Vehicle rearview mirror adjusting method, system and equipment based on multi-source information fusion

Abstract

The invention discloses a vehicle rearview mirror adjusting method, system and device based on multi-source information fusion. The method comprises the steps that road scene sensing information, driver state sensing information, vehicle state sensing information and roadside collaborative sensing information are obtained; performing weight distribution and data fusion on the road scene sensing information, the driver state sensing information, the vehicle state sensing information and the roadside collaborative sensing information to obtain target fusion data; inputting the target fusion data into a pre-trained scene stage recognition model to obtain a current driving scene and a corresponding current scene stage; determining a corresponding target rearview mirror adjusting strategy according to the current driving scene and the current scene stage, determining a target rearview mirror angle according to the road scene sensing information, the driver state sensing information and the target rearview mirror adjusting strategy, and adjusting the rearview mirror according to the target rearview mirror angle. The method improves the accuracy and timeliness of vehicle rearview mirror adjustment, and can be applied to the technical field of vehicle control.

Description

Technical Field

[0001] This invention relates to the field of vehicle control technology, and in particular to a method, system and device for adjusting vehicle rearview mirrors based on multi-source information fusion. Background Technology

[0002] Existing vehicle rearview mirror adjustment solutions have the following drawbacks: 1) Delayed triggering mechanism and poor scenario adaptability, failing to avoid blind spots in advance: Existing automatic rearview mirror adjustment relies heavily on single, immediate signals such as "turn signal on" and "shift to R," without incorporating scenario prediction logic for "same-direction angled roads." For example, when a vehicle approaches a green-covered area on a same-direction angled road (such as the angle between a main road and a parallel auxiliary road, or the angle between lanes turning in the same direction), the system only initiates adjustment after the driver activates the turn signal. However, by this time, the vehicle is less than 30 meters from the angle point, and the green-covered blind spot has already formed, resulting in an adjustment action that lags behind the required field of vision. Furthermore, it does not integrate the "same-direction angled blind spot coordinates" from high-precision maps or pre-warning signals from roadside equipment, failing to activate adjustment preparation 50-100 meters before the vehicle enters the angle point. Consequently, when the driver enters the blind spot, the rearview mirror remains at the normal angle for straight road sections, making it impossible to observe road targets on the angled side in time.

[0003] 2) Weak anti-obstruction perception capability and failure to detect targets in blind spots: Existing systems mostly rely on ordinary vehicle-mounted cameras or low-frequency millimeter-wave radar. When faced with obstructions such as greenery and guardrails between roads in the same direction at an angle, their perception capability is significantly insufficient. Ordinary cameras are easily completely obstructed by greenery and guardrails, making it impossible to capture the dynamics of vehicles and pedestrians on the angled side of the road. Although conventional 24GHz millimeter-wave radar can penetrate some greenery, it is affected by clutter from branches and leaves and metal reflections from guardrails, misidentifying stationary obstructions as obstacles, or failing to distinguish between stationary greenery and moving targets on the angled side (such as electric vehicles on the auxiliary road or pedestrians in the separate lane), resulting in the system being unable to obtain key information such as the distance, speed, and orientation of the real targets in the blind spot. For example, when a bicycle approaches on the angled side of the road in the same direction, the existing perception system cannot identify it due to obstruction, and the rearview mirror will not adjust accordingly, directly causing a blind spot.

[0004] 3) Fixed adjustment strategies, failing to achieve dynamic adaptation through "phased + target priority": Existing automatic adjustments mostly adopt "fixed angle presets" (e.g., tilting down 15° when reversing) or "single parameter association" (e.g., fixed wide-range angle when vehicle speed > 40km / h), without designing layered strategies for dynamic scenarios of roads with angles in the same direction. At long distances (80-50m from the angle point), the angle is not prioritized to avoid obstruction on the angle side; at medium distances (50-20m), the focus is not directed based on the target position in the blind spot on the angle side; at close distances (<20m), the angle is not fine-tuned to meet the needs of entering the lane on the angle side. Furthermore, target priority is not distinguished, such as treating stationary greenery on the angle side and non-motorized vehicles approaching at high speeds in the same way, resulting in rearview mirrors either having a wide enough angle to distract the field of vision or a narrow enough angle to miss key targets, failing to meet the scenario-based field of vision needs for different angles (e.g., 30°, 60°).

[0005] 4) Lack of vehicle-road coordination and closed-loop adjustment feedback, resulting in insufficient reliability: Existing systems mostly follow a one-way process of "vehicle-only perception-adjustment," failing to coordinate with roadside equipment (such as RSUs and roadside cameras) on the same-direction, angled roads. When vehicle-mounted perception is completely obscured by dense greenery or high barriers between the same-direction, angled roads, the system cannot obtain the blind-spot overhead image and target data transmitted by the roadside equipment, directly falling into an adjustment stagnation state of "no perception to rely on." Furthermore, there is no mechanism to verify the effect after adjustment: the system does not track the driver's gaze, nor does it detect the target's display proportion in the rearview mirror. If the initial adjustment does not allow the driver to see the target in the blind spot on the angled side, the system will not correct the angle a second time. Moreover, the reset logic is simple (e.g., relying solely on vehicle speed), forcibly restoring the normal angle before confirming the vehicle has fully entered the same-direction, angled lane, easily leading to sudden changes in field of vision or premature reset, affecting subsequent driving safety on angled roads.

[0006] The above problems urgently need to be addressed. Summary of the Invention

[0007] The purpose of this invention is to at least partially solve one of the technical problems existing in the prior art.

[0008] Therefore, one objective of this invention is to provide a vehicle rearview mirror adjustment method based on multi-source information fusion, which improves the accuracy and timeliness of vehicle rearview mirror adjustment, thereby enhancing vehicle driving safety and the user's driving experience.

[0009] Another objective of this invention is to provide a vehicle rearview mirror adjustment system based on multi-source information fusion.

[0010] To achieve the above-mentioned technical objectives, the technical solutions adopted in the embodiments of the present invention include: On one hand, embodiments of the present invention provide a method for adjusting vehicle rearview mirrors based on multi-source information fusion, including the following steps: Acquire road scene perception information, driver status perception information, vehicle status perception information, and roadside cooperative perception information; Weights are assigned and data fusion is performed on the road scene perception information, the driver state perception information, the vehicle state perception information, and the roadside cooperative perception information to obtain target fused data; The target fusion data is input into a pre-trained scene stage recognition model to obtain the current driving scene and the corresponding current scene stage. The target rearview mirror adjustment strategy is determined based on the current driving scenario and the current scenario stage. The target rearview mirror angle is determined based on the road scene perception information, the driver state perception information, and the target rearview mirror adjustment strategy. Then, the rearview mirror is adjusted based on the target rearview mirror angle.

[0011] Furthermore, in one embodiment of the present invention, the acquisition of road scene perception information, driver state perception information, vehicle state perception information, and roadside cooperative perception information specifically includes: Based on road images, radar detection information, and high-precision maps, road types, occlusion conditions, blind spot target types, and blind spot target locations are identified to obtain the road scene perception information; Based on seat sensor data, driver images, and driver operation information, the driver's physiological state and driver's operation intentions are identified to obtain the driver state perception information. The vehicle status perception information is obtained by acquiring real-time vehicle status and vehicle driving mode through the body controller; The system receives regional global images and regional vehicle positioning data sent by roadside equipment to obtain the roadside collaborative perception information.

[0012] Furthermore, in one embodiment of the present invention, the weighting and data fusion of the road scene perception information, the driver state perception information, the vehicle state perception information, and the roadside cooperative perception information to obtain target fused data specifically includes: Determine the weight parameters for the road scene perception information, the driver state perception information, the vehicle state perception information, and the roadside cooperative perception information; The road scene perception information, driver state perception information, vehicle state perception information, and roadside cooperative perception information are cleaned and time-aligned, and then time-series fusion is performed based on the weight parameters to obtain the target fused data.

