving trajectory at time t. The electronic device calculates the risk assessment value corresponding to the proposed driving trajectory based on the total potential energy at each time point on the proposed driving trajectory.

[0118] The risk assessment value is calculated as follows:

[0119] Risk=∫0ᵀU(x(t),y(t))×w(t)dt.

[0120] Where Risk is the risk assessment value, T is the total duration of the proposed driving trajectory, and w(t) = exp(-t / 2.0) is the recent weighting function, indicating that the more recent the moment, the greater the weight. The preset safety threshold Risk_threshold_safe can be 3.0 (offline calibration). It is understood that the above potential energy parameters and the preset safety threshold are only one embodiment, and those skilled in the art can adjust them according to actual needs.

[0121] In some embodiments, when the risk assessment value is less than or equal to a preset safety threshold, the electronic device controls the target vehicle to drive based on the target driving action information. When the risk assessment value is greater than the preset safety threshold, the electronic device rejects the current target driving action information and determines the candidate driving action with the highest basic decision probability from the first driving tendency information, excluding the candidate driving action corresponding to the current target driving action information, as the updated target driving action information, and re-performs safety verification.

[0122] For example, in an overtaking scenario, if the total risk assessment value Risk=3.7>3.0 corresponding to the proposed overtaking trajectory, overtaking is prohibited, and the lane-changing action with the second highest probability of basic decision in the first driving tendency information is determined as the updated target driving action information.

[0123] The vehicle control method provided in this application obtains information about the surrounding environment of the target vehicle, determines the target scene and potential interactive vehicles, and determines the interaction intensity information based on the feature information of the potential interactive vehicles. This allows for the quantification of the strength of the interaction intentions between the surrounding vehicles and the target vehicle. Simultaneously, by combining the driver's historical driving style information in the target scene with the interaction intensity information, the target driving action information is determined. Risk assessment ensures driving safety, making the final driving decision consider both the interaction status of surrounding vehicles in the current traffic environment and the driver's personal driving style. This achieves safe and personalized vehicle control, effectively avoiding problems such as mismatch between decision and user preferences and frequent user takeover caused by all drivers sharing a unified decision-making logic in the prior art. This improves the user experience and user trust of the driving assistance function.

[0124] For example, Figure 3 According to some embodiments of this application, a schematic flowchart of a vehicle control method is shown, wherein the executing entity of the vehicle control method can be an electronic device. For example... Figure 3 As shown, the method includes:

[0125] 301: Obtain information about the surrounding environment of the target vehicle.

[0126] In some embodiments, the electronic device acquires surrounding environmental information in real time through sensors configured on the target vehicle and a map. The specific content of the surrounding environmental information can be referred to step 201 above, and will not be repeated here.

[0127] 302: Determine the target interaction intensity information K_agg corresponding to the potential interactive vehicle based on the surrounding environment information.

[0128] In some embodiments, the electronic device filters out at least one potential interactive vehicle from surrounding vehicles based on surrounding environmental information, which has the possibility of interacting with the target vehicle, and determines the target interaction intensity information K_agg based on the feature information of each potential interactive vehicle. The target interaction intensity information is used to quantitatively characterize the strength of the interaction intention between surrounding vehicles and the target vehicle in the current traffic environment. Its specific calculation method can be referred to step 203 above, and will not be repeated here.

[0129] 303: Determine if the NOA function is activated.

[0130] In some embodiments, the electronic device determines in real time whether the NOA function of the target vehicle is activated. If the NOA function is not activated (i.e., the target vehicle is in manual driving mode), the process proceeds to step 313 to collect the driver's driving decision events and perform driving style learning; if the NOA function is activated, the process proceeds to step 304 to make personalized driving decisions based on the stored historical driving style information.

[0131] It's understandable that the learning process of historical driving style information is only triggered in manual driving mode. When the NOA function is activated, style updates are no longer performed; instead, the previously learned historical driving style information is directly invoked. This avoids instability in decision-making caused by frequent changes in style parameters during the NOA decision-making process.

[0132] 304: Determine the target scene corresponding to the target vehicle based on surrounding environmental information.

