Vehicle navigation service resource management method and system

Abstract

The invention relates to the technical field of vehicle-mounted navigation service resource management, in particular to a vehicle-mounted navigation service resource management method and system, and the method comprises the following steps: sensing the signal quality of a global navigation satellite system in an environment where a vehicle-mounted navigation system is located and the load state of a central processing unit in the system; when the signal of the global navigation satellite system is poor and the load of a central processing unit in the system is saturated, adaptively adjusting the calculation complexity of the data processing algorithm of the inertial measurement unit to obtain the adjusted data processing algorithm of the inertial measurement unit; executing path planning in stages, and preferentially generating a coarse-grained path; allocating high-priority resources for the voice guidance task, and loading voice data in advance; predictively triggering voice guidance; and starting a real-time sensing and risk identification process aiming at the key decision point on the coarse-grained path. The real-time performance of the navigation service and the user experience feeling can be improved.

Description

Technical Field

[0001] This invention relates to the technical field of in-vehicle navigation service resource management, and specifically to an in-vehicle navigation service resource management method and system. Background Technology

[0002] In modern in-vehicle infotainment systems, navigation is a core function, typically providing real-time route guidance and map display while the vehicle is in motion. However, in practice, in-vehicle navigation systems often encounter a series of complex and interconnected technical challenges. For example, when a driver actively switches to a high-resolution 3D city model view, the system needs to render more detailed building models, terrain features, texture maps, and richer point-of-interest information. This requires the system to read and cache a much larger amount of detailed map tile data from its built-in flash memory chip than is required for conventional 2D maps. This data must be loaded into the system memory before being rendered by the graphics processing unit. Therefore, this request for a high-resolution 3D view immediately puts significant pressure on system memory, manifesting as a sharp increase in memory usage and a high density of memory allocation requests, leading to sluggish navigation response. Summary of the Invention

[0003] The purpose of this invention is to address the aforementioned shortcomings by proposing a method and system for managing vehicle navigation service resources.

[0004] The present invention adopts the following technical solution: A method for managing in-vehicle navigation service resources, the method comprising the following steps: The system senses the signal quality of the global navigation satellite system in the environment in which the vehicle navigation system is located, as well as the load status of the central processing unit inside the system. When the signal of the global navigation satellite system is poor and the central processing unit inside the system is saturated, the computational complexity of the inertial measurement unit data processing algorithm is adaptively adjusted to obtain the adjusted inertial measurement unit data processing algorithm. Based on the adjusted inertial measurement unit data processing algorithm, path planning is performed in stages, with priority given to generating coarse-grained paths; Allocate high-priority resources to voice guidance tasks and preload voice data; By combining vehicle speed and positioning module confidence, and based on coarse-grained path, voice guidance is predictively triggered using pre-loaded voice data. A real-time perception and risk identification process is initiated for key decision points on the coarse-grained path. The system uses onboard cameras to identify physical obstacles, lane changes, and traffic signs at these key decision points, generating visual perception results. Onboard millimeter-wave radar and ultrasonic sensors assess the traffic conditions at these key decision points, providing a trafficability assessment result. This assessment includes the presence of a large number of stationary or extremely slow-moving vehicles in the key lanes. If the visual perception results conflict with the static map information used for coarse-grained path planning, or if a large number of stationary or extremely slow-moving vehicles are present in the key lanes, a risk is identified at the key decision point. If a risk is identified, customized warning information is generated, including the location of the conflict point and suggested driving behavior. The generation, display, and voice broadcast of warning information are assigned the highest priority. The warning information is displayed graphically on the navigation screen, highlighting the adjacent lanes to which the driver should change lanes. Voice guidance synchronized with actual road conditions is issued via the voice broadcast module. The triggering timing of the warning information is dynamically calculated based on the vehicle's current speed and distance from the key decision point.

[0005] This technical solution can effectively solve the problems of slow navigation guidance response and delayed path planning updates in existing technologies by adaptively adjusting the computational complexity of the inertial measurement unit data processing algorithm, optimizing resource allocation, and performing path planning in stages under complex environments with poor global navigation satellite system signals and saturated central processing unit load. It prioritizes the generation of coarse-grained paths, allocates high-priority resources for voice guidance tasks, and predictively triggers them.