[0013] Furthermore, in one embodiment of the present invention, the scene stage recognition model is trained through the following steps: Obtain road scene perception samples, driver state perception samples, vehicle state perception samples, and roadside cooperative perception samples in the test scenario, and generate sample fusion data based on the road scene perception samples, driver state perception samples, vehicle state perception samples, and roadside cooperative perception samples. The primary labels for driving scenarios and the secondary labels for driving stages corresponding to the sample fusion data are determined by manual annotation. A training dataset is then generated based on the sample fusion data and the corresponding primary labels for driving scenarios and secondary labels for driving stages. The training samples are input into a pre-built convolutional neural network to obtain the predicted scene type and the predicted stage type; The loss value is determined based on the predicted scenario type, the predicted stage type, the primary label of the driving scenario, and the secondary label of the driving scenario. The parameters of the convolutional neural network are updated based on the loss value to obtain the trained scene stage recognition model; The secondary stage label of the scenario is a stage subdivision label under the primary stage label of the driving scenario.

[0014] Furthermore, in one embodiment of the present invention, the step of determining the corresponding target rearview mirror adjustment strategy based on the current driving scenario and the current scenario stage, and determining the target rearview mirror angle based on the road scene perception information, the driver state perception information, and the target rearview mirror adjustment strategy, specifically includes: The target rearview mirror adjustment strategy is obtained by matching the current driving scenario and the current scenario stage from a pre-defined rearview mirror adjustment strategy library. The blind spot target type and location are determined based on the road scene perception information, and the blind spot target risk level is determined based on the blind spot target type and location. The driver's posture is determined based on the driver's state perception information; Based on the target risk level in the blind spot and the driver's posture, the target rearview mirror adjustment strategy is queried to obtain the target rearview mirror angle.

[0015] Furthermore, in one embodiment of the present invention, the adjustment of the rearview mirror according to the target rearview mirror angle specifically includes: Obtain the current rearview mirror angle, and determine the rearview mirror angle adjustment amount based on the current rearview mirror angle and the target rearview mirror angle; The horizontal and vertical adjustment angles are determined based on the rearview mirror angle adjustment amount, and the angle adjustment rate is also determined. Horizontal motor control commands and vertical motor control commands are generated based on the horizontal adjustment angle, the vertical adjustment angle, and the angle adjustment rate. The horizontal and vertical adjustment motors are driven to adjust the target rearview mirror according to the horizontal motor control command and the vertical motor control command.

[0016] On the other hand, embodiments of the present invention provide a vehicle rearview mirror adjustment system based on multi-source information fusion, comprising: The information acquisition module is used to acquire road scene perception information, driver status perception information, vehicle status perception information, and roadside cooperative perception information. The data fusion module is used to perform weight allocation and data fusion on the road scene perception information, the driver state perception information, the vehicle state perception information and the roadside cooperative perception information to obtain target fused data; The scene stage recognition module is used to input the target fusion data into the pre-trained scene stage recognition model to obtain the current driving scene and the corresponding current scene stage; The adjustment module is used to determine the corresponding target rearview mirror adjustment strategy based on the current driving scenario and the current scenario stage, determine the target rearview mirror angle based on the road scene perception information, the driver state perception information and the target rearview mirror adjustment strategy, and then adjust the rearview mirror according to the target rearview mirror angle.

[0017] On the other hand, embodiments of the present invention provide an electronic device, which includes a memory, a processor, a program stored in the memory and executable on the processor, and a data bus for enabling communication between the processor and the memory. When the program is executed by the processor, it implements the vehicle rearview mirror adjustment method based on multi-source information fusion as described above.

[0018] On the other hand, embodiments of the present invention also provide a vehicle, the vehicle including a vehicle rearview mirror adjustment system or electronic device based on multi-source information fusion as described above.

[0019] On the other hand, embodiments of the present invention also provide a storage medium, which is a computer-readable storage medium for computer-readable storage. The storage medium stores one or more programs, which can be executed by one or more processors to implement the vehicle rearview mirror adjustment method based on multi-source information fusion as described above.

[0020] On the other hand, embodiments of the present invention also provide a computer program product, including a computer program that, when executed by a processor, implements the vehicle rearview mirror adjustment method based on multi-source information fusion as described above.

[0021] The advantages and beneficial effects of the present invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention: This invention acquires road scene perception information, driver state perception information, vehicle state perception information, and roadside cooperative perception information. It then performs weight allocation and data fusion on these information to obtain target fused data. This target fused data is input into a pre-trained scene stage recognition model to obtain the current driving scene and its corresponding stage. Based on the current driving scene and stage, a corresponding target rearview mirror adjustment strategy is determined. The target rearview mirror angle is then determined based on the road scene perception information, driver state perception information, and the target rearview mirror adjustment strategy. Finally, the rearview mirror is adjusted according to the target rearview mirror angle. This invention identifies the current driving scene and stage based on road scene perception information, driver state perception information, vehicle state perception information, and roadside cooperative perception information, then matches the corresponding rearview mirror adjustment strategy, and finally determines the corresponding target rearview mirror angle by combining the road scene perception information and driver state perception information. This improves the accuracy and timeliness of vehicle rearview mirror adjustment, thereby enhancing vehicle driving safety and the user's driving experience. Attached Figure Description

[0022] To more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings used in the embodiments of the present invention are described below. It should be understood that the drawings described below are only for the convenience of clearly describing some embodiments of the technical solutions of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 A flowchart illustrating the steps of a vehicle rearview mirror adjustment method based on multi-source information fusion provided in an embodiment of the present invention; Figure 2 A schematic diagram of an embodiment of the present invention is provided; Figure 3 A schematic diagram of the structure of a vehicle rearview mirror adjustment system based on multi-source information fusion provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0024] The embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application. It should be noted that although functional modules are divided in the system schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the system schematic diagram or the order in the flowchart. The step numbers in the following embodiments are only set for ease of explanation and do not limit the order between steps. The execution order of each step in the embodiments can be adaptively adjusted according to the understanding of those skilled in the art.

[0025] In the description of this invention, "multiple" means two or more. The use of "first" and "second" is for distinguishing technical features only and should not be construed as indicating or implying relative importance, the number of indicated technical features, or the order of the indicated technical features. Furthermore, unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0026] It should be noted that in all specific embodiments of this application, when processing data related to user identity or characteristics, such as user information, user behavior data, user historical data, and user location information, user permission or consent is obtained first. Furthermore, the collection, use, and processing of this data comply with relevant laws, regulations, and standards of the relevant countries and regions. In addition, when embodiments of this application require access to sensitive personal information of users, separate permission or consent from the user is obtained through pop-ups or redirects to confirmation pages. Only after obtaining the user's separate permission or consent is the necessary user-related data for the proper functioning of the embodiments of this application obtained.

[0027] This invention aims to solve the following core problems existing in current automatic rearview mirror adjustment technology: 1) This invention addresses the problem of delayed adjustment caused by existing systems that do not incorporate scenario prediction of road angles in the same direction and rely solely on a single real-time signal. Existing technologies only initiate adjustment when the vehicle is less than 30m from the angle point. This invention integrates high-precision map blind spot coordinates with roadside pre-warnings, activating adjustment preparation 50-100m in advance to avoid adjustment delays after the blind spot has formed.

[0028] 2) This invention addresses the problem of existing sensing devices having weak anti-obstruction capabilities and being unable to obtain target information in blind spots due to interference from greenery and guardrails. By combining anti-interference millimeter-wave radar with radar-vision fusion technology, this invention penetrates obstructions and filters clutter, accurately acquiring the distance, speed, and azimuth of targets on the included side, thus avoiding detection failure.