[0133] In some embodiments, when the NOA function is activated, the electronic device identifies the target scenario in which the target vehicle is currently located based on surrounding environmental information. The target scenario may include any of the following: adjacent lane vehicle insertion scenario, overtaking scenario, ramp merging scenario, and congestion interaction scenario. The specific identification method of the target scenario can be referred to step 202 above, and will not be repeated here.

[0134] 305: Obtain historical driving style information K_style for the target scenario.

[0135] In some embodiments, the electronic device independently maintains corresponding historical driving style information for each scenario type. After identifying the current target scenario, the electronic device retrieves the historical driving style information K_style corresponding to the current target scenario from the pre-stored historical driving style information corresponding to each of the multiple scenario types. The historical driving style information is used to characterize the driver's personalized driving preferences in similar traffic scenarios, and its learning process can be referred to step 204 above, which will not be repeated here.

[0136] 306: Obtain the first driving preference information output by the end-to-end decision network.

[0137] In some embodiments, the first model is an end-to-end decision network pre-trained based on historical driving data. The electronic device can input the surrounding environment information of the target vehicle into the first model to obtain the first driving tendency information output by the first model. The first driving tendency information includes the basic decision probabilities corresponding to multiple candidate driving actions, including overtaking, yielding, and lane changing actions. A detailed description of the first model can be found in step 204 above, and will not be repeated here.

[0138] 307: Calculate the modulation factor based on the target interaction intensity information K_agg and the historical driving style information K_style.

[0139] In some embodiments, the target interaction intensity information is an intensity parameter, and the historical driving style information is a style parameter; both are quantifiable values ​​between 0 and 1. The electronic device determines driving action adjustment information (i.e., modulation factor) based on the historical driving style information and the target interaction intensity information. The specific calculation method for the driving action adjustment information can be referred to step 204 above, and will not be repeated here.

[0140] 308: Adjust the first driving tendency information based on the modulation factor to determine the target driving action information.

[0141] In some embodiments, after determining the modulation factor, the electronic device performs offset correction on the original output values ​​corresponding to each candidate driving action in the first driving tendency information based on the modulation factor and the corresponding action indication information of each candidate driving action, to obtain the corrected output value. The corrected output value is then normalized to obtain the target decision probability of each candidate driving action, i.e., the second driving tendency information. The specific methods of offset correction and normalization can be referred to in step 204 above, and will not be repeated here.

[0142] In some embodiments, the electronic device determines the candidate driving action with the highest probability of target decision among multiple candidate driving actions as the target driving action information.

[0143] 309: Determine the risk assessment value of the proposed driving trajectory corresponding to the target driving action information.

[0144] In some embodiments, the electronic device generates a corresponding proposed driving trajectory based on the target driving action information, and calculates the risk assessment value corresponding to the proposed driving trajectory based on the constructed vehicle risk field. The specific calculation method of the risk assessment value can be referred to step 205 above, and will not be repeated here.

[0145] 310: Determine whether the risk assessment value is greater than the preset safety threshold.

[0146] In some embodiments, the electronic device determines whether the risk assessment value is greater than a preset safety threshold. If the risk assessment value is greater than the preset safety threshold, the process proceeds to step 311, where a new target driving action information is selected; if the risk assessment value is less than or equal to the preset safety threshold, the process proceeds to step 312, where a decision is made and the execution speed is adjusted.

[0147] It is understandable that the preset safety threshold is the risk limit determined offline. When the risk assessment value of the proposed driving trajectory exceeds this threshold, it indicates that there is a safety hazard in the driving action and a more conservative decision needs to be made.

[0148] 311: The candidate driving action with the highest basic decision probability among the first driving tendency information, excluding the current driving action, is identified as the new target driving action information.

[0149] In some embodiments, if the risk assessment value exceeds a preset safety threshold, the electronic device rejects the current target driving action information and identifies the candidate driving action with the highest basic decision probability (excluding the candidate driving action corresponding to the current target driving action information) from the first driving tendency information as the new target driving action information. The candidate driving action with the highest basic decision probability (excluding the candidate driving action corresponding to the current target driving action information) is typically a more conservative candidate driving action compared to the current target driving action information, such as changing lanes to observe or slowing down to yield; its corresponding proposed driving trajectory risk assessment value is usually lower than the preset safety threshold. In some embodiments, the electronic device may also reassess the risk of the new target driving action information to ensure its safety before execution.

