TJA control misjudgment reduction method and device for night curve cruise

Abstract

The invention provides a TJA control misjudgment reduction method and device for night curve cruise, and relates to the field of auxiliary intelligent driving. The method comprises the steps of obtaining multi-source data and inputting the multi-source data into a multi-filtering strategy model; first filtering operation is carried out through the multi-filtering strategy model so as to eliminate false targets and first misrecognition vehicles in the target to-be-recognized vehicles; and calculating a motion fluctuation characteristic threshold value corresponding to the remaining target to-be-recognized vehicles based on the multi-source data through second filtering operation, judging the remaining target to-be-recognized vehicles as effective TJA control targets when determining that the motion fluctuation characteristic threshold value is smaller than a preset reference threshold value, and performing TJA braking operation based on the effective TJA control targets. The problem that the driving experience of a driver is poor due to the fact that existing TJA braking control is prone to misjudgment in a night curve scene is solved.

Description

Technical Field

[0001] This application relates to the field of assisted driving, and in particular to a method and apparatus for reducing TJA control misjudgment during nighttime curve cruise. Background Technology

[0002] Currently, Traffic Jam Assist (TJA) braking, as an important driver assistance function in low-speed congestion scenarios and some curve cruise scenarios, has been widely used in mass-produced vehicles. Its core objective is to achieve smooth following and automatic braking in complex traffic environments, thereby reducing the driver's operational burden in congested areas. How to achieve precise, stable, and error-free TJA braking control under different road structures, lighting conditions, and traffic behaviors has become an important research direction in the field of intelligent driving.

[0003] Existing technologies typically trigger TJA braking by monitoring the relative distance, relative speed, and lane space structure of the vehicle ahead. This method usually relies on millimeter-wave radar and cameras to continuously track the target vehicle ahead, calculating the rate of change in distance, speed difference, and vision-based lane position relationships to determine whether braking is necessary, thus achieving automatic following. However, this type of method performs well on straight roads or in well-lit scenarios with high information acquisition stability, but is significantly limited in low-light conditions or on roads with complex morphology.

[0004] Especially in nighttime curves, the significantly reduced light intensity degrades camera image quality, causing the outline of the vehicle in front to become fragmented or blurred. Simultaneously, the spatial structure of the curve causes overlapping projections of vehicles in adjacent lanes in the image, making it easy to misidentify the target's lateral position and heading angle. Furthermore, vehicles typically experience some degree of yaw angle change in curves, making it easier to lose the correlation between the predicted trajectory and the perceived target. The combination of these factors significantly increases the probability of false targets or abnormally close targets being misidentified as real vehicles in front, causing the TJA system to trigger unnecessary braking actions in curves, resulting in a poor driving experience for the user.

[0005] Therefore, there is an urgent need for a method and device to reduce TJA control misjudgment for nighttime curve cruise. Summary of the Invention

[0006] This application provides a method and apparatus for reducing misjudgment in TJA control for nighttime curve cruise, which solves the problem that existing TJA braking control is prone to misjudgment in nighttime curve scenarios, resulting in a poor driving experience for the driver.

[0007] The first aspect of this application provides a method for reducing TJA control misjudgment during nighttime curve cruise. The method includes: acquiring multi-source data corresponding to a target vehicle to be identified, a target vehicle, and the nighttime curve environment; inputting the multi-source data into a multi-filtering strategy model and performing a first filtering operation through the multi-filtering strategy model; identifying false targets and first misidentified vehicles among the target vehicles to be identified through the first filtering operation; removing false targets and first misidentified vehicles from the target vehicles to be identified, and performing a second filtering operation on the remaining target vehicles to be identified; calculating a motion fluctuation feature threshold corresponding to the remaining target vehicles to be identified based on the second filtering operation and the multi-source data, and determining whether the motion fluctuation feature threshold is less than a preset reference threshold; if it is confirmed that the motion fluctuation feature threshold is less than the preset reference threshold, determining that the remaining target vehicles to be identified are valid TJA control targets, and performing TJA braking operation based on the valid TJA control targets.

[0008] Optionally, multi-source data corresponding to the target vehicle to be identified, the target self-vehicle, and the nighttime curve environment are acquired, specifically including: acquiring the full perception data of the preceding vehicle corresponding to the target vehicle to be identified, the self-vehicle motion data corresponding to the target self-vehicle, and the nighttime curve environment data corresponding to the nighttime curve environment; and using the set of the full perception data of the preceding vehicle, the self-vehicle motion data, and the nighttime curve environment data as multi-source data.

[0009] Optionally, after acquiring multi-source data corresponding to the target vehicle to be identified, the target vehicle, and the nighttime curve environment, the method further includes: predicting the vehicle trajectory based on the vehicle motion data and constructing a path funnel on the vehicle trajectory; filtering the target vehicle to be identified on the path funnel and selecting the TJA control target.

[0010] Optionally, the first filtering operation identifies false targets among the target vehicles to be identified, specifically including: acquiring a first vehicle to be identified, wherein the first vehicle to be identified is any vehicle among the target vehicles to be identified; determining whether the on-site light intensity value is lower than a preset light intensity value based on the light intensity data in the nighttime curve environment data; determining whether the confidence value corresponding to the target vehicle to be identified is lower than a preset confidence value based on the point cloud image data in the full perception data of the preceding vehicle; and identifying the first vehicle to be identified whose on-site light intensity value is lower than the preset light intensity value or whose confidence value is lower than the preset confidence value as a false target.

[0011] Optionally, the first filtering operation identifies the remaining target vehicles to be identified as the first misidentified vehicle, specifically including: acquiring a second vehicle to be identified, which is any vehicle among the remaining target vehicles to be identified; determining whether the distance of the second vehicle to be identified is less than or equal to a preset distance corresponding to the vehicle type based on vehicle type data and vehicle distance data in the full perception data of the preceding vehicle; determining whether the driving speed of the target vehicle is greater than or equal to a preset driving speed based on the driving speed data of the vehicle in the vehicle motion data; determining whether the lane curvature is within a preset curvature range based on the lane curvature data of the nighttime curve environment data; if it is confirmed that the vehicle distance is less than or equal to the preset distance corresponding to the vehicle type, or that the driving speed is greater than or equal to the preset driving speed, or that the lane curvature is within a preset curvature range, then the second vehicle to be identified is identified as the first misidentified vehicle.

[0012] Optionally, the motion fluctuation feature threshold includes a longitudinal distance fluctuation threshold and a speed fluctuation standard deviation threshold. The motion fluctuation feature threshold for the remaining target vehicles to be identified is calculated based on multi-source data through a second filtering operation, and it is determined whether the motion fluctuation feature threshold is less than a preset reference threshold. Specifically, this includes: acquiring a third target vehicle, which is any vehicle among the remaining target vehicles to be identified; calculating the longitudinal distance fluctuation threshold based on the target longitudinal distance sequence in the full perception data of the preceding vehicle; calculating the speed fluctuation standard deviation threshold based on the target speed sequence in the full perception data of the preceding vehicle; determining whether the motion fluctuation feature threshold is less than a preset distance fluctuation threshold, and determining whether the speed fluctuation standard deviation threshold is less than a preset speed fluctuation threshold.