[0029] 3) This invention addresses the problem of existing adjustment strategies being rigid, lacking phased approaches and target priorities, and failing to adapt to different angle scenarios. It adjusts in stages based on distance from the angle point (far, medium, and near), and prioritizes key targets according to target risk, meeting the field-of-view requirements of different angles such as 30° and 60°.

[0030] 4) This invention addresses the problems of existing systems lacking vehicle-road cooperation and feedback loops, resulting in perception failures and improper resets. It integrates roadside equipment to supplement perception, verifies the adjustment effect through driver eye tracking and target display ratio, and only resets the system after confirming the vehicle has fully entered the included lane, thus avoiding sudden changes in field of vision.

[0031] like Figure 1 The diagram shows a flowchart of a vehicle rearview mirror adjustment method based on multi-source information fusion provided in an embodiment of the present invention. (Refer to...) Figure 1 This invention provides a method for adjusting vehicle rearview mirrors based on multi-source information fusion, specifically including the following steps: S101. Acquire road scene perception information, driver status perception information, vehicle status perception information, and roadside cooperative perception information. S102. Weighting and data fusion are performed on road scene perception information, driver state perception information, vehicle state perception information and roadside cooperative perception information to obtain target fused data. S103. Input the target fusion data into the pre-trained scene stage recognition model to obtain the current driving scene and the corresponding current scene stage; S104. Determine the corresponding target rearview mirror adjustment strategy based on the current driving scenario and the current scenario stage. Determine the target rearview mirror angle based on the road scenario perception information, driver state perception information and the target rearview mirror adjustment strategy, and then adjust the rearview mirror according to the target rearview mirror angle.

[0032] Specifically, the embodiments of the present invention are based on multi-source information deep fusion technology. Through collaborative perception and accurate identification of typical complex road scenarios such as main road turning into auxiliary road and roads with angles in the same direction, the invention simultaneously collects key factors in multiple dimensions such as driver physiological characteristics (such as posture and line of sight) and driving state (such as operating intention), vehicle real-time dynamics (such as vehicle speed and orientation) and driving mode, and constructs a scenario-human-vehicle collaborative decision-making model. This enables dynamic adaptive adjustment of the vehicle's rearview mirror, effectively eliminating blind spots caused by obstruction, reducing interactive driving risks, and significantly improving driving safety and ease of operation.

[0033] It can be recognized that the embodiments of the present invention identify the current driving scene and the current scene stage based on road scene perception information, driver state perception information, vehicle state perception information and roadside cooperative perception information, then match the corresponding rearview mirror adjustment strategy, and then combine the road scene perception information and driver state perception information to determine the corresponding target rearview mirror angle, thereby improving the accuracy and timeliness of vehicle rearview mirror adjustment, thereby improving vehicle driving safety and user driving experience.

[0034] This invention aims to "eliminate blind spots caused by road obstructions and improve driving safety." It constructs a four-layer vertical collaborative architecture of "perception-fusion decision-making-execution-feedback," achieving scenario-based and personalized automatic adaptation of the rearview mirror through multi-dimensional information collection, intelligent decision-making, and dynamic adjustment. The details are explained below.

[0035] As an optional implementation, road scene perception information, driver state perception information, vehicle state perception information, and roadside cooperative perception information are acquired, specifically including: S201. Based on road images, radar detection information, and high-precision maps, identify road types, occlusion conditions, blind spot target types, and blind spot target locations to obtain road scene perception information; S202. Based on seat sensor data, driver image and driver operation information, identify the driver's physiological state and driver's operation intention to obtain driver state perception information; S203. Obtain real-time vehicle status and vehicle driving mode through the body controller to obtain vehicle status perception information; S204. Receive the regional global image and regional vehicle positioning data sent by the roadside equipment to obtain roadside collaborative perception information.

[0036] Specifically, in this embodiment of the invention, the perception layer is used for multi-source information collection (core function: comprehensively acquire the basic data required for adjustment, and solve the problem of "incomplete data leading to adjustment deviation"). The perception layer collects key data of roads, drivers, vehicles and roadside through multi-dimensional sub-modules, covering all elements of "scene-people-vehicle-road", and provides a basis for subsequent fusion decision-making.

[0037] 1. Road Scene Perception Module Function: Accurately identifies road types, obstructions, and blind spot locations to support scene prediction.

[0038] 1) High-precision map sub-module: Pre-stores blind spot coordinates (such as the location of the corner point and the range of green cover) for scenarios such as "roads with the same direction and angle (30° / 60°, etc.)" and "main road turning into auxiliary road" nationwide. When a vehicle enters the "100m warning circle in front of the blind spot", the scene activation signal is automatically triggered to solve the problem of "adjustment lag".

[0039] 2) Visual Recognition Submodule: It uses front-view / side-view cameras to recognize lane line changes (such as from solid to blurred), auxiliary road entrance signs, and greenery occlusion outlines to help confirm the current road scene (such as "whether it has entered a road segment with an angle in the same direction") and correct the positioning deviation of the high-precision map.

[0040] 3) Anti-interference perception submodule: It adopts a 24.15GHz anti-clutter millimeter-wave radar, combined with a radar-view fusion algorithm (radar data + camera image), to penetrate obstructions such as greenery and guardrails, filter out clutter from branches and leaves, and output the distance (e.g., 35m), speed (e.g., 25km / h), and azimuth angle (e.g., 45° to the right of the vehicle) of targets in the blind spot (vehicles / pedestrians / non-motorized vehicles), thus solving the problem of "weak anti-obstruction capability of existing perception".

[0041] 2. Driver Status Perception Module Function: Matches the driver's personalized needs and operating intentions to achieve "one policy per person" adjustment.

[0042] 1) Driver Physiological Characteristics Submodule: Combining seat sensors (seat height / fore-and-aft distance) with the driver monitoring system (DMS), it locates the three-dimensional coordinates of the driver's eyes, calculates height and posture parameters (e.g., tall drivers have higher eye positions), and provides a basis for fine-tuning the vertical angle of the rearview mirror (e.g., tall drivers → slightly downward vertically), adapting to personalized vision needs.

[0043] 2) Driving Operation Intent Submodule: Collects turn signal on status (e.g., right turn signal on → anticipating entering the right-hand corner road), steering wheel angle (small angle steering → confirming lane change intention), accelerator / brake pedal depth (emergency braking → triggering emergency visibility adjustment), to avoid "accidental adjustment" (e.g., only passing the corner point without intending to enter).

[0044] 3. Vehicle Status Perception Module Function: Acquire vehicle dynamic data to adapt to the visibility requirements of different driving stages.

[0045] 1) Vehicle Dynamics Submodule: Collects real-time vehicle speed (e.g., low speed ≤30km / h → close-range view, high speed ≥80km / h → long-range view), gearbox gear (R gear → triggers reversing and tilting), and vehicle tilt (tilt / pitch → compensates for vertical angle), providing basic parameters for phased adjustments.

[0046] 2) Driving mode submodule: Identifies vehicle preset modes (such as automatic parking → rearview mirror tilts down significantly, off-road → horizontal angle widens), and optimizes adjustment strategies in conjunction with scenario requirements (such as in off-road mode, even when entering a road with an angle, priority is still given to ensuring the observation of obstacles to the side and rear).

[0047] 4. Roadside Cooperative Sensing Module Function: To supplement blind spots in vehicle-mounted sensing and solve the problem of "single vehicle-mounted sensing failure".

[0048] 1) Roadside Target Transmission Submodule: Receives blind spot overhead images and target data (such as "a truck is approaching 50m from the corner side" identified by the roadside equipment (RSU) on the auxiliary road / angle road) and serves as a backup for perception when the vehicle radar is obscured by dense greenery.

[0049] 2) Vehicle-to-everything (V2X) communication submodule: It enables wireless matching between on-board equipment and roadside equipment through V2X technology. When the matching is successful, it confirms that the vehicle has "entered the roadside sensing coverage area", triggers the adjustment preparation procedure, and activates the system response in advance.