[0150] For example, if the current target driving action information is an overtaking maneuver, and the risk assessment value of the proposed overtaking trajectory exceeds a preset safety threshold, the electronic equipment will prohibit overtaking and downgrade to a conservative decision to detour and observe. After downgrading to the new target driving action information, proceed to step 312 to execute the decision and adjust the execution speed.

[0151] 312: Control the target vehicle's movement based on target driving action information and speed execution strategy.

[0152] In some embodiments, when the risk assessment value is less than or equal to a preset safety threshold, the electronic device controls the target vehicle to perform the corresponding driving operation based on the target driving action information; when the risk assessment value is greater than the preset safety threshold, the electronic device controls the target vehicle to perform the corresponding driving operation based on the downgraded target driving action information. During the execution of the driving operation, the electronic device dynamically adjusts the execution speed according to the type of target driving action information and the current traffic environment through a speed execution strategy, such as reducing the execution speed in congested scenarios and increasing the execution speed in smooth scenarios, to ensure the smoothness and safety of the driving operation.

[0153] After making a decision and adjusting the execution speed, return to step 301 to continue acquiring information about the surrounding environment of the target vehicle and proceed with the next round of perception-decision-control loop.

[0154] 313: Collect driving decision events of drivers and calculate instantaneous driving style information based on decision actions and driving characteristics.

[0155] In some embodiments, when the NOA function is not activated and the target vehicle is in manual driving mode, the electronic device continuously collects the driver's driving operations. Whenever the driver completes a preset driving action (such as overtaking, yielding, or changing lanes), the electronic device records the scenario type corresponding to the driving action and the decision action selected by the driver, and stores it as a historical driving decision event.

[0156] For a current historical driving decision event, the electronic device calculates the instantaneous driving style information corresponding to the event based on the action type of the decision action and the driving characteristic information during the execution of the decision action. The instantaneous driving style information is used to characterize the driving tendency of the driver in a single driving decision event. Its calculation method can refer to the κ_game calculation formula in step 204 above, and will not be repeated here.

[0157] 314: Update and store the historical driving style information K_style for the corresponding scenario based on the instantaneous driving style information.

[0158] In some embodiments, the electronic device obtains updated historical driving style information by using a moving average update based on instantaneous driving style information and existing historical driving style information under the corresponding scenario type. The updated historical driving style information is then associated with and stored with the corresponding scenario type for personalized decision-making when the NOA function is activated subsequently. The specific method of the moving average update and the adjustment of the adaptive learning rate can be found in the K_style update formula in step 204 above, and will not be repeated here.

[0159] After the update is complete, return to step 301 to continue acquiring information about the surrounding environment of the target vehicle in order to continuously learn the driving style.

[0160] It is understandable that, through the aforementioned vehicle control process, the electronic device continuously learns the driver's historical driving style information in different scenarios under manual driving mode. When the NOA function is activated, it combines the target interaction intensity information of surrounding vehicles to personalize the output of the first model, and ensures the safety of the decision through risk field safety verification. When the risk exceeds the threshold, it downgrades to a conservative / suboptimal decision before execution. In this way, the driving decision of the NOA system can not only adapt to the driver's personal driving style, but also dynamically adjust according to the interaction situation of surrounding vehicles, while ensuring driving safety. This effectively avoids the problems of decision mismatch with user preferences and frequent user takeover caused by all drivers sharing a unified decision logic in existing technologies, thus improving the user experience and user trust of driving assistance functions.

[0161] This application provides an electronic device, including one or more processors and one or more memories; the one or more memories store one or more programs, and when one or more programs are executed by the processor, the electronic device performs the vehicle control method mentioned in this application.

[0162] It should be noted that the electronic device mentioned in this application can be any vehicle equipment such as an in-vehicle computer, vehicle infotainment system, or in-vehicle electronic device. Alternatively, the electronic device mentioned in this application can also be any terminal device capable of data interaction with the vehicle, such as a mobile phone, wearable device, tablet computer, or terminal in self-driving vehicles. This application does not limit the specific form of the electronic device in its embodiments.