[0013] Optionally, after determining whether the motion fluctuation feature threshold is less than the preset distance fluctuation threshold and whether the speed fluctuation standard deviation threshold is less than the preset speed fluctuation threshold, the method further includes: if it is confirmed that the motion fluctuation feature threshold is greater than or equal to the preset distance fluctuation threshold and the speed fluctuation standard deviation threshold is greater than or equal to the preset speed fluctuation threshold, then the third vehicle to be identified is regarded as the second misidentified vehicle; the second misidentified vehicle is removed from the remaining target vehicles to be identified.

[0014] A second aspect of this application provides a TJA control misjudgment reduction device for nighttime curve cruise, the device including an acquisition module and a processing module, wherein, The acquisition module is used to acquire multi-source data corresponding to the target vehicle to be identified, the target vehicle itself, and the nighttime curve environment.

[0015] The processing module is used to input multi-source data into a multi-filtering strategy model and perform a first filtering operation through the multi-filtering strategy model; identify false targets and first misidentified vehicles among the target vehicles to be identified through the first filtering operation; remove false targets and first misidentified vehicles from the target vehicles to be identified, and perform a second filtering operation on the remaining target vehicles to be identified; calculate the motion fluctuation feature threshold corresponding to the remaining target vehicles to be identified through the second filtering operation and based on the multi-source data, and determine whether the motion fluctuation feature threshold is less than a preset reference threshold; if it is confirmed that the motion fluctuation feature threshold is less than the preset reference threshold, then determine that the remaining target vehicles to be identified are valid TJA control targets, and perform TJA braking operation based on the valid TJA control targets.

[0016] A third aspect of this application provides an electronic device including a processor, a memory, a user interface, and a network interface. The memory is used to store instructions, the user interface and the network interface are used to communicate with other devices, and the processor is used to execute the instructions stored in the memory to cause the electronic device to perform the method as described above.

[0017] A fourth aspect of this application provides a computer-readable storage medium storing a computer program, which is executed by a processor using the method described in any of the foregoing descriptions.

[0018] One or more technical solutions provided in the embodiments of this application have at least the following technical effects or advantages: 1. Acquire multi-source data and input it into a multi-filtering strategy model; perform a first filtering operation on the target vehicles to be identified using the multi-filtering strategy model; remove the first misidentified vehicle from the remaining target vehicles to be identified using the first filtering operation; perform a second filtering operation and calculate the motion fluctuation feature threshold corresponding to the remaining target vehicles to be identified based on the multi-source data, and when it is confirmed that the motion fluctuation feature threshold is less than the preset reference threshold, determine that the remaining target vehicles to be identified are valid TJA control targets, and perform TJA braking operation based on the valid TJA control targets. In this way, in nighttime curve cruise scenarios, it effectively filters out false targets and unstable front vehicle targets introduced by insufficient light, curve geometry and interference from adjacent lanes, significantly reducing the misjudgment rate and false braking frequency of TJA control, and improving the safety, control stability and driving comfort of the automatic cruise process.

[0019] 2. Acquire the full perception data of the preceding vehicle corresponding to the target vehicle, the vehicle motion data corresponding to the target vehicle, and the nighttime curve environment data corresponding to the nighttime curve environment. Use the collection of the full perception data of the preceding vehicle, the vehicle motion data, and the nighttime curve environment data as multi-source data to build a complete scene information foundation covering the target vehicle, the dynamic behavior of the vehicle, and the characteristics of the curve environment in the nighttime curve cruise scenario. This provides stable, continuous, and structured data support for subsequent multi-filtering strategy models, ensuring sufficient input basis for false target identification, abnormal approach vehicle elimination, and motion fluctuation feature judgment, further guaranteeing the accurate screening of TJA control targets and the reliable execution of TJA braking operations.

[0020] 3. By constructing a multi-filtering strategy model and performing three filtering operations, false targets, abnormally approaching targets, and targets with unstable motion states are eliminated step by step. This ensures that only stable, reliable, and real forward targets related to the vehicle's trajectory are used as effective TJA control targets in subsequent braking decisions, thereby improving the accuracy and safety of TJA control in nighttime curve cruise scenarios. Attached Figure Description

[0021] Figure 1 This is a flowchart illustrating a method for reducing TJA control misjudgment during nighttime curve cruise, provided in an embodiment of this application. Figure 2 This is a schematic diagram of a TJA control misjudgment reduction device for nighttime curve cruise provided in an embodiment of this application; Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application.

[0022] Explanation of reference numerals in the attached figures: 21. Acquisition module; 22. Processing module; 301. Processor; 302. Communication bus; 303. User interface; 304. Network interface; 305. Memory. Detailed Implementation

[0023] To enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments.

[0024] The terminology used in the following embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. As used in the specification of this application, the singular expressions “a,” “an,” “the,” “the,” “the,” and “this” are intended to include the plural expressions as well, unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used in this application refers to and includes any or all possible combinations of one or more of the listed items.

[0025] Hereinafter, the terms "first" and "second" are used for descriptive purposes only and should not be construed as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature, and in the description of the embodiments of this application, unless otherwise stated, "multiple" means two or more.

[0026] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings.

[0027] Please refer to Figure 1 The diagram shows a flowchart of a method for reducing TJA control misjudgment during nighttime curve cruise provided in this application embodiment. The flowchart mainly includes the following steps: S101 to S106.

[0028] Step S101: Obtain multi-source data corresponding to the target vehicle to be identified, the target vehicle itself, and the nighttime curve environment.

[0029] Specifically, the operation of acquiring multi-source data corresponding to the target vehicle to be identified, the target vehicle itself, and the nighttime curve environment is applied to scenarios where a driver is driving the target vehicle on a curve at night. In this situation, insufficient nighttime lighting leads to a decrease in the quality of the perceived image, and the curve geometry causes significant curvature changes in lane lines. The vehicle may also exhibit yaw drift when turning outwards, making the outline of the detected vehicle in the image discontinuous and its position more sensitive to changes. Simultaneously, the vehicle in front may exhibit lateral deviations, cutting in and out, and significant distance fluctuations on the curve, all of which make the perception system more prone to misidentifying adjacent lanes or unstable point cloud segments as the real vehicle in front, thus triggering false braking. Therefore, to avoid these misjudgments, it is necessary to comprehensively acquire data related to the state of the vehicle in front, the vehicle's state, and the nighttime curve environment as multi-source inputs. In this application, a curve scenario refers to a road area where the vehicle's driving path has continuous or discontinuous curvature changes, the lane lines have an arc-shaped orientation, and may be accompanied by an asymmetrical structure of the inner and outer spaces. Examples include driving at night on highway ramps, curve areas with medium or small radii on urban roads, continuous curve sections on mountain roads, or compound curve roads with obvious turning transitions.