[0050] As an optional implementation, road scene perception information, driver state perception information, vehicle state perception information, and roadside cooperative perception information are weighted and fused to obtain target fused data, which specifically includes: S301. Determine the weight parameters for road scene perception information, driver state perception information, vehicle state perception information, and roadside cooperative perception information. S302. Clean and time-align the road scene perception information, driver state perception information, vehicle state perception information, and roadside cooperative perception information data, and perform time-series fusion based on weight parameters to obtain the target fused data.

[0051] Specifically, the fusion decision layer is used for data processing and strategy generation (core function: to achieve multi-source data collaboration, output precise adjustment instructions, and solve the problem of "fixed strategy"). Through data fusion and intelligent decision-making, the fusion decision layer transforms the fragmented data from the perception layer into executable rearview mirror adjustment instructions, reflecting the logic of "safety first, dynamic adaptation".

[0052] 1. Multi-source information fusion module Function: Integrates multi-dimensional data, eliminates information redundancy and conflict, and generates a unified understanding of the scene and the target.

[0053] 1) Data cleaning and alignment submodule: Performs time synchronization (e.g., unification to millisecond level) and format conversion on heterogeneous data (such as radar distance data, camera image data, map coordinate data) collected by the perception layer, removes abnormal data (such as false detections of radar clutter), and ensures data consistency.

[0054] 2) Priority Weight Allocation Submodule: Data weights are allocated according to "Safety > Scenario > Personalization"—high priority (high-risk targets transmitted from the roadside, reverse gear signals) accounts for 60% of the weight, medium priority (vehicle speed, scenario stage) accounts for 30%, and low priority (driver height, historical preferences) accounts for 10%, avoiding "non-critical data affecting safety decisions". 2. Adjustment Strategy Decision Module Function: Based on the fused data, generate rearview mirror adjustment instructions in stages and for different objectives to solve the problem of "fixed existing strategies".

[0055] 1) Scene Stage Judgment Submodule: Based on the distance between the vehicle and the corner point / auxiliary road entrance, it is divided into three stages: "long distance (80-50m) → medium distance (50-20m) → short distance (<20m)", and the corresponding output is the adjustment direction of "avoiding occlusion → focusing on target → adapting to merge" (e.g., long distance → horizontal expansion 15°, short distance → horizontal return to center and expansion 5°).

[0056] 2) Target Risk Ranking Submodule: Classify blind spot targets according to risk level - high risk (non-motorized vehicles / pedestrians, distance <30m), medium risk (small cars, distance 30-50m), low risk (stationary targets, distance >50m). Prioritize aligning the rearview mirror angle with high-risk targets (such as placing electric vehicles in the center of the mirror) to avoid "distracted vision and missing key targets".

[0057] 3) Angle instruction generation submodule: Combines scene stage, target risk, and driver parameters to calculate specific adjustment angles (e.g., medium distance + high-risk electric vehicle + short driver → horizontal expansion of 12° + vertical slight upward 3°), and outputs them to the execution layer.

[0058] As an optional implementation, the scene stage recognition model is trained through the following steps: S401. Obtain road scene perception samples, driver state perception samples, vehicle state perception samples and roadside cooperative perception samples in the test scenario, and generate sample fusion data based on the road scene perception samples, driver state perception samples, vehicle state perception samples and roadside cooperative perception samples. S402. Determine the primary labels of driving scenarios and secondary labels of scene stages corresponding to the sample fusion data through manual annotation, and generate a training dataset based on the sample fusion data and the corresponding primary labels of driving scenarios and secondary labels of scene stages. S403. Input the training samples into the pre-built convolutional neural network to obtain the predicted scene type and the predicted stage type; S404. Determine the loss value based on the prediction scenario type, prediction stage type, primary driving scenario label, and secondary driving scenario label. S405. Update the parameters of the convolutional neural network based on the loss value to obtain the trained scene stage recognition model; Among them, the secondary tags of the scene stage are the stage sub-tags under the primary tags of the corresponding driving scene.

[0059] Specifically, the process involves acquiring road scene perception samples, driver state perception samples, vehicle state perception samples, and roadside cooperative perception samples in the test scenario. Sample fusion data is generated based on these samples. The primary driving scene label and secondary scene stage label corresponding to the sample fusion data are determined through manual annotation. A training dataset is then generated based on the sample fusion data and the corresponding primary and secondary driving scene labels. The training samples are input into a pre-constructed convolutional neural network to obtain the predicted scene type and predicted stage type. A loss value is determined based on the predicted scene type, predicted stage type, primary driving scene label, and secondary driving scene label. The parameters of the convolutional neural network are updated based on the loss value, completing one round of iterative training. Training stops when the number of iterations reaches a preset threshold or the loss value falls below the preset threshold, resulting in a well-trained scene stage recognition model.

[0060] As an optional implementation, a corresponding target rearview mirror adjustment strategy is determined based on the current driving scenario and the current scenario stage. The target rearview mirror angle is determined based on road scene perception information, driver state perception information, and the target rearview mirror adjustment strategy. Specifically, this includes: S501. Based on the current driving scenario and the current scenario stage, the target rearview mirror adjustment strategy is obtained by matching from the pre-calibrated rearview mirror adjustment strategy library. S502. Determine the type and location of blind spot targets based on road scene perception information, and determine the risk level of blind spot targets based on the type and location of blind spot targets; S503. Determine the driver's posture based on the driver's state perception information; S504. Based on the target risk level in the blind spot and the driver's posture, query the target rearview mirror adjustment strategy to obtain the target rearview mirror angle.

[0061] Specifically, rearview mirror adjustment strategies corresponding to different driving scenarios and different scenario stages are pre-calibrated to form a rearview mirror adjustment strategy library. The rearview mirror adjustment strategies include the blind spot target risk level and the mapping relationship between driver posture and rearview mirror angle.

[0062] As a further optional implementation, the rearview mirror is adjusted according to the target rearview mirror angle, specifically including: S601. Obtain the current rearview mirror angle and determine the rearview mirror angle adjustment amount based on the current rearview mirror angle and the target rearview mirror angle. S602. Determine the horizontal and vertical adjustment angles based on the rearview mirror angle adjustment amount, and determine the angle adjustment rate. S603. Generate horizontal motor control commands and vertical motor control commands based on the horizontal adjustment angle, vertical adjustment angle, and angle adjustment rate. S604. Drive the horizontal adjustment motor and the vertical adjustment motor to adjust the target rearview mirror according to the horizontal motor control command and the vertical motor control command.

[0063] Specifically, the execution layer is used for the dynamic adjustment of the rearview mirror (core function: accurately execute adjustment commands, ensure smooth transition of vision, and solve the problem of "improper reset"). Through hardware drive and control logic, it transforms the commands of the decision layer into changes in the physical angle of the rearview mirror.

[0064] 1. Rearview mirror drive submodule Function: Controls the horizontal / vertical angle adjustment of the rearview mirror. It adopts a high-precision stepper motor (adjustment accuracy ±0.5°) and supports a horizontal adjustment range of -30° to +30° and a vertical adjustment range of -20° to +20° to ensure precise angle control and adapt to the vision requirements of different scenarios (such as reversing down → vertical -15°, high speed → vertical +5°).

[0065] 2. Gradual Adjustment Control Submodule Function: To avoid visual discomfort to the driver caused by sudden angle changes and achieve a "smooth transition" - the adjustment process is gradually changed at a rate of 5° / second (e.g., it takes 3 seconds to expand from the normal angle to 15°); when resetting (after the vehicle has fully entered the lane on the side of the angle), it also returns to the normal angle at a gradual rate, solving the problem of "the existing system resetting too quickly and causing sudden changes in vision".