[0163] The functional modules of the electronic device are briefly described below. In some embodiments, the electronic device may include:

[0164] A perception module is used to acquire information about the surrounding environment of the target vehicle. In some embodiments, the perception module may include an environment and traffic participant perception module and a navigation and road structure perception module. The environment and traffic participant perception module is used to acquire traffic participant information. For example, the environment and traffic participant perception module is connected to sensors such as a forward-looking camera, a surround-view camera, an angle radar, and a lidar to acquire information such as the position, speed, acceleration, yaw rate, and vehicle type of surrounding vehicles, and to calculate the interactive dynamic characteristics of surrounding vehicles in real time, such as the cut-in angle, time to collision (TTC), headway (THW), and its rate of change. The navigation and road structure perception module is used to acquire road structure information, such as determining the number of lanes, lane type, merging and merging point locations, curvature, etc., based on high-precision maps or navigation SD map information.

[0165] The first determining module is used to determine the target scene corresponding to the target vehicle and at least one potential interactive vehicle corresponding to the target vehicle based on the surrounding environment information.

[0166] The second determining module is used to determine the target interaction intensity information corresponding to at least one potential interactive vehicle based on the feature information of at least one potential interactive vehicle.

[0167] A decision module is used to determine target driving action information based on target interaction intensity information and the historical driving style information of the driver of the target vehicle in the target scenario. In some embodiments, the decision module includes an end-to-end decision network and a personalized adaptation layer. The end-to-end decision network is used to output first driving tendency information (i.e., the basic decision probability corresponding to each candidate driving action) based on surrounding environment information. The personalized adaptation layer is used to dynamically correct the output of the end-to-end decision network based on historical driving style information K_style and target interaction intensity information K_agg to determine second driving tendency information, and then determine the target driving action information based on the second driving tendency information.

[0168] The control module is used to control the driving of the target vehicle based on the target driving action information.

[0169] The driver operation recording and scene segmentation module is used to continuously collect the driver's driving operations in manual driving mode. Whenever the driver completes a preset driving action (such as overtaking, yielding, or changing lanes), the module records the scene type corresponding to the driving action and the decision action selected by the driver. Each decision event is labeled with a scene tag for subsequent driving style learning.

[0170] The driver interaction game style learning module is used to determine the driver's historical driving style information in different scenarios based on the driver's operation records and the historical driving decision events recorded by the scenario segmentation module. Specifically, the driver interaction game style learning module internally runs a style quantification model based on exponential moving average or Bayesian update, divides the scenario into different traffic flow density and relative speed ranges, and statistically analyzes the driver's historical driving style information K_style (with a value ranging from 0 to 1, where closer to 0 indicates a more conservative driving style and a tendency to yield, and closer to 1 indicates a more aggressive driving style and a tendency to overtake) in the scenario, and independently maintains the corresponding historical driving style information for each scenario type.

[0171] This application provides a vehicle including an electronic device for executing the vehicle control method mentioned in this application. In some embodiments, the electronic device may be a computer system within the vehicle.

[0172] For example, Figure 4 According to some embodiments of this application, a functional framework diagram of a vehicle is shown. For example... Figure 4 As shown, the functional framework of vehicle 00 may include various subsystems, such as Figure 4 The sensor system 10, control system 20, and one or more peripheral devices 30 shown are described. Figure 4 (Taking one as an example) , power supply 40, computer system 50. Optionally, vehicle 00 may also include other functional systems, such as an engine system that provides power to vehicle 00, etc., which are not limited in this application.

[0173] The sensor system 10 may include several detection devices that can sense the measured information and convert the sensed information into electrical signals or other required forms of information output according to a certain rule. For example... Figure 4 As shown, these detection devices may include a global positioning system (GPS), a vehicle speed sensor (12), an inertial measurement unit (IMU), etc., which are not limited in this application.

[0174] The Global Positioning System (GPS) 11 is a system that uses GPS positioning satellites to perform real-time positioning and navigation globally. In this application, the GPS 11 system can be used to achieve real-time positioning of vehicle 00, providing the geographical location information of vehicle 00. The vehicle speed sensor 12 is used to detect the driving speed of vehicle 00.