[0030] In step S101, the target vehicle to be identified is any one or more candidate vehicles that may be within the predicted path area of ​​the vehicle ahead. This includes the target vehicle detected by a camera, millimeter-wave radar, or fusion perception system. Its identity has not yet been determined as a real or false target and needs to be identified step by step in subsequent filtering steps. The target vehicle is the currently driving vehicle. Its own operating status information includes the vehicle speed, vehicle heading angle, vehicle lateral position, vehicle curvature, and dynamic stability data of the vehicle in the curve. This information is used to determine the actual driving trend of the vehicle in the curve and to determine whether the target vehicle ahead may be located within the future trajectory range of the vehicle. This information serves as an important component of the multi-source data.

[0031] In one possible implementation, step S101 further includes: acquiring the full perception data of the preceding vehicle corresponding to the target vehicle to be identified, the vehicle motion data corresponding to the target vehicle, and the nighttime curve environment data corresponding to the nighttime curve environment; and using the set of the full perception data of the preceding vehicle, the vehicle motion data, and the nighttime curve environment data as multi-source data.

[0032] Specifically, the full-scale perception data of the vehicle ahead can be acquired through, but is not limited to, forward-facing cameras, millimeter-wave radar, lidar, and multi-sensor fusion perception units. In nighttime curve scenarios, the perception unit performs target detection and tracking on multiple candidate vehicles ahead, obtaining the detection box information, category label, and spatial position of each target in the vehicle's coordinate system for each frame. Trajectory correlation is then performed in the temporal dimension to form target sequence data covering a continuous time window. The full-scale perception data of the vehicle ahead includes, but is not limited to, target category data of the vehicle ahead, used to distinguish different vehicle types such as cars, motorcycles, and trucks. The longitudinal distance and lateral offset data are used to describe the distance and lane offset of the preceding vehicle relative to the following vehicle in the forward coordinate system; the relative speed and relative acceleration data of the preceding vehicle are used to describe the change in the following proximity of the preceding vehicle in the longitudinal direction; the heading angle data of the preceding vehicle is used to represent the angle between the heading direction of the preceding vehicle and the road reference direction; the point cloud image data of the preceding vehicle is used to characterize the outline shape and spatial structure density of the preceding vehicle; and the detection confidence data and target duration data of the preceding vehicle are used to reflect the reliability of the perception system for the preceding vehicle target and the continuous existence of the target on the time axis.

[0033] Vehicle motion data can be acquired through, but is not limited to, the onboard CAN bus, the vehicle stability control system, the inertial measurement unit, and wheel speed sensors. During nighttime cornering, the vehicle controller periodically reads the target vehicle's operating status parameters and synchronizes them to the perception and decision-making timeline. Vehicle motion data includes, but is not limited to, longitudinal speed data, which characterizes the current cruising speed and the speed range that triggers different filtering strategies; lateral position offset data, which describes the degree of offset of the vehicle relative to the lane centerline; yaw angle data and yaw rate data, which describe the trend of the angle change between the vehicle's body posture and the road reference direction; curvature data, which characterizes the curvature of the vehicle's current trajectory and its consistency with the lane line curvature; and longitudinal acceleration data and braking status data, which reflect whether the vehicle is currently accelerating, moving at a constant speed, or decelerating, thus providing a basis for subsequent abnormal approach judgment and braking strategies.

[0034] Nighttime curve environment data can be acquired through, but is not limited to, ambient light sensors, lane line recognition algorithms, and high-precision map data. The environmental perception module combines real-time detection results with prior map information to model the surrounding road environment on nighttime curve sections. This nighttime curve environment data includes, but is not limited to, light intensity data to characterize the current ambient brightness level and whether it is in low-light conditions; lane line curvature data to describe the curvature radius and curvature variation trend of the curve; curve spatial geometry data to describe the number of lanes, lane width, curve direction, and the location of the inner and outer road boundaries within the curve; and distribution data of highly reflective signs or roadside facilities to assist in determining the spatial location of potential obstacles on the outer side of the curve. The aforementioned comprehensive perception data of the preceding vehicle, the vehicle's motion data, and the nighttime curve environment data are synchronized along the timeline and combined to form multi-source data. This provides a complete and continuous foundation of scene information for subsequent multi-filtering strategy models in the processes of false target identification, abnormally approaching vehicle identification, and motion fluctuation feature discrimination.

[0035] In one possible implementation, step S101 further includes: predicting the vehicle trajectory based on the vehicle motion data and constructing a path funnel on the vehicle trajectory; filtering the target vehicles to be identified on the path funnel and selecting TJA control targets.

[0036] Specifically, based on vehicle motion data such as vehicle speed, yaw angle, curvature, and lane position, a predicted trajectory of the vehicle that evolves over time or longitudinal distance is first generated in a nighttime curve scenario. This predicted trajectory is used as a reference line, and a pair of symmetrical trajectory boundaries are initialized on both sides of the reference line according to a preset lateral offset distance to obtain an initial path funnel profile. Subsequently, combined with lane curvature data in the nighttime curve environment data, the path funnel is subjected to the first layer of geometric correction. When the curve curvature increases, the direction of the inside and outside of the curve is identified, the path boundary on the outside of the curve is tightened, and the path boundary on the inside of the curve is appropriately widened, so that the path funnel presents an asymmetrical shape in the curve area to reflect the physical characteristics that the vehicle is more likely to deviate to the outside in the curve and that there is a higher risk of crossing the boundary on the outside.

[0037] After completing the geometric correction based on the curvature of the curve, a second layer of dynamic correction is performed on the path funnel by combining the vehicle's yaw angle and the trend of yaw angle changes. When the yaw angle error of the vehicle increases or the lateral drift trend of the vehicle intensifies, it is assumed that there is a deviation between the vehicle's attitude and the road reference direction. At this time, the boundary width of the path funnel is tightened on both sides at the same time. By reducing the available lateral space, the subsequent trajectory planning is forced to be more conservative, thereby reducing the risk of going out of bounds caused by the superposition of attitude error and curve geometry. After the yaw angle error recovers to a small range and the vehicle attitude is realigned with the road reference direction, the normal width of the path funnel is gradually restored. The path funnel follows the principle of "difficult to enter and difficult to exit" during the dynamic tightening and restoration process. That is, the action is faster when the boundary is tightened and smoother when it is widened, so as to avoid the problem of oscillation or frequent entry and exit of the target in the funnel during the frequent tightening and widening of the boundary.