[0066] The feedback optimization layer is used for effect verification and parameter iteration (core function: forming closed-loop optimization, improving system reliability, and solving the problem of "no feedback closed loop"). By verifying the adjustment effect and iterating parameters, it continuously improves the adjustment accuracy and personalized adaptation capability.

[0067] 1. Adjustment effect verification submodule Function: Confirms whether the rearview mirror adjustment meets the requirement of "visible to the driver and clear target".

[0068] 1) Driver eye tracking submodule: Tracks the driver's eye position through the DMS system. If the eye stays in the rearview mirror area for more than 1 second, it is determined that "the driver has observed the target"; if the stay time is less than 0.5 seconds, an audio-visual prompt is triggered (such as "Please check the rearview mirror field of view"), and the angle is finely adjusted again.

[0069] 2) Target display ratio detection submodule: Analyzes the display ratio of the target in the rearview mirror through the side-view camera. If the ratio is <10% (the target is too small and easy to miss), the adjustment angle is automatically increased (such as horizontal expansion of 3°) until the ratio is ≥10% to ensure that the target is visible.

[0070] 2. Parameter Iteration Optimization Submodule Function: Optimize adjustment strategies based on historical data to enhance personalization and scenario adaptability.

[0071] 1) Historical data storage submodule: Records the driver's adjustment preferences (e.g., a driver habitually expands the left rearview mirror 2° outward each time they enter a road with a 30° angle) and adaptation parameters for different scenarios (e.g., rainy days → increase the horizontal adjustment range by 5°), forming a user-specific database.

[0072] 2) Adaptive Correction Submodule: Iteratively optimizes adjustment instructions based on historical data (e.g., for this driver, when entering a road with a 30° angle next time, directly default to expanding outward by 2° to the left), reducing redundant adjustments and improving the system's intelligence and personalization level.

[0073] like Figure 2 The diagram shown is an instruction illustration of an embodiment of the present invention. This process is a complete closed loop of "trigger-collection-fusion-execution-verification-iteration", covering scenarios such as roads with the same direction and angle, and main road turning into auxiliary road with greenery obstruction. It avoids ineffective adjustments and ensures the effectiveness of the field of vision through two key judgment nodes, and finally achieves dynamic adaptation of the rearview mirror, as detailed below: 1. Start: The system powers on and initializes, activating the perception / decision module and waiting for triggering. Purpose: To ensure the module is ready.

[0074] 2. Step 1: Scene Triggering: High-precision map + visual recognition to identify blind spots, vehicle-road matching to collect pre-warning data, and data collection of turn signals / vehicle speed / turning angle. Purpose: To accurately locate the scene and prevent accidental triggering.

[0075] 3. Judgment 1: Confirm entry into a narrow road angle? Conditions: turn signal on + vehicle speed ≤ 30km / h + small turning angle (< 30°). Branches: "No → Return to step 1, Yes → Proceed to data collection". Purpose: To eliminate accidental touches and prevent invalid adjustments.

[0076] 4. Step 2: Data Acquisition: Radar + Vision acquires road target / occlusion data, driver features data, and vehicle / roadside data. Objective: To obtain comprehensive data across all dimensions of "scene-person-vehicle-road".

[0077] 5. Step 3: Data Fusion: Clean and synchronize data, allocate weights according to "Security > Scenario > Personalization", and divide into long / medium / near scenario stages. Purpose: To unify data and clarify adjustment priorities.

[0078] 6. Step 4: Generate adjustment instructions: Rank target risks, calculate rearview mirror angle + gradient rate, and send execution layer. Purpose: Output precise and executable instructions.

[0079] 7. Step 5: Perform adjustments: adjust the motor angle, gradually change angles at 5° / second, and monitor the angle deviation (≤0.5°). Purpose: To ensure smooth and precise performance for comfort.

[0080] 8. Judgment 2: Is the adjustment effect satisfactory? Conditions include: gaze duration > 1 second + target percentage ≥ 10%. Branches are: "No → fine-tune angle and return to step 5; Yes → proceed to reset iteration." Purpose: To prevent blind spots after adjustment.

[0081] 9. Step 6: Reset + Iteration: After the vehicle enters, gradually reset and save the data to optimize the next strategy. Purpose: To restore normal visibility and improve subsequent accuracy.

[0082] 10. End: The process is closed, and the system returns to monitoring mode for the next trigger. Purpose: To adapt to multiple scenarios for repeated use.

[0083] It can be recognized that the embodiments of the present invention solve the problems of weak anti-obstruction and lagging adjustment of existing technologies by using multi-source fusion perception (penetrating green cover and filtering clutter) + phased adjustment of "far / medium / near" + dual verification of "eye tracking + target proportion". This ensures 100% visibility of blind spot targets in scenarios such as roads with the same direction angle and main road turning into auxiliary road, significantly reducing the risk of collision with non-motorized vehicles and pedestrians. Combining the driver's height, sitting posture and other physiological characteristics + historical adjustment preferences, a "one-person-one-policy" adjustment plan is generated (e.g., tall people slightly downward vertically, short people slightly upward), avoiding the limitations of fixed angle adjustment. At the same time, through 5° / second gradual control, visual abrupt changes in field of view are prevented. The system is designed to balance safety and comfort. It utilizes a feedback optimization layer with a closed loop of "effect verification - data storage - strategy iteration" to record adaptation parameters for different scenarios (such as a 30° angle or 1.2m of greenery obstruction). When entering the same scenario again, it automatically calls the optimization instructions, eliminating the need for recalculation and achieving "increasing accuracy with use." This addresses the issues of lack of iteration and poor long-term reliability in existing systems. By pre-storing multi-scenario blind spot coordinates (main road to auxiliary road, roads with different angles in the same direction) on high-precision maps and using visual recognition for dynamic scene matching, it is not limited to a single road type and is compatible with reversing, low-speed, and high-speed driving modes. Its applicability covers multiple scenarios, including urban roads and suburban highways, enhancing the solution's feasibility and promotional value.

[0084] The present invention will be further described below with reference to two specific embodiments.

[0085] 1. Scenario 1: Same-direction angled road scenario. At 3:00 PM on Sunday, the weather is sunny. A vehicle is traveling on a two-lane road in the suburbs (speed limit 60km / h). 300m ahead, there is a road that forms a 30° angle with the main road (connecting to a rural homestay area). There are 1.2m high bushes on both sides of the road, which partially obstruct the view. The driver plans to drive into the angled road to reach the homestay.

[0086] 1) Road Scene Perception Module High-precision map submodule: Pre-stores the coordinates of the road with the same angle (300m from the current location, angle 30°), road width (4m), and shrub obstruction range (only the 1m area to the left of the entrance at the angle is obstructed). The visual recognition submodule identifies the "homestay guide sign" and "20km / h speed limit sign" at the entrance of the angled road, confirms that there are no fallen bushes blocking the road, and that there are no potholes on the road surface at the angled road. Anti-interference sensing submodule (millimeter-wave radar): Penetrates the obstruction of bushes and detects a sightseeing electric vehicle (speeding 10km / h, traveling along the middle lane) 20m inside the angled road, with no other moving targets.

[0087] 2) Driver Status Perception Module Driving operation intention submodule: Collects turn signal status (not turned on, driver not operating), steering wheel angle (0°, no steering), and determines "there is an intention to drive in but the operation has not been fully started"; Driver physiological characteristics submodule: DMS detects the driver's forward-leaning posture (body slightly tilted towards the angle) and frequent glances towards the entrance of the angle, confirming "paying attention to the road conditions at the angle and not being fatigued or distracted".

[0088] 3) Vehicle Status Perception Module Vehicle Dynamics Submodule: Collects vehicle speed (from 60km / h to 30km / h), brake pedal opening (10%, smooth deceleration), gear (D), and vehicle attitude (no pitch, stable driving). Driving mode submodule: Identifies the current mode as "suburban leisure mode" (system preset, adapted for low-interference scenarios where attention needs to be paid to side roads).