[0175] The inertial measurement unit 13 may include a combination of an accelerometer and a gyroscope, and is a device for measuring the angular rate and acceleration of the vehicle 00. For example, during the movement of the vehicle 00, the inertial measurement unit 13 can measure the changes in the position and angle of the vehicle body based on the inertial acceleration of the vehicle 00, such as measuring the acceleration and angular rate of the vehicle 00.

[0176] The control system 20 may include a steering unit 21, a braking unit 22, etc.

[0177] Steering unit 21 can represent a system for adjusting the direction of travel of vehicle 00, which may include, but is not limited to, a steering wheel or other structural devices for adjusting or controlling the direction of travel of vehicle 00. Braking unit 22 can represent a system for slowing down the speed of vehicle 00, and may also be called a vehicle braking system. It may include, but is not limited to, a brake controller, a reducer, or other structural devices for slowing down vehicle 00. In practical applications, braking unit 22 can use friction to slow down the tires of vehicle 00, thereby slowing down the speed of vehicle 00.

[0178] Peripheral device 30 may include several components, such as Figure 4The diagram shows a communication system 31, a touchscreen 32, a user interface 33, etc. The communication system 31 is used to enable network communication between the vehicle 00 and other devices besides the vehicle 00. In practical applications, the communication system 31 can employ wireless communication technology or wired communication technology to achieve network communication between the vehicle 00 and other devices. This wired communication technology can refer to communication between the vehicle 00 and other devices via network cables or optical fibers. This wireless communication technology includes, but is not limited to, Global System for Mobile Communications (GSM), General Packet Radio Service (GPRS), Code Division Multiple Access (CDMA), Wide Band Code Division Multiple Access (WCDMA), Time Division Code Division Multiple Access (TD-SCDMA), Long Term Evolution (LTE), Wireless Local Area Networks (WLAN) (such as Wireless Fidelity (Wi-Fi) networks), Bluetooth (BT), Global Navigation Satellite System (GNSS), Frequency Modulation (FM), Near Field Communication (NFC), and Infrared (IR) technology, etc.

[0179] The touchscreen 32 can be used to detect operation commands on the touchscreen 32. For example, the user can perform touch operations on the content data displayed on the touchscreen 32 according to actual needs to achieve the corresponding function, such as playing music, video, or other multimedia files. The user interface 33 can specifically be a touch panel, used to detect operation commands on the touch panel. The user interface 33 can also be a physical button or a mouse. The user interface 33 can also be a display screen, used to output data and display images or data. Optionally, the user interface 33 can also be at least one device belonging to the category of peripheral devices, such as a touchscreen, microphone, and speaker.

[0180] Several functions of vehicle 00 are controlled and implemented by computer system 50. Computer system 50 may include multiple processors such as processor 51, continuous damping control (CDC) 52, mobile data center (MDC) 53, telematics box (T-BOX) 54, as well as memory 55 (also referred to as storage device) and gateway 56. In practical applications, memory 55 may be located inside computer system 50 or outside computer system 50, for example, as a cache in vehicle 00, etc., and this application does not limit this.

[0181] Among them, processor 51 can be a graphics processing unit (GPU), etc. Processor 51, CDC 52, MDC 53, and T-BOX 54 can be used to run relevant programs or corresponding instructions stored in memory 55 to realize the corresponding functions of vehicle 00, such as network switching function based on service.

[0182] The memory 55 may include volatile memory, such as RAM; it may also include non-volatile memory, such as ROM, flash memory, HDD, or SSD; or it may include a combination of the above types of memory. The memory 55 can be used to store a set of program code or instructions corresponding to program code, so that the processor 51 can call the program code or instructions stored in the memory 55 to implement the corresponding functions of the vehicle 00. This function includes, but is not limited to, […]. Figure 4 The functional framework diagram shown includes some or all of the functions. In this application, the memory 55 can store a set of program code for vehicle control. The processor 51, CDC 52, MDC 53, and T-BOX 54 can call this program code to control the vehicle 00 to execute the vehicle control method mentioned in the embodiments of this application.