[0038] Based on the path funnel corrected by the vehicle's state and curve geometry, the position of the target vehicle is mapped to the vehicle's coordinate system. Combined with the lateral offset data and target heading angle data from the full perception data of the preceding vehicle, the third layer of target-related path funnel refinement is completed: when the lateral position of a target vehicle falls entirely in the adjacent lane and is always outside the funnel, it is determined that the target has no direct conflict with the vehicle's reference driving path and is directly marked as a background target unrelated to the vehicle's trajectory at the path funnel level; when the target partially intrudes into the vehicle's lane or is near the lane boundary, the funnel boundary on one side of the target is further compressed to a safe distance position close to the target's outline, so that the path funnel forms a local "bottleneck" at the target position. Combined with the target heading angle, it is determined whether the target has a tendency to cut into or out of the vehicle's lane. For targets with a cutting-in tendency, the funnel boundary on the side closer to it is further contracted and the longitudinal influence range is extended. For targets with a cutting-out tendency, the corresponding boundary is appropriately widened while ensuring a safe distance. Through the above multi-level corrections, a path funnel structure that matches the nighttime curve geometry, the vehicle's dynamic attitude, and the preceding vehicle's behavioral intentions is finally obtained. Within the channel defined by this path funnel, the spatial and temporal evolution relationships of all target vehicles to be identified are determined. Only vehicles whose trajectories are located inside the funnel for a long time and have a potential following or collision relationship with the vehicle's future driving path are marked as candidate TJA control targets. Vehicles located outside the funnel or that only enter momentarily but do not have a continuous correlation are removed from the control target set. In this way, in the subsequent multi-filtering strategy model and TJA longitudinal control process, the misidentification and false braking caused by adjacent lane targets, distant targets, or targets with unstable trajectories in nighttime curve scenarios are effectively reduced.

[0039] Step S102: Input multi-source data into the multi-filtering strategy model, and perform the first filtering operation through the multi-filtering strategy model.

[0040] Specifically, the multi-filtering strategy model consists of a first filtering strategy, a second filtering strategy, and a third filtering strategy that act sequentially. The first filtering strategy, the second filtering strategy, and the third filtering strategy are used to perform the first filtering operation, the first filtering operation, and the second filtering operation, respectively. The first filtering strategy is used to perform preliminary reliability screening of the target vehicle to be identified based on the reduced perception reliability caused by insufficient light in the nighttime curve environment and the unstable confidence in the perception data of the preceding vehicle. This identifies false targets caused by weak light, incomplete outlines, or insufficient detection continuity, so as to avoid such targets from entering the subsequent decision chain and reduce the risk of false braking caused by misidentification in the nighttime curve section.

[0041] Step S103: Identify false targets and first misidentified vehicles among the target vehicles through the first filtering operation.

[0042] Specifically, based on the light intensity data in the nighttime curve environment and the confidence and target continuity information in the full perception data of the preceding vehicle, a preliminary confidence judgment is made for each target vehicle to be identified. In cases where insufficient light or low target confidence leads to unstable perception results, targets that cannot be continuously and stably detected are marked as false targets. This eliminates non-real preceding vehicle targets caused by dark environment, incomplete outline, or discontinuous detection before entering the subsequent filtering process, thus avoiding interference from such false targets to subsequent following judgment and TJA control.

[0043] In one possible implementation, step S103 further includes: acquiring a first vehicle to be identified, wherein the first vehicle to be identified is any vehicle among the target vehicles to be identified; determining whether the on-site light intensity value is lower than a preset light intensity value based on the light intensity data in the nighttime curve environment data; determining whether the confidence value corresponding to the target vehicle to be identified is lower than a preset confidence value based on the point cloud image data in the full perception data of the preceding vehicle; and identifying the first vehicle to be identified whose on-site light intensity value is lower than the preset light intensity value or whose confidence value is lower than the preset confidence value as a false target.

[0044] Specifically, after acquiring the first vehicle to be identified in a nighttime curve scenario, the current road environment's light intensity is obtained through a light sensor or camera brightness analysis module. This light intensity is then compared with a preset minimum identifiable light threshold. When the light level is lower than this preset light intensity value, it indicates increased imaging noise in the visual perception link, and missing, blurred, or broken point cloud or image features, making it easier for the target ahead to have incomplete outlines. Subsequently, the confidence level of the first vehicle to be identified is judged by combining the point cloud image data from the full perception data of the preceding vehicle. When the confidence level is less than the preset confidence threshold, it indicates that the detection results of the vehicle in the perception system are discontinuous, the target association stability is poor, or the spatial structure exhibits abnormal jumps. Such targets cannot maintain a stable existence in continuous frames. If either of the above two unstable factors is present, the first vehicle to be identified is determined to be a false target and removed from the target vehicle identification set. This avoids visual noise or occasional point cloud fragments being mistakenly treated as real vehicles in subsequent vehicle following judgments and participating in TJA control target selection, thereby improving the reliability of target recognition and control safety in nighttime curve scenarios.

[0045] In one possible implementation, step S103 further includes: acquiring a second vehicle to be identified, wherein the second vehicle to be identified is any vehicle among the remaining target vehicles to be identified; determining whether the vehicle distance of the second vehicle to be identified is less than or equal to a preset distance corresponding to the vehicle type based on vehicle type data and vehicle distance data in the full perception data of the preceding vehicle; determining whether the driving speed of the target vehicle is greater than or equal to a preset driving speed based on driving speed data in the vehicle motion data; determining whether the lane curvature is within a preset curvature range based on lane curvature data in the nighttime curve environment data; and if it is confirmed that the vehicle distance is less than or equal to the preset distance corresponding to the vehicle type, or that the driving speed is greater than or equal to the preset driving speed, or that the lane curvature is within a preset curvature range, then the second vehicle to be identified is designated as the first misidentified vehicle.

[0046] Specifically, a second vehicle to be identified is acquired. This second vehicle can be any vehicle among the remaining target vehicles. For this second vehicle, its vehicle type and distance data are read from the full-scale perception data of the preceding vehicle. These are then compared against preset safety distance thresholds corresponding to different vehicle types. For example, the preset distance can be set to no more than 50m for a sedan and no more than 60m for a motorcycle or truck. By determining whether the distance to the second vehicle is less than or equal to the preset distance corresponding to its vehicle type, a target that is abnormally close to the vehicle's longitudinal distance in the current nighttime curve scenario is identified. Simultaneously, the vehicle's speed data is read from the vehicle's motion data to determine whether the target vehicle's speed is greater than... Alternatively, it can be equal to a preset vehicle speed threshold, such as 30 kph. When the vehicle speed is not lower than 30 kph, it is considered to be in a medium-to-high speed driving state. At this time, any longitudinal abnormality approaching the vehicle has a more significant impact on safety. Furthermore, lane curvature data is read from the nighttime curve environment data to determine whether the current lane curvature is within the preset curve curvature range. For example, by limiting the curve radius R to between 150m and 1000m, when the curve radius corresponding to the lane curvature falls within the range of 150≤R≤1000m, it is considered to be in a medium to sharp curve scenario. In this type of curve, the vehicle is more likely to deviate to the outside due to centrifugal force, and the target in the adjacent lane is more likely to visually overlap with the target in this lane, thereby amplifying the risk of misidentification. If it is confirmed that the distance to the vehicle is less than or equal to the corresponding preset distance, such as a car distance less than or equal to 50m or a motorcycle or truck distance less than or equal to 60m, or if it is confirmed that the target vehicle's speed is greater than or equal to 30kph, or if it is confirmed that the current lane curvature is within the range of a curve radius of 150m to 1000m, then the second vehicle to be identified is considered to have a high risk of false triggering under any of the above conditions. The second vehicle to be identified is marked as the first falsely identified vehicle and removed from the set of remaining target vehicles in subsequent processing. This is to avoid misidentifying high-risk targets that are not suitable as stable TJA following targets as effective control targets under the combined conditions of medium-speed night curves, small curve radii, and excessively close longitudinal distances, thereby reducing the risk of false braking.