[0089] 4) Roadside Cooperative Sensing Module Vehicle-to-Road Communication Submodule: Matches with roadside equipment along the angled road to receive "No pedestrians crossing the angled road" alerts; Roadside target transmission submodule: Receives the angled road view sent by the roadside camera to confirm that "the electric vehicle is not carrying passengers and its driving trajectory is stable".

[0090] 5) Scene Judgment: The scene stage judgment submodule of the decision layer is integrated with the perception data: "30° angle coordinates of high-precision map + vehicle speed reduced to 30km / h (pre-adapted speed of angled road) + radar detection of electric vehicle within the angle + driver's line of sight focusing angle". When the vehicle is 50m away from the entrance of the angle, it is judged as "same direction angled road scene (pre-entry stage)" and "the driver needs to be reminded to turn on the turn signal and adjust the rearview mirror to avoid bushes obstructing the view".

[0091] 6) Strategy matching and execution Multi-source information fusion module: Cleaning and aligning perception data (filtering radar clutter reflected from bushes), weight allocation - electric vehicle data within the included angle (60%), bush occlusion data (25%), driver operation preparation data (15%). Adjustment strategy decision module: Invoke the "same direction angle road adaptation strategy" - the left side rearview mirror needs to "expand outward to penetrate the bushes and slightly upward to cover the far field of the angle", the right side rearview mirror maintains the normal angle, the adjustment rate is 5° / second, and at the same time, the space for "secondary fine adjustment after the turn signal is turned on" is reserved.

[0092] 7) Execute actions (execution layer) Rearview mirror drive sub-module: Left rearview mirror (angle side): Horizontal angle expands from the normal -5° to +8° (penetrating 1m of bush obstruction on the left and covering electric vehicles 20m away on the angled road), vertical angle slightly increases from the normal 0° to 3° (avoiding bush tops obstructing the far-field view at the angle); Right rearview mirror: Maintains the normal angle (horizontal +5°, vertical 0°) to ensure no gaps in the rear view of the main road; Gradual adjustment control submodule: Adjustments are performed at a rate of 5° / second (it takes 2.6 seconds to expand the left rearview mirror outward by 13° and 0.6 seconds to adjust it vertically upward by 3°). During the adjustment process, the "angle execution progress" is fed back to the fusion decision layer to facilitate subsequent fine-tuning.

[0093] 8) Feedback optimization (feedback optimization layer) Adjustment effect verification submodule: Target display ratio detection: The side-view camera confirms that "the electric vehicle in the angle has a display ratio of 15% (≥10% threshold) in the left rearview mirror" and there are no targets missing in the shrub-covered area; Driver gaze tracking: If the DMS detects that the driver's gaze lingers on the left rearview mirror for 1.5 seconds (>1 second threshold), it determines that "the adjustment is effective and the blind spot is eliminated"; The parameter iteration optimization submodule stores the parameters "30° same-direction angle, 1m shrub obstruction, and adjustment angle +8° +3°", and associates them with "suburban leisure mode". The adjustment logic will be automatically adapted to the next road with the same angle without repeated calculation.

[0094] This specific embodiment relies on the core module of the system to accurately solve the pain points of "blind spots and difficulty in capturing low-speed targets" on roads with the same angle: the rearview mirror angle penetrates the obstruction and adapts to low-speed targets, the gradual adjustment is combined with dynamic optimization based on roadside data, the effect verification ensures that the target is visible, the parameter iteration improves the subsequent response speed, and safe entry can be achieved without external modules, thus enhancing the system's autonomous adjustment capability.

[0095] 2. Scenario 2: Main Road to Side Road. Monday morning rush hour at 7:30, during the city expressway commuting hours, the vehicle is traveling in the middle lane of the main road. 500m ahead is a clearly marked exit from the main road to the side road (the side road connects to a residential area, with moderate non-motorized vehicle traffic during the morning rush hour). The driver has already turned onto the side road according to the navigation plan and is preparing to go to the residential area.

[0096] 1) Road Scene Perception Module High-precision map submodule: Pre-stores the coordinates of the main road exit (500m from the current location), auxiliary road lane parameters (2 lanes, 4.5m wide), and the height difference between the main and auxiliary road surfaces (0.1m, no steps), and simultaneously marks the "high-frequency non-motorized vehicle traffic zone" within 100m of the auxiliary road entrance; Visual recognition submodule: Identifies the "blue auxiliary road sign", "white lane reduction marking (main road 4 → auxiliary road 2)" and "green isolation belt boundary (0.8m high, unobstructed auxiliary road view)" at the main road exit, and confirms that there are no temporary construction barriers; Anti-interference sensing submodule (millimeter-wave radar): detects targets within 100m of the auxiliary road - there is an electric bicycle (speed 15km / h, no tendency to change lanes) 15m to the right of the auxiliary road, and a car (speed 70km / h, safe following distance) 30m behind the main road. No pedestrians are crossing the intersection of the main and auxiliary roads.

[0097] 2) Driver Status Perception Module Driving operation intention submodule: Collects the status of the right turn signal (continuously on for 8 seconds) and the steering wheel angle (slightly turn right by 10°, stabilize for 5 seconds), and determines "clear intention to turn onto the auxiliary road"; Driver physiological characteristics submodule: The DMS system detects the driver's sitting posture (back straight, no excessive forward or backward leaning) and line of sight (focused on the main and auxiliary intersections ahead, no distraction) to confirm that the driving state is stable.

[0098] 3) Vehicle Status Perception Module Vehicle Dynamics Submodule: Collects vehicle speed (smoothly decreasing from 80km / h to 40km / h), brake pedal opening (15%, light braking deceleration), accelerator pedal opening (0%, no acceleration action), and vehicle posture (no body roll, stable driving). Driving mode submodule: Identifies the current mode as "city commuter mode" (system preset, adapted for low-speed driving during morning rush hour).

[0099] 4) Roadside Cooperative Sensing Module Vehicle-to-everything (V2X) submodule: Matched with the roadside unit (RSU) at the entrance of the auxiliary road, it receives data on "real-time traffic density of the auxiliary road (medium)" and "trajectory of electric bicycles (along the right lane)" sent by the roadside unit; Roadside target transmission submodule: Receives the overhead view of the auxiliary road entrance sent by the roadside camera to further verify "no illegal parking occupying the auxiliary road lane".

[0100] 5) Scene Judgment The scenario stage judgment submodule of the integrated decision layer, combined with the perception layer data: "high-precision map main road exit coordinates + right turn signal on for ≥5 seconds + vehicle speed reduced to the auxiliary road adaptation speed (40km / h) + roadside feedback that the auxiliary road is not occupied + driver's line of sight focused on the intersection", when the vehicle is 30m away from the auxiliary road entrance, it is clearly judged as "main road to auxiliary road scenario (safe turning stage)" and abnormal situations such as "temporary lane change hesitation" and "complex road conditions requiring waiting" are excluded.

[0101] 6) Strategy matching and execution Multi-source information fusion module: Cleans and aligns the perception data (removes radar clutter data), and assigns weights according to "safety > scenario > personalization" - roadside electric bicycle data (weight 60%), vehicle steering operation data (weight 30%), and driver's posture data (weight 10%). Adjustment strategy decision module: Query the "Scene-Strategy" mapping library and call the "Main Road to Auxiliary Road Adaptation Strategy" - the right rearview mirror needs to "expand outward to avoid the blind spot of the main road + slightly downward to cover the near field of the auxiliary road", the left rearview mirror needs to "fine-tune to maintain the rear view of the main road", and at the same time set the adjustment rate to 5° / second (to avoid sudden changes in the field of view).