[0183] Optionally, in addition to storing program code or instructions, memory 55 may also store information such as road maps, driving routes, and sensor data. Computer system 50 can be combined with other components in the vehicle functional framework diagram, such as sensors in the sensor system and GPS, to realize the relevant functions of vehicle 00. For example, computer system 50 can control the driving direction or speed of vehicle 00 based on data input from sensor system 10; this application does not limit this.

[0184] It should be noted that the above Figure 4This is merely a schematic diagram of one possible functional framework for vehicle 00. In practical applications, a vehicle may include more or fewer systems or components, and this application does not impose any limitations.

[0185] This application provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a device, the vehicle control method mentioned in this application is implemented.

[0186] This application provides a computer program product, including a program or instructions, which, when executed, cause the vehicle control method mentioned in this application to be implemented.

[0187] This application provides a chip that includes a program or instructions. When the program or instructions are executed, the vehicle control method mentioned in this application is implemented.

[0188] Various embodiments of the mechanisms disclosed in this application can be implemented in hardware, software, firmware, or combinations of these implementation methods. Embodiments of this application can be implemented as computer programs or program code executable on a programmable system, the programmable system including at least one processor, a storage system (including volatile and non-volatile memory and / or storage elements), at least one input device, and at least one output device.

[0189] Program code can be applied to input instructions to execute the functions described in this application and generate output information. The output information can be applied to one or more output devices in a known manner. For the purposes of this application, the processing system includes any system having a processor such as, for example, a digital signal processor, a microcontroller, an application-specific integrated circuit, or a microprocessor.

[0190] The program code can be implemented using a high-level procedural language or an object-oriented programming language to communicate with the processing system. Assembly language or machine language can also be used when needed. The mechanisms described in this application are not limited to any particular programming language. In either case, the language can be a compiled language or an interpreted language.

[0191] In some cases, the disclosed embodiments may be implemented in hardware, firmware, software, or any combination thereof. The disclosed embodiments may also be implemented as instructions carried or stored on or on one or more temporary or non-temporary machine-readable (e.g., computer-readable) storage media, which may be read and executed by one or more processors. For example, the instructions may be distributed via a network or through other computer-readable media. Therefore, machine-readable media may include any mechanism for storing or transmitting information in a machine-readable (e.g., computer-readable) form, including but not limited to floppy disks, optical disks, optical discs, read-only memory, magneto-optical disks, random access memory, erasable programmable read-only memory, electrically erasable programmable read-only memory, magnetic cards or optical cards, flash memory, or tangible machine-readable storage for transmitting information (e.g., carrier waves, infrared signals, digital signals, etc.) using the Internet in the form of electrical, optical, acoustic, or other propagation signals. Therefore, machine-readable media include any type of machine-readable medium suitable for storing or transmitting electronic instructions or information in a machine-readable (e.g., computer-readable) form.

[0192] In the accompanying drawings, some structural or methodological features may be shown in a specific arrangement and / or order. However, it should be understood that such a specific arrangement and / or order may not be necessary. Rather, in some embodiments, these features may be arranged in a manner and / or order different from that shown in the illustrative drawings. Furthermore, the inclusion of structural or methodological features in a particular figure does not imply that such features are required in all embodiments, and in some embodiments, these features may be omitted or may be combined with other features.

[0193] It should be noted that all units / modules mentioned in the device embodiments of this application are logical units / modules. Physically, a logical unit / module can be a physical unit / module, a part of a physical unit / module, or a combination of multiple physical units / modules. The physical implementation of these logical units / modules themselves is not the most important factor; the combination of functions implemented by these logical units / modules is the key to solving the technical problems proposed in this application. Furthermore, to highlight the innovative aspects of this application, the above-described device embodiments of this application have not introduced units / modules that are not closely related to solving the technical problems proposed in this application. This does not mean that the above-described device embodiments do not contain other units / modules.

[0194] It should be noted that in the examples and description of this patent, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0195] Although this application has been illustrated and described with reference to certain preferred embodiments thereof, those skilled in the art should understand that various changes in form and detail may be made thereto without departing from the spirit and scope of this application.