[0006] This application also discloses an in-vehicle navigation service resource management system, applied to an in-vehicle navigation service resource management method, the system comprising: The perception module senses the signal quality of the global navigation satellite system in the environment in which the vehicle navigation system is located, as well as the load status of the central processing unit inside the system. The adjustment module adaptively adjusts the computational complexity of the inertial measurement unit (IMU) data processing algorithm when the global navigation satellite system signal is poor and the internal central processing unit is saturated, thus obtaining the adjusted IMU data processing algorithm. The planning module, based on the adjusted inertial measurement unit data processing algorithm, performs path planning in stages, prioritizing the generation of coarse-grained paths; The allocation module allocates high-priority resources to voice guidance tasks and preloads voice data; The triggering module, combining vehicle speed and positioning module confidence, predictively triggers voice guidance based on coarse-grained path and using pre-loaded voice data; The processing module initiates a real-time perception and risk identification process for key decision points on the coarse-grained path. It uses onboard cameras to identify physical obstacles, lane changes, and traffic signs at these key decision points, generating visual perception results. It then assesses the traffic conditions at these key decision points using onboard millimeter-wave radar and ultrasonic sensors, obtaining a trafficability assessment result. This assessment includes whether there are a large number of stationary or extremely slow-moving vehicles in the key lanes. When the visual perception results conflict with the static map information used for coarse-grained path planning, or when there are a large number of stationary or extremely slow-moving vehicles in the key lanes, the module determines that the key decision point poses a risk. If a risk is identified, customized warning information is generated, including the location of the conflict point and suggested driving behavior. The module assigns the highest priority to the generation, display, and voice broadcast of warning information. The warning information is displayed graphically on the navigation screen, highlighting the adjacent lanes to which the driver should change lanes. Voice guidance synchronized with actual road conditions is issued via the voice broadcast module. Finally, the module dynamically calculates the triggering timing of the warning information based on the vehicle's current speed and distance from the key decision point.

[0007] Through this technical solution, the system can intelligently adjust the data processing algorithm of the inertial measurement unit, optimize the path planning strategy, and prioritize resources for voice guidance tasks when the vehicle navigation system faces challenges such as poor global navigation satellite system signals and central processor load saturation through the coordinated work of various modules. This effectively improves the real-time performance, accuracy, and user experience of navigation services and solves the problem of slow navigation response in existing technologies.

[0008] This application adaptively adjusts the computational complexity of the inertial measurement unit data processing algorithm, enabling the system to dynamically optimize the computational load based on actual resource conditions and avoid overloading the central processing unit. Phased path planning prioritizes the generation of coarse-grained paths, quickly providing basic navigation information and ensuring navigation continuity. High-priority resources are allocated to voice guidance tasks, and voice data is pre-loaded. Predictive triggering, combined with vehicle speed and positioning module confidence levels, ensures the timely and accurate delivery of critical navigation information, significantly improving the real-time performance and user experience of the navigation service.

[0009] To further understand the features and technical content of the present invention, please refer to the following detailed description and drawings of the present invention. However, the drawings provided are for reference and illustration only and are not intended to limit the present invention. Attached Figure Description

[0010] Figure 1 This is a flowchart of the in-vehicle navigation service resource management method of the present invention; Figure 2 This is a schematic diagram of the structure of the vehicle navigation service resource management system of the present invention. Detailed Implementation

[0011] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can understand the advantages and effects of the present invention from the content disclosed in this specification. The present invention can be implemented or applied through other different specific embodiments, and various details in this specification can also be modified and changed based on different viewpoints and applications without departing from the spirit of the present invention. Furthermore, the accompanying drawings of the present invention are for simple illustrative purposes only and are not depictions of actual dimensions; this is stated in advance. The following embodiments will further describe the relevant technical content of the present invention in detail, but the disclosed content is not intended to limit the scope of protection of the present invention.

[0012] This embodiment provides a method and system for managing vehicle navigation service resources, combined with... Figure 1 and Figure 2 As shown.