[0102] 7) Execute actions (execution layer) Rearview mirror drive submodule: (1) Right rearview mirror: The horizontal angle is expanded by 10° from the normal 0° (to cover the blind spot at the intersection of the main and auxiliary roads), and the vertical angle is slightly lowered by 2° from the normal 0° (to clearly show electric bicycles on the auxiliary road), with an adjustment accuracy of ±0.5°; (2) Left rearview mirror: The horizontal angle is slightly adjusted from the normal -5° to -3° (to maintain the visibility of cars 30m behind on the main road and avoid missing vehicles behind when turning), while the vertical angle remains unchanged; Gradual adjustment control submodule: Adjustments are performed at a rate of 5° / second (it takes 2 seconds to expand the right rearview mirror outward by 10° and 0.4 seconds to adjust it vertically downward by 2°). The motor position signal is fed back in real time during the adjustment process to ensure that the angle deviation is ≤0.5° without any stuttering or sudden changes.

[0103] 8) Feedback optimization (feedback optimization layer) Target display percentage detection: Confirm "the display percentage of the electric bicycle in the right rearview mirror is 18% (≥10% threshold)" using a side-view camera; Driver gaze tracking: The DMS system detects that the driver's gaze lingers on the right-side rearview mirror for 1.2 seconds (>1 second threshold), and determines that "the adjustment is effective and the field of vision coverage meets the standard".

[0104] This specific embodiment solves the problems of "difficulty in taking into account both the rear view of the main road and the near field of the auxiliary road" and "missing non-motorized vehicles in the blind spot" when turning from the main road to the auxiliary road by coordinating the core modules of the system: the right rearview mirror is expanded outward to accurately cover the blind spot at the intersection of the main road and the auxiliary road, and electric bicycles are clearly visible; the gradual adjustment avoids visual discomfort for the driver; the feedback optimization makes the adjustment more efficient in the next scene, upgrading from "passive adaptation" to "active and precise matching".

[0105] The method steps of the embodiments of the present invention have been described above. It is understood that the embodiments of the present invention identify the current driving scenario and current scenario stage based on road scene perception information, driver state perception information, vehicle state perception information, and roadside cooperative perception information. Then, they match the corresponding rearview mirror adjustment strategy, and further combine the road scene perception information and driver state perception information to determine the corresponding target rearview mirror angle. This improves the accuracy and timeliness of vehicle rearview mirror adjustment, thereby enhancing vehicle driving safety and the user's driving experience.

[0106] Compared with the prior art, the embodiments of the present invention have the following advantages: 1) Pre-activation adjustment allows drivers to observe targets on the angled side of the road through the rearview mirror before entering the blind spot of the road with an angle in the same direction, eliminating the visual gap caused by adjustment lag, greatly reducing the risk of collision with vehicles and pedestrians on the angled side, and improving driving safety.

[0107] 2) Anti-occlusion perception technology ensures that even in scenarios where greenery or guardrails obstruct the view, it can still accurately identify targets in blind spots, avoid mistaking stationary obstructions for obstacles or missing moving targets, and enable drivers to keep abreast of the dynamics of the road on the angled side and deal with complex road conditions with ease.

[0108] 3) A phased and target priority adjustment strategy allows the rearview mirror angle to dynamically change with the distance between the vehicle and the corner point and the target risk. This avoids both an overly wide field of vision that distracts attention and the omission of high-risk targets, adapting to different corner scenarios and optimizing the driver's observation experience.

[0109] 4) Vehicle-road cooperation supplements the blind spots of perception, and the feedback mechanism verifies the adjustment effect and accurately resets the system. This avoids the adjustment from stalling when the vehicle-mounted perception fails alone, and also prevents premature reset from affecting the field of vision. This ensures the reliability of the adjustment and the stability of subsequent driving, and improves the overall practicality of the system.

[0110] like Figure 3 The diagram shown is a structural schematic of a vehicle rearview mirror adjustment system based on multi-source information fusion provided in an embodiment of the present invention. (Refer to...) Figure 3 This invention provides a vehicle rearview mirror adjustment system based on multi-source information fusion, comprising: The information acquisition module is used to acquire road scene perception information, driver status perception information, vehicle status perception information, and roadside cooperative perception information. The data fusion module is used to perform weight allocation and data fusion on road scene perception information, driver state perception information, vehicle state perception information and roadside cooperative perception information to obtain target fused data. The scene stage recognition module is used to input the target fusion data into the pre-trained scene stage recognition model to obtain the current driving scene and the corresponding current scene stage; The adjustment module is used to determine the corresponding target rearview mirror adjustment strategy based on the current driving scenario and the current scenario stage. It determines the target rearview mirror angle based on road scene perception information, driver state perception information and the target rearview mirror adjustment strategy, and then adjusts the rearview mirror according to the target rearview mirror angle.

[0111] It is understood that the content of the above method embodiments is applicable to this system embodiment. The specific functions implemented by this device embodiment are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0112] This invention also provides an electronic device, comprising: a memory, a processor, a program stored in the memory and executable on the processor, and a data bus for communication between the processor and the memory. When the program is executed by the processor, it implements the aforementioned vehicle rearview mirror adjustment method based on multi-source information fusion. This electronic device can be any smart terminal, including a tablet computer or an in-vehicle computer.

[0113] like Figure 4 The diagram shown is a hardware structure schematic of an electronic device provided in an embodiment of the present invention. (Refer to...) Figure 4 This invention provides an electronic device, comprising: The processor 401 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present invention. The memory 402 can be implemented as a read-only memory (ROM), static storage device, dynamic storage device, or random access memory (RAM). The memory 402 can store the operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 402 and is called and executed by the processor 401 to implement the vehicle rearview mirror adjustment method based on multi-source information fusion according to the embodiments of the present invention. Input / output interface 403 is used to implement information input and output; The communication interface 404 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). Bus 405 transmits information between various components of the device (e.g., processor 401, memory 402, input / output interface 403, and communication interface 404); The processor 401, memory 402, input / output interface 403 and communication interface 404 are connected to each other within the device via bus 405.

[0114] It is understood that the content of the above method embodiments is applicable to this device embodiment. The specific functions implemented by this device embodiment are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0115] This invention also provides a vehicle, which includes the electric drive assembly of the vehicle rearview mirror adjustment system or electronic device based on multi-source information fusion.

[0116] The vehicle can be a private car, such as a sedan, SUV, MPV, or pickup truck. It can also be a commercial vehicle, such as a van, bus, small truck, or large semi-trailer. The vehicle must have an electric motor capable of outputting power or acting as a generator to store mechanical energy. When the vehicle is a new energy vehicle, it can be a hybrid or a pure electric vehicle.

[0117] Since the vehicle applies all the technical solutions of the above-mentioned vehicle rearview mirror adjustment system or electronic device based on multi-source information fusion, it has at least all the beneficial effects brought about by the technical solutions of the above embodiments, which will not be repeated here.

[0118] This invention also provides a storage medium, which is a computer-readable storage medium for computer-readable storage. The storage medium stores one or more programs, which can be executed by one or more processors to implement the above-described vehicle rearview mirror adjustment method based on multi-source information fusion.

[0119] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0120] It is understood that the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0121] This invention also provides a computer program product or computer program, which includes computer instructions stored in a computer-readable storage medium. A processor of a computer device can read the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, causing the computer device to perform... Figure 1 The method shown.

[0122] It is understood that the content of the above method embodiments is applicable to the embodiments of this program product. The specific functions implemented by the embodiments of this program product are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0123] The embodiments described in this invention are for the purpose of more clearly illustrating the technical solutions of the embodiments of this invention, and do not constitute a limitation on the technical solutions provided by the embodiments of this invention. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this invention are also applicable to similar technical problems.