[0047] Based on the safety distance requirements corresponding to vehicle type, the current speed range of the vehicle, and the lane curvature characteristics in the nighttime curve environment, the remaining target vehicles to be identified are further judged for proximity and curve adaptability. When it is determined that the longitudinal distance between a vehicle and the vehicle is insufficient, or that the relative position of the vehicle has an abnormally close trend under the curve curvature conditions, the vehicle is identified as the first misidentified vehicle. Thus, before entering the next filtering stage, these unstable or unreliable vehicles caused by excessively close distance, speed mismatch, or curve space limitations are eliminated, ensuring that the vehicles finally used for TJA control decisions only include real and valid targets that meet the curve safety conditions.

[0048] Step S104: Remove the false target and the first misidentified vehicle from the target vehicles to be identified, and perform a second filtering operation on the remaining target vehicles to be identified.

[0049] Specifically, if any judgment condition of the first filtering strategy is not met, it indicates that the ambient light intensity is within the normal identifiable range and the confidence level of the corresponding target is stable and the outline is continuous. The perception system can reliably identify the target in consecutive frames. In this case, the first vehicle to be identified is regarded as a normal target, no filtering operation is performed on it, and it is allowed to enter the next stage of the filtering judgment process. At this time, false targets are removed from the target vehicles to be identified, and the first filtering operation is performed again on the remaining target vehicles to be identified through the second filtering strategy. The second filtering strategy is used to further screen out targets that may be too close in longitudinal distance or have potential collision risks under specific curve conditions based on the relationship between the vehicle type of the preceding vehicle, the vehicle distance, and the vehicle speed and curve geometry features. This identifies abnormally close vehicles that are not suitable as the basis for following judgment in nighttime curve scenarios, and prevents such abnormal targets from continuing to participate in subsequent trajectory prediction and TJA control decisions.

[0050] When the remaining target vehicles do not meet the second filtering conditions, it means that the vehicle has not reached the preset trigger threshold in any of the following aspects: safe distance corresponding to the vehicle type, self-vehicle speed trigger condition, and curve curvature requirement. Its longitudinal approach, safety impact level, and collision risk in the curve environment are all within an acceptable range. At this time, it is considered that the vehicle does not have abnormal approach characteristics, and the vehicle is regarded as a normal target, not as the first misidentified vehicle, and is allowed to continue to the subsequent second filtering operation to further judge its stability through motion fluctuation characteristics. This ensures that only the target vehicles that may cause misjudgment or misbraking are eliminated when driving on curves at medium and high speeds at night. At this time, the first misidentified vehicle is removed from the remaining target vehicles, and the second filtering operation is performed on the remaining target vehicles through the third filtering strategy. The third filtering strategy is used to perform a statistical robustness assessment of the motion state of the remaining targets based on the longitudinal distance fluctuation characteristics and relative speed stability of the vehicle in continuous time sequence. This identifies the second misidentified vehicles that have abnormal fluctuation trends in distance or speed changes and cannot maintain a stable following relationship, further improving the accuracy and reliability of TJA control target selection in night curve scenarios.

[0051] Step S105: Calculate the motion fluctuation feature threshold corresponding to the remaining target vehicles to be identified through the second filtering operation and based on multi-source data, and determine whether the motion fluctuation feature threshold is less than the preset reference threshold.

[0052] In one possible implementation, the motion fluctuation feature threshold includes a longitudinal distance fluctuation threshold and a speed fluctuation standard deviation threshold. Step S107 further includes: acquiring a third vehicle to be identified, wherein the third vehicle to be identified is any vehicle among the remaining target vehicles to be identified; calculating a longitudinal distance fluctuation threshold based on the target longitudinal distance sequence in the full perception data of the preceding vehicle; calculating a speed fluctuation standard deviation threshold based on the target speed sequence in the full perception data of the preceding vehicle; determining whether the motion fluctuation feature threshold is less than a preset distance fluctuation threshold, and determining whether the speed fluctuation standard deviation threshold is less than a preset speed fluctuation threshold.

[0053] Specifically, within the current time window, a target longitudinal distance sequence of length N is constructed for the third vehicle to be identified, denoted as... , ,…, ,in Indicates time The longitudinal distance between the vehicle and the target at each moment. To emphasize the importance of the most recent time data, an exponentially decaying weight is applied to this sequence in the direction of decreasing time. The weighted mean of the vertical distance sequence is calculated using a weighted method. The calculation formula is:

[0054] in, For the corresponding The weights, which can be defined as follows: , This is the weight decay coefficient, a constant with a value greater than 0. The larger the value, the more it emphasizes distance data from the most recent moment; The time window length can be set within several sampling periods based on the system sampling period and TJA control response requirements; for example, the value range can be 5 to 50 sampling points. The weighted average mentioned above is used to characterize the average longitudinal distance of the target over a recent period, thereby eliminating the influence of the fixed bias in subsequent fluctuation calculations.

[0055] Obtaining the weighted mean Then, the weighted standard deviation is calculated based on the target longitudinal distance sequence within the same time window. This is used to characterize the degree of fluctuation in longitudinal distance, and the calculation formula is: in, This represents the weighted standard deviation of the longitudinal distance to the target within the current time window, reflecting the stability of the longitudinal proximity relationship between the target and the vehicle. A larger value indicates that the target's distance fluctuates, and its approach characteristics are unstable. This is combined with the longitudinal acceleration data from the vehicle's motion. and lane curvature in nighttime curve environment data Further construct longitudinal distance fluctuation thresholds related to operating conditions. This allows for a more rigorous amplification and evaluation of distance fluctuations under conditions of rapid acceleration or deceleration, as well as on curves with significant curvature. The calculation formula can be set as follows:

[0056] in, The vertical distance fluctuation threshold, This is the benchmark amplification factor for longitudinal distance fluctuations, used to map the standard deviation scale to the threshold scale; it is a constant with a value greater than 0. This is the acceleration sensitivity coefficient, used to reflect the influence of the vehicle's longitudinal acceleration and deceleration on the sensitivity to distance fluctuations. A constant whose value is greater than 0. The average longitudinal acceleration of the vehicle within the current time window can be obtained by averaging the longitudinal acceleration signals in the CAN bus within the window. This is a nominal acceleration reference value used to normalize the effects of different accelerations. Its value can be set based on the actual acceleration performance of the vehicle. This is the curvature sensitivity coefficient, used to characterize how much the vehicle is affected by centrifugal force when the curvature of the curve is greater, leading to distance fluctuations that require more stringent consideration. A constant whose value is greater than 0. The current lane curvature can be obtained by matching the lane line recognition algorithm or a high-precision map with the vehicle's current location. The reference curvature value is used to normalize the impact of different curve radii. Through the combination of the above coefficients, the standard deviation of the longitudinal distance is calculated. Under the same conditions, when the vehicle exhibits significant longitudinal acceleration and deceleration behavior, that is... When the value is large, or when the current location is in a curve region with significant curvature, that is... When the value is large, because and All exhibit an amplification factor greater than 1, making Obtaining larger values ​​under these conditions improves the sensitivity of motion fluctuation characteristic thresholds in rapid acceleration, sudden braking, and high-curvature road sections, enabling more rigorous screening of distance fluctuations of target vehicles in high-risk environments.

[0057] In obtaining and After that, Considered as a threshold for motion fluctuation characteristics, it is compared with a preset distance fluctuation threshold. When comparing, When the distance fluctuation is less than a preset threshold, the vertical distance change of the target within the current window is considered to be within a stable and acceptable range; similarly, With preset speed fluctuation threshold When comparing, If the velocity fluctuation is less than the preset velocity fluctuation threshold, it is considered that the target does not have obvious abnormal fluctuations in the velocity direction. and It can be determined offline during the calibration phase according to different target types and operating conditions, and is used to limit the maximum acceptable range of longitudinal distance and speed fluctuations. and If the corresponding preset threshold requirements are met, the third vehicle to be identified is regarded as a target with a stable motion state in the third filtering strategy; if any one of them exceeds the preset threshold, the target is marked as a second misidentified vehicle and removed.

[0058] In practical implementation, based on the above continuous fluctuation characteristic calculation, a discrete rule threshold can be introduced for supplementary determination of different vehicle types, combining continuous statistical results with preset empirical thresholds. For example, when the vehicle speed is 60 kph and the target type is a sedan, a longitudinal distance fluctuation threshold can be introduced for longitudinal distance, and this threshold can be set as follows:

[0059] When the vehicle speed is 50 kph and the target is a motorcycle, the longitudinal distance fluctuation threshold can be expressed as: When the vehicle speed is 50 kph and the target is a truck, the longitudinal distance fluctuation threshold can also be expressed as: This reflects the characteristics of trucks, such as their larger size and higher requirements for distance safety margins. The standard deviation threshold for speed fluctuations can be set as follows: Based on this, type-related lower limits are set. For example, in a 60kph scenario where the target type is a sedan, the target speed fluctuation standard deviation threshold can be greater than or equal to 60. In a 50kph scenario, if the target type is a motorcycle, the target speed fluctuation standard deviation threshold can be greater than or equal to 50, and if the target type is a truck, the target speed fluctuation standard deviation threshold can be greater than or equal to 60. By using the speed fluctuation standard deviation thresholds corresponding to different target types, the differences in speed fluctuation behavior between sedans, motorcycles, and trucks can be further distinguished. This allows the third filtering strategy to more precisely identify target vehicles with unstable motion states in nighttime curve scenarios under the joint constraints of continuous statistical features and discrete rule features, thereby ensuring that the targets ultimately retained are more suitable as effective control targets for TJA.

[0060] Step S106: If it is confirmed that the motion fluctuation feature threshold is less than the preset reference threshold, then the remaining target vehicle to be identified is determined to be a valid TJA control target, and TJA braking operation is performed based on the valid TJA control target.

[0061] Specifically, if the motion fluctuation feature threshold is confirmed to be less than a preset reference threshold, the remaining target vehicle to be identified is determined to be a valid TJA control target, and TJA braking operation is performed based on the valid TJA control target. Specifically, after jointly judging the longitudinal distance fluctuation feature and speed fluctuation feature of the third target vehicle, when the longitudinal distance fluctuation threshold corresponding to that vehicle... The distance fluctuation is less than the preset distance fluctuation threshold, and its speed fluctuation standard deviation threshold is also less than the preset distance fluctuation threshold. When the speed fluctuation is less than the preset threshold, the vehicle can be considered to have maintained a stable distance change trajectory and a smooth speed change trend in nighttime curve scenarios. The vehicle has high reliability in the time-series dimension and can continuously maintain clear forward motion characteristics within the range of the vehicle's future predicted trajectory. In this case, the vehicle is identified as an effective TJA control target and the effective target is input into the TJA control link. Through the combined action of the vehicle speed, vehicle trajectory prediction, and path funnel constraints, a longitudinal control quantity suitable for curve conditions is generated to perform smooth deceleration, following, or braking actions, thereby maintaining the safety and comfort of nighttime curve cruise without causing false braking.

[0062] If the threshold of motion fluctuation characteristics is confirmed Greater than or equal to the preset distance fluctuation threshold, and confirming the speed fluctuation standard deviation threshold. If the speed fluctuation is greater than or equal to the preset speed fluctuation threshold, it indicates that the third vehicle to be identified has a large fluctuation in the distance change trend, or its speed sequence shows obvious jump characteristics. The vehicle cannot form a stable and predictable motion state in front of the vehicle, and may trigger unnecessary deceleration actions in nighttime curve scenarios. In this case, the vehicle is regarded as the second misidentified vehicle, and the second misidentified vehicle is removed from the remaining target vehicles to be identified, so that the subsequent TJA control only performs braking operations based on real and stable front vehicle targets, further reducing the probability of false braking in situations with large curvature changes, insufficient lighting, or multiple vehicle interference.

[0063] Please refer to Figure 2 The diagram illustrates a module schematic of a TJA control misjudgment reduction device for nighttime curve cruise provided in an embodiment of this application. The device includes an acquisition module 21 and a processing module 22, wherein... The acquisition module 21 is used to acquire multi-source data corresponding to the target vehicle to be identified, the target vehicle itself, and the nighttime curve environment.

[0064] The processing module 22 is used to input multi-source data into a multi-filtering strategy model and perform a first filtering operation through the multi-filtering strategy model; identify false targets and first misidentified vehicles among the target vehicles to be identified through the first filtering operation; remove false targets and first misidentified vehicles among the target vehicles to be identified, and perform a second filtering operation on the remaining target vehicles to be identified; calculate the motion fluctuation feature threshold corresponding to the remaining target vehicles to be identified through the second filtering operation and based on the multi-source data, and determine whether the motion fluctuation feature threshold is less than a preset reference threshold; if it is confirmed that the motion fluctuation feature threshold is less than the preset reference threshold, then determine that the remaining target vehicles to be identified are valid TJA control targets, and perform TJA braking operation based on valid TJA control targets.