[0013] refer to Figure 1 A method for managing vehicle navigation service resources, the method comprising the following steps: The system senses the signal quality of the global navigation satellite system in the environment in which the vehicle navigation system is located, as well as the load status of the central processing unit inside the system. When the signal of the global navigation satellite system is poor and the central processing unit inside the system is saturated, the computational complexity of the inertial measurement unit data processing algorithm is adaptively adjusted to obtain the adjusted inertial measurement unit data processing algorithm. Based on the adjusted inertial measurement unit data processing algorithm, path planning is performed in stages, with priority given to generating coarse-grained paths; Allocate high-priority resources to voice guidance tasks and preload voice data; By combining vehicle speed and positioning module confidence, and based on coarse-grained path, voice guidance is predictively triggered using pre-loaded voice data. A real-time perception and risk identification process is initiated for key decision points on the coarse-grained path. The system uses onboard cameras to identify physical obstacles, lane changes, and traffic signs at these key decision points, generating visual perception results. Onboard millimeter-wave radar and ultrasonic sensors assess the traffic conditions at these key decision points, providing a trafficability assessment result. This assessment includes the presence of a large number of stationary or extremely slow-moving vehicles in the key lanes. If the visual perception results conflict with the static map information used for coarse-grained path planning, or if a large number of stationary or extremely slow-moving vehicles are present in the key lanes, a risk is identified at the key decision point. If a risk is identified, customized warning information is generated, including the location of the conflict point and suggested driving behavior. The generation, display, and voice broadcast of warning information are assigned the highest priority. The warning information is displayed graphically on the navigation screen, highlighting the adjacent lanes to which the driver should change lanes. Voice guidance synchronized with actual road conditions is issued via the voice broadcast module. The triggering timing of the warning information is dynamically calculated based on the vehicle's current speed and distance from the key decision point.

[0014] Among them, "Global Navigation Satellite System (GNSS) signal quality" refers to indicators such as the strength, signal-to-noise ratio, and number of available GNSS signals received by the vehicle-mounted navigation system. These indicators directly affect the accuracy and reliability of positioning. "Central Processing Unit (CPU) load status" refers to the utilization rate of the CPU within the vehicle-mounted navigation system, reflecting the workload of its processing tasks. "Inertial Measurement Unit (INS) data processing algorithm" refers to the algorithm used to process INS sensor data for dead reckoning; its computational complexity directly affects processing speed and accuracy. "Coarse-grained path" refers to a simplified path generated in the early stages of path planning, containing major road segments and key decision points; its level of detail is lower than the final fine-grained path. "Positioning module confidence level" refers to the reliability assessment of the vehicle-mounted navigation system's estimation of the current vehicle position, usually expressed as probability or error range. The implementation environment of this application is typically a vehicle equipped with an INS navigation system, a GNSS receiver, an INS, a CPU, a memory, a display screen, and a voice broadcast module.

[0015] The in-vehicle navigation service resource management method proposed in this application firstly senses the signal quality of the Global Navigation Satellite System (GNSS) in the environment in which the in-vehicle navigation system operates, as well as the load status of the central processing unit (CPU) within the system. For example, it can monitor parameters such as signal strength, number of satellites, and positioning accuracy in real time through a GNSS receiver to assess the GNSS signal quality. When the signal strength is below a preset threshold or the number of available satellites decreases, the GNSS signal quality can be determined to be poor. Simultaneously, it can obtain information such as CPU utilization and task queue length in real time through interfaces provided by the operating system or hardware monitoring modules to assess the CPU load status. When the CPU utilization consistently exceeds a certain threshold, the CPU load can be determined to be saturated.

[0016] Secondly, when the global navigation satellite system signal is poor and the internal central processing unit (CPU) is saturated, this method can adaptively adjust the computational complexity of the inertial measurement unit (IMU) data processing algorithm to obtain an adjusted IMU data processing algorithm. For example, multiple processing modes with different computational complexities can be preset in the IMU data processing algorithm. When system resources are scarce, the sampling frequency of the IMU data fusion algorithm can be reduced, or the model complexity of state estimation algorithms such as Kalman filtering can be simplified to reduce the computational burden on the CPU. This adjustment can be dynamic, switching in real time according to changes in signal quality and CPU load.

[0017] Furthermore, based on the adjusted inertial measurement unit (IMU) data processing algorithm, this method can perform path planning in stages, prioritizing the generation of coarse-grained paths. For example, in the initial stage of path planning, a simplified IMU data processing algorithm can be used to quickly generate a coarse path containing only main roads and key turning points. This coarse-grained path requires less computation and can quickly provide navigation direction even with limited resources. Subsequently, when system resources are more readily available, or when the vehicle approaches a key decision point, the path is gradually refined to generate a fine-grained path containing lane-level information, specific steering maneuvers, and other details.