Claims

1. A vehicle control method, characterized in that, For use in electronic devices, the method includes: Obtain information about the surrounding environment of the target vehicle; Based on the surrounding environment information, determine the target scene corresponding to the target vehicle and at least one potential interactive vehicle corresponding to the target vehicle; Based on the feature information of the at least one potential interactive vehicle, determine the target interaction intensity information corresponding to the at least one potential interactive vehicle; Target driving action information is determined based on the target interaction intensity information and the historical driving style information of the driver of the target vehicle in the target scenario; The target vehicle is controlled to drive based on the target driving action information.

2. The method according to claim 1, characterized in that, The target scenarios include any of the following: adjacent lane vehicle insertion scenario, overtaking scenario, ramp merging scenario, and congestion interaction scenario.

3. The method according to claim 1, characterized in that, The characteristic information of the potential interactive vehicle includes at least one of the relative speed of the potential interactive vehicle relative to the target vehicle, lateral acceleration, and collision time change rate. Determining the target interaction intensity information corresponding to the at least one potential interactive vehicle based on the feature information of the at least one potential interactive vehicle includes: The feature information of each potential interactive vehicle in the at least one potential interactive vehicle is weighted and fused to determine the interaction intensity information of each potential interactive vehicle. The target interaction intensity information is determined based on the interaction intensity information of each of the potential interactive vehicles.

4. The method according to claim 3, characterized in that, Determining the target interaction intensity information based on the interaction intensity information of each of the potential interactive vehicles includes: When there are multiple potential interactive vehicles, the maximum value of the interaction intensity information corresponding to each potential interactive vehicle is determined as the target interaction intensity information. When there is only one potential interactive vehicle, the interaction intensity information corresponding to the potential interactive vehicle is determined as the target interaction intensity information.

5. The method according to any one of claims 1-4, characterized in that, The historical driving style information of the driver of the target vehicle in the target scenario is determined based on the driver's historical driving behavior in the target scenario and in manual driving mode.

6. The method according to any one of claims 1-4, characterized in that, The determination of target driving action information based on the target interaction intensity information and the historical driving style information of the driver of the target vehicle in the target scenario includes: The surrounding environment information of the target vehicle is input into the first model to obtain the first driving tendency information output by the first model; Based on the historical driving style information and the target interaction intensity information, driving action adjustment information is determined; Based on the driving action adjustment information, the first driving tendency information is adjusted to obtain the second driving tendency information; The target driving action information is determined based on the second driving tendency information.

7. The method according to claim 6, characterized in that, The first driving tendency information includes the basic decision probabilities corresponding to multiple candidate driving actions, and the second driving tendency information includes the target decision probabilities corresponding to the multiple candidate driving actions.

8. The method according to claim 7, characterized in that, Determining the target driving action information based on the second driving tendency information includes: The candidate driving action with the highest probability of target decision among the multiple candidate driving actions is determined as the target driving action information.

9. The method according to claim 7 or 8, characterized in that, The candidate driving actions include overtaking, yielding, and changing lanes.

10. The method according to claim 9, characterized in that, The method of controlling the target vehicle's movement based on the target driving action information includes: Based on the proposed driving trajectory corresponding to the target driving action information, obtain the risk assessment value corresponding to the proposed driving trajectory; If the risk assessment value is less than or equal to a preset safety threshold, the target vehicle is controlled to drive based on the target driving action information.

11. The method according to claim 10, characterized in that, The method further includes: If the risk assessment value is greater than the preset safety threshold, the candidate driving action with the highest basic decision probability among the first driving tendency information, excluding the candidate driving action corresponding to the target driving action information, is determined as the updated target driving action information.

12. An electronic device, characterized in that, include: One or more processors; One or more memories; the one or more memories storing one or more programs that, when executed by the processor, cause the electronic device to perform the method of any one of claims 1-11.

13. A vehicle, characterized in that, Includes an electronic device for performing the method according to any one of claims 1-11.

14. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a program or instructions that, when executed by an electronic device, implement the method as claimed in any one of claims 1-11.

15. A computer program product, characterized in that, Includes instructions that, when executed, cause the method of any one of claims 1-11 to be implemented.

Images