[0065] In one possible implementation, the acquisition module 21 is used to acquire multi-source data corresponding to the target vehicle to be identified, the target vehicle, and the nighttime curve environment. Specifically, it includes: acquiring the full perception data of the preceding vehicle corresponding to the target vehicle to be identified, the vehicle motion data corresponding to the target vehicle, and the nighttime curve environment data corresponding to the nighttime curve environment; and using the set of the full perception data of the preceding vehicle, the vehicle motion data, and the nighttime curve environment data as multi-source data.

[0066] In one possible implementation, after acquiring multi-source data corresponding to the target vehicle to be identified, the target vehicle, and the nighttime curve environment, the acquisition module 21 further includes: predicting the vehicle trajectory based on the vehicle motion data and constructing a path funnel on the vehicle trajectory; filtering the target vehicle to be identified on the path funnel and selecting the TJA control target.

[0067] In one possible implementation, the processing module 22 is used to identify false targets among the target vehicles to be identified through a first filtering operation, specifically including: acquiring a first vehicle to be identified, wherein the first vehicle to be identified is any vehicle among the target vehicles to be identified; determining whether the on-site light intensity value is lower than a preset light intensity value based on the light intensity data in the nighttime curve environment data; determining whether the confidence value corresponding to the target vehicle to be identified is less than a preset confidence value based on the point cloud image data in the full perception data of the preceding vehicle; and identifying the first vehicle to be identified whose on-site light intensity value is lower than the preset light intensity value or whose confidence value is less than the preset confidence value as a false target.

[0068] In one possible implementation, the processing module 22 is used to identify the first misidentified vehicle among the remaining target vehicles to be identified through a first filtering operation. Specifically, this includes: acquiring a second vehicle to be identified, wherein the second vehicle to be identified is any vehicle among the remaining target vehicles to be identified; determining whether the distance of the second vehicle to be identified is less than or equal to a preset distance corresponding to the vehicle type based on vehicle type data and vehicle distance data in the full perception data of the preceding vehicle; determining whether the driving speed of the target vehicle is greater than or equal to a preset driving speed based on the driving speed data of the vehicle in the vehicle motion data; determining whether the lane curvature is within a preset curvature range based on the lane curvature data in the nighttime curve environment data; and if it is confirmed that the vehicle distance is less than or equal to the preset distance corresponding to the vehicle type, or that the driving speed is greater than or equal to the preset driving speed, or that the lane curvature is within a preset curvature range, then the second vehicle to be identified is identified as the first misidentified vehicle.

[0069] In one possible implementation, the motion fluctuation feature threshold includes a longitudinal distance fluctuation threshold and a speed fluctuation standard deviation threshold. The processing module 22 is used to calculate the motion fluctuation feature threshold corresponding to the remaining target vehicles to be identified through a second filtering operation and based on multi-source data, and to determine whether the motion fluctuation feature threshold is less than a preset reference threshold. Specifically, this includes: acquiring a third vehicle to be identified, wherein the third vehicle to be identified is any vehicle among the remaining target vehicles to be identified; calculating the longitudinal distance fluctuation threshold based on the target longitudinal distance sequence in the full perception data of the preceding vehicle; calculating the speed fluctuation standard deviation threshold based on the target speed sequence in the full perception data of the preceding vehicle; determining whether the motion fluctuation feature threshold is less than a preset distance fluctuation threshold, and determining whether the speed fluctuation standard deviation threshold is less than a preset speed fluctuation threshold.

[0070] In one possible implementation, after determining whether the motion fluctuation feature threshold is less than a preset distance fluctuation threshold and whether the speed fluctuation standard deviation threshold is less than a preset speed fluctuation threshold, the processing module 22 further includes: if it is confirmed that the motion fluctuation feature threshold is greater than or equal to the preset distance fluctuation threshold and the speed fluctuation standard deviation threshold is greater than or equal to the preset speed fluctuation threshold, then the third vehicle to be identified is identified as the second misidentified vehicle; and the second misidentified vehicle is removed from the remaining target vehicles to be identified.

[0071] It should be noted that the above embodiments of the apparatus are only illustrated by the division of the above functional modules. In practical applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the apparatus and method embodiments provided in the above embodiments belong to the same concept, and the specific implementation process can be found in the method embodiments, which will not be repeated here.

[0072] This application also provides an electronic device. (See reference...) Figure 3 , Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include: at least one processor 301, at least one communication bus 302, a user interface 303, at least one network interface 304, and a memory 305.

[0073] The communication bus 302 is used to enable communication between these components.

[0074] The user interface 303 may include a display screen and a camera. Optionally, the user interface 303 may also include a standard wired interface and a wireless interface.

[0075] The network interface 304 may optionally include a standard wired interface or a wireless interface (such as a Wi-Fi interface).

[0076] The processor 301 may include one or more processing cores. The processor 301 connects to various parts of the server using various interfaces and lines, and performs various server functions and processes data by running or executing instructions, programs, code sets, or instruction sets stored in memory 305, and by calling data stored in memory 305. Optionally, the processor 301 may be implemented using at least one hardware form of Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), or Programmable Logic Array (PLA). The processor 301 may integrate one or a combination of several of the following: Central Processing Unit (CPU), Graphics Processing Unit (GPU), and modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing the content required for display; and the modem handles wireless communication. It is understood that the modem may also not be integrated into the processor 301 and may be implemented as a separate chip.

[0077] The memory 305 may include random access memory (RAM) or read-only memory. Optionally, the memory 305 may include a non-transitory computer-readable storage medium. The memory 305 may be used to store instructions, programs, code, code sets, or instruction sets. The memory 305 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as touch function, sound playback function, image playback function, etc.), instructions for implementing the above-described method embodiments, etc.; the data storage area may store data involved in the above-described method embodiments, etc. Optionally, the memory 305 may also be at least one storage device located remotely from the aforementioned processor 301. (Refer to...) Figure 3 The memory 305, which serves as a computer storage medium, may include an operating system, a network communication module, a user interface module, and a TJA control misjudgment reduction application for nighttime curve cruise.

[0078] exist Figure 3 In the illustrated electronic device, the user interface 303 is primarily used to provide an input interface for the user and acquire user input data; while the processor 301 can be used to call the TJA control misjudgment reduction application stored in the memory 305 for nighttime curve cruise. When executed by one or more processors 301, the electronic device performs one or more of the methods described in the above embodiments. It should be noted that, for the foregoing method embodiments, for the sake of simplicity, they are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, because according to this application, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also understand that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0079] This application also provides a computer-readable storage medium storing instructions. When executed by one or more processors, these instructions cause an electronic device to perform one or more of the methods described in the above embodiments.

[0080] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0081] In the various embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some service interface; the indirect coupling or communication connection between apparatuses or units may be electrical or other forms.