[0124] The terms "first," "second," "third," "fourth," etc. (if present) in the specification and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0125] In some alternative embodiments, the functions / operations mentioned in the block diagrams may not occur in the order shown in the operation diagrams. For example, depending on the functions / operations involved, two consecutively shown blocks may actually be executed substantially simultaneously, or the aforementioned blocks may sometimes be executed in reverse order. Furthermore, the embodiments presented and described in the flowcharts of this invention are provided by way of example to provide a more comprehensive understanding of the technology. The disclosed methods are not limited to the operations and logic flows presented herein. Alternative embodiments are contemplated in which the order of various operations is changed and sub-operations described as part of a larger operation are executed independently.

[0126] Furthermore, although the invention has been described in the context of functional modules, it should be understood that, unless otherwise stated, one or more of the aforementioned functions and / or features may be integrated into a single physical device and / or software module, or one or more functions and / or features may be implemented in a separate physical device or software module. It is also understood that a detailed discussion of the actual implementation of each module is unnecessary for understanding the invention. Rather, given the properties, functions, and internal relationships of the various functional modules in the apparatus disclosed herein, the actual implementation of the module will be understood within the scope of conventional skill of an engineer. Therefore, those skilled in the art can implement the invention as set forth in the claims using ordinary techniques without excessive experimentation. It is also understood that the specific concepts disclosed are merely illustrative and not intended to limit the scope of the invention, which is determined by the full scope of the appended claims and their equivalents.

[0127] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium 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 described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0128] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-including system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0129] More specific examples (a non-exhaustive list) of computer-readable media include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the aforementioned program can be printed, because the aforementioned program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0130] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0131] In the foregoing description of this specification, references to terms such as "one embodiment," "another embodiment," or "some embodiments" indicate that a specific feature, structure, material, or characteristic described in connection with an embodiment or example is included in at least one embodiment or example of the present invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0132] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

[0133] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the above embodiments. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A method for adjusting vehicle rearview mirrors based on multi-source information fusion, characterized in that, Includes the following steps: Acquire road scene perception information, driver status perception information, vehicle status perception information, and roadside cooperative perception information; Weights are assigned and data fusion is performed on the road scene perception information, the driver state perception information, the vehicle state perception information, and the roadside cooperative perception information to obtain target fused data; The target fusion data is input into a pre-trained scene stage recognition model to obtain the current driving scene and the corresponding current scene stage. The target rearview mirror adjustment strategy is determined based on the current driving scenario and the current scenario stage. The target rearview mirror angle is determined based on the road scene perception information, the driver state perception information, and the target rearview mirror adjustment strategy. Then, the rearview mirror is adjusted based on the target rearview mirror angle.

2. The vehicle rearview mirror adjustment method based on multi-source information fusion according to claim 1, characterized in that, The acquisition of road scene perception information, driver state perception information, vehicle state perception information, and roadside cooperative perception information specifically includes: Based on road images, radar detection information, and high-precision maps, road types, occlusion conditions, blind spot target types, and blind spot target locations are identified to obtain the road scene perception information; Based on seat sensor data, driver images, and driver operation information, the driver's physiological state and driver's operation intentions are identified to obtain the driver state perception information. The vehicle status perception information is obtained by acquiring real-time vehicle status and vehicle driving mode through the body controller; The system receives regional global images and regional vehicle positioning data sent by roadside equipment to obtain the roadside collaborative perception information.

3. The vehicle rearview mirror adjustment method based on multi-source information fusion according to claim 1, characterized in that, The step of weighting and fusing the road scene perception information, driver state perception information, vehicle state perception information, and roadside cooperative perception information to obtain target fused data specifically includes: Determine the weight parameters for the road scene perception information, the driver state perception information, the vehicle state perception information, and the roadside cooperative perception information; The road scene perception information, driver state perception information, vehicle state perception information, and roadside cooperative perception information are cleaned and time-aligned, and then time-series fusion is performed based on the weight parameters to obtain the target fused data.

4. The vehicle rearview mirror adjustment method based on multi-source information fusion according to claim 1, characterized in that, The scene stage recognition model is trained through the following steps: Obtain road scene perception samples, driver state perception samples, vehicle state perception samples, and roadside cooperative perception samples in the test scenario, and generate sample fusion data based on the road scene perception samples, driver state perception samples, vehicle state perception samples, and roadside cooperative perception samples. The primary labels for driving scenarios and the secondary labels for driving stages corresponding to the sample fusion data are determined by manual annotation. A training dataset is then generated based on the sample fusion data and the corresponding primary labels for driving scenarios and secondary labels for driving stages. The training samples are input into a pre-built convolutional neural network to obtain the predicted scene type and the predicted stage type; The loss value is determined based on the predicted scenario type, the predicted stage type, the primary label of the driving scenario, and the secondary label of the driving scenario. The parameters of the convolutional neural network are updated based on the loss value to obtain the trained scene stage recognition model; The secondary stage label of the scenario is a stage subdivision label under the primary stage label of the driving scenario.

5. The vehicle rearview mirror adjustment method based on multi-source information fusion according to claim 1, characterized in that, The step of determining the corresponding target rearview mirror adjustment strategy based on the current driving scenario and the current scenario stage, and determining the target rearview mirror angle based on the road scene perception information, the driver state perception information, and the target rearview mirror adjustment strategy, specifically includes: The target rearview mirror adjustment strategy is obtained by matching the current driving scenario and the current scenario stage from a pre-defined rearview mirror adjustment strategy library. The blind spot target type and location are determined based on the road scene perception information, and the blind spot target risk level is determined based on the blind spot target type and location. The driver's posture is determined based on the driver's state perception information; Based on the target risk level in the blind spot and the driver's posture, the target rearview mirror adjustment strategy is queried to obtain the target rearview mirror angle.

6. A method for adjusting a vehicle rearview mirror based on multi-source information fusion according to any one of claims 1 to 5, characterized in that, The adjustment of the rearview mirror according to the target rearview mirror angle specifically includes: Obtain the current rearview mirror angle, and determine the rearview mirror angle adjustment amount based on the current rearview mirror angle and the target rearview mirror angle; The horizontal and vertical adjustment angles are determined based on the rearview mirror angle adjustment amount, and the angle adjustment rate is also determined. Horizontal motor control commands and vertical motor control commands are generated based on the horizontal adjustment angle, the vertical adjustment angle, and the angle adjustment rate. The horizontal and vertical adjustment motors are driven to adjust the target rearview mirror according to the horizontal motor control command and the vertical motor control command.

7. A vehicle rearview mirror adjustment system based on multi-source information fusion, characterized in that, include: The information acquisition module is used to acquire road scene perception information, driver status perception information, vehicle status perception information, and roadside cooperative perception information. The data fusion module is used to perform weight allocation and data fusion on the road scene perception information, the driver state perception information, the vehicle state perception information and the roadside cooperative perception information to obtain target fused data; The scene stage recognition module is used to input the target fusion data into the pre-trained scene stage recognition model to obtain the current driving scene and the corresponding current scene stage; The adjustment module is used to determine the corresponding target rearview mirror adjustment strategy based on the current driving scenario and the current scenario stage, determine the target rearview mirror angle based on the road scene perception information, the driver state perception information and the target rearview mirror adjustment strategy, and then adjust the rearview mirror according to the target rearview mirror angle.

8. An electronic device, characterized in that, The electronic device includes a memory, a processor, a program stored in the memory and executable on the processor, and a data bus for enabling communication between the processor and the memory. When the program is executed by the processor, it implements the vehicle rearview mirror adjustment method based on multi-source information fusion as described in any one of claims 1 to 6.

9. A vehicle, characterized in that, The vehicle includes the vehicle rearview mirror adjustment system based on multi-source information fusion as described in claim 7 or the electronic device as described in claim 8.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by the processor, it implements the vehicle rearview mirror adjustment method based on multi-source information fusion as described in any one of claims 1 to 6.

Images