[0082] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0083] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0084] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, portable hard drives, magnetic disks, or optical disks.

[0085] The above description is merely an exemplary embodiment disclosed in this application and should not be construed as limiting the scope of this application. Any equivalent changes and modifications made in accordance with the teachings of this application shall still fall within the scope of this application.

[0086] This application is intended to cover any variations, uses, or adaptations disclosed herein that follow the general principles disclosed herein and include common knowledge or customary technical means in the art that are not described in this application.

Claims

1. A method for reducing TJA control misjudgment for night-time curved cruising, characterized in that, The method comprises: obtaining multi-source data corresponding to a target vehicle to be identified, a target ego vehicle and a night-time curved road environment; inputting the multi-source data into a multi-filtering strategy model and performing a first filtering operation through the multi-filtering strategy model; identifying false targets and first misidentified vehicles in the target vehicle to be identified through the first filtering operation; eliminating the false targets and the first misidentified vehicles in the target vehicle to be identified, and performing a second filtering operation on the remaining target vehicle to be identified; calculating a motion fluctuation feature threshold corresponding to the remaining target vehicle to be identified through the second filtering operation and based on the multi-source data, and determining whether the motion fluctuation feature threshold is less than a preset reference threshold; if it is confirmed that the motion fluctuation feature threshold is less than the preset reference threshold, determining that the remaining target vehicle to be identified is an effective TJA control target, and performing a TJA braking operation based on the effective TJA control target.

2. The method of claim 1, wherein, The obtaining of the multi-source data corresponding to the target vehicle to be identified, the target ego vehicle and the night-time curved road environment specifically comprises: obtaining front vehicle full perception data corresponding to the target vehicle to be identified, ego vehicle motion data corresponding to the target ego vehicle and night-time curved road environment data corresponding to the night-time curved road environment; collecting the front vehicle full perception data, the ego vehicle motion data and the night-time curved road environment data as the multi-source data.

3. The method of claim 2, wherein, After the obtaining of the multi-source data corresponding to the target vehicle to be identified, the target ego vehicle and the night-time curved road environment, the method further comprises: predicting an ego vehicle trajectory based on the ego vehicle motion data, and constructing a path funnel on the ego vehicle trajectory; filtering the target vehicle to be identified on the path funnel and selecting a TJA control target.

4. The method of claim 2, wherein, The identification of the false targets in the target vehicle to be identified through the first filtering operation specifically comprises: obtaining a first vehicle to be identified, the first vehicle to be identified being any vehicle in the target vehicle to be identified; determining whether an on-site illumination intensity value is lower than a preset illumination intensity value based on illumination intensity data in the night-time curved road environment data; determining whether a confidence value corresponding to the target vehicle to be identified is less than a preset confidence value based on point cloud image data in the front vehicle full perception data; regarding the first vehicle to be identified whose on-site illumination intensity value is lower than the preset illumination intensity value or whose confidence value is less than the preset confidence value as the false target.

5. The method of claim 2, wherein, The identification of the first misidentified vehicles in the target vehicle to be identified through the first filtering operation specifically comprises: obtaining a second vehicle to be identified, the second vehicle to be identified being any vehicle in the target vehicle to be identified after the false targets are eliminated; determining whether a vehicle distance of the second vehicle to be identified is less than or equal to a preset distance corresponding to a vehicle type based on vehicle type data and vehicle distance data in the front vehicle full perception data; determining whether a driving speed of the target ego vehicle is greater than or equal to a preset ego vehicle speed based on ego vehicle speed data in the ego vehicle motion data; determine whether the lane line curvature is in a preset curve curvature interval based on lane line curvature data in the night curve environment data; if it is determined that the vehicle distance is less than or equal to the preset distance corresponding to the vehicle type, or that the driving speed is greater than or equal to the preset ego vehicle speed, or that the lane line curvature is in the preset curve curvature interval, the second to-be-identified vehicle is determined as the first misidentified vehicle.

6. The method of claim 2, wherein, The motion fluctuation feature threshold value includes a longitudinal distance fluctuation threshold value and a speed fluctuation standard deviation threshold value. The motion fluctuation feature threshold value corresponding to the remaining target to-be-identified vehicle is calculated based on the multi-source data through the second filtering operation, and it is determined whether the motion fluctuation feature threshold value is less than a preset reference threshold value. Specifically, the method comprises: obtaining a third to-be-identified vehicle, the third to-be-identified vehicle being any vehicle in the remaining target to-be-identified vehicle; calculating the longitudinal distance fluctuation threshold value based on a target longitudinal distance sequence in the front vehicle full perception data; calculating the speed fluctuation standard deviation threshold value based on a target speed sequence in the front vehicle full perception data; determining whether the motion fluctuation feature threshold value is less than a preset distance fluctuation threshold value, and determining whether the speed fluctuation standard deviation threshold value is less than a preset speed fluctuation threshold value.

7. The method of claim 6, wherein, After determining whether the motion fluctuation feature threshold value is less than a preset distance fluctuation threshold value, and determining whether the speed fluctuation standard deviation threshold value is less than a preset speed fluctuation threshold value, the method further comprises: if it is determined that the motion fluctuation feature threshold value is greater than or equal to the preset distance fluctuation threshold value, and that the speed fluctuation standard deviation threshold value is greater than or equal to the preset speed fluctuation threshold value, the third to-be-identified vehicle is determined as a second misidentified vehicle; eliminating the second misidentified vehicle from the remaining target to-be-identified vehicle.

8. A TJA control misjudgment reduction device for night-time curved cruising, characterized by, The device comprises an obtaining module and a processing module, wherein, the obtaining module is configured to obtain multi-source data corresponding to a target to-be-identified vehicle, a target ego vehicle, and a night curve environment; the processing module is configured to input the multi-source data into a multi-filtering strategy model, and perform a first filtering operation through the multi-filtering strategy model; identify false targets and first misidentified vehicles in the target to-be-identified vehicle through the first filtering operation; eliminate the false targets and the first misidentified vehicles in the target to-be-identified vehicle, and perform a second filtering operation on the remaining target to-be-identified vehicle; calculate a motion fluctuation feature threshold value corresponding to the remaining target to-be-identified vehicle based on the multi-source data through the second filtering operation, and determine whether the motion fluctuation feature threshold value is less than a preset reference threshold value; if it is determined that the motion fluctuation feature threshold value is less than the preset reference threshold value, the remaining target to-be-identified vehicle is determined as an effective TJA control target, and a TJA braking operation is performed based on the effective TJA control target.

9. An electronic device, comprising: The electronic device comprises a processor, a communication bus, a user interface, a network interface, and a memory for storing instructions, the user interface and the network interface being configured to communicate with other devices, and the processor being configured to execute the instructions stored in the memory to cause the electronic device to perform the method of any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores instructions that, when executed, perform the method of any one of claims 1 to 7.

Images