[0018] Furthermore, this method can allocate high-priority resources to voice guidance tasks and preload voice data. For example, voice guidance tasks can be set as high priority in the operating system scheduler to ensure they receive priority access to CPU time slices and memory resources. Simultaneously, based on coarse-grained paths or predicted driving trajectories, potentially needed voice guidance data can be preloaded from storage into memory, avoiding loading only when needed and thus reducing latency.

[0019] Finally, this method combines vehicle speed and positioning module confidence levels, and based on coarse-grained paths, uses pre-loaded voice data to predictively trigger voice guidance. For example, when the vehicle speed is high or the positioning module confidence level is high, voice guidance can be triggered earlier, allowing the driver more reaction time. Conversely, when the vehicle speed is slow or the positioning module confidence level is low, the triggering of voice guidance can be appropriately delayed to ensure accuracy. This predictive triggering mechanism, combining the vehicle's real-time status and the navigation path, makes voice guidance more timely and effective.

[0020] Critical decision points refer to locations where the driver needs to make important judgments or take significant actions during vehicle operation, such as intersections, highway entrances and exits, lane merging points, or lane diverging points. The real-time perception and risk identification process aims to continuously monitor the environment and assess potential hazards at these critical points.

[0021] Specifically, vehicle-mounted cameras can capture images or video streams of the environment in front of or around the vehicle, and use image processing and computer vision algorithms to analyze the image content in real time. Physical obstacles can include, but are not limited to, fallen objects, construction barriers, and accident vehicles; lane line changes can refer to the disappearance, addition, narrowing, or widening of lane lines; traffic signs include speed limit signs, no-entry signs, and turn indicator signs. The visual perception result is the identification and location information of these elements.

[0022] Millimeter-wave radar can accurately measure the distance, speed, and angle of objects ahead by transmitting and receiving millimeter-wave signals, and it excels particularly in adverse weather conditions. Ultrasonic sensors are suitable for detecting obstacles at close range. Traffic condition assessment refers to the comprehensive use of data from these sensors to determine whether key lanes are unobstructed. For example, by using radar to detect the number of vehicles, speed distribution, and location information, it is possible to determine whether there are a large number of stationary or very slow-moving vehicles, thereby assessing the degree of congestion or traffic obstruction in the lanes.

[0023] There may be discrepancies between visual perception results and static map information. For example, a static map may show that a lane is clear, but the camera may detect construction barriers in that lane; or a static map may show that lane lines should be present at a certain location, but the lane lines on the actual road surface may be blurred or missing. When these real-time perceived dynamic information do not match the preset static information, or when sensors detect obstacles that seriously affect traffic flow in key lanes (such as congestion caused by a large number of stationary or extremely low-speed vehicles), then the key decision point is considered to have a potential risk.

[0024] Customized warning messages are generated based on the specific type, location, and severity of the risk, aiming to provide drivers with clear and targeted guidance. The conflict point location can precisely indicate the exact location where the risk occurs, such as "There is an obstacle in the left lane at the intersection 200 meters ahead." Suggested driving behaviors are specific operational suggestions given to the driver in response to this risk situation, such as "Please slow down," "Suggest changing lanes to the right lane," or "Be cautious and avoid the obstacle."

[0025] When system resources are limited, tasks related to risk warnings will be prioritized over other non-urgent tasks to ensure that warning information can be generated and presented to the driver as quickly as possible, thereby gaining valuable reaction time.

[0026] Graphical overlay display refers to presenting warning information directly on the map or traffic view in a prominent manner on the regular navigation interface (e.g., using specific colors, flashing icons, or animation effects), closely integrating it with the actual traffic situation. Highlighting the adjacent lane to which the driver is advised to change lanes further guides the driver to take the recommended driving actions intuitively; for example, highlighting the recommended lane change lane in green or blue.

[0027] The voice broadcast module generates corresponding voice prompts based on real-time perceived changes in road conditions and risk assessment results. Synchronization with actual road conditions means that the content, timing, and speed of the voice guidance are consistent with the vehicle's current environment and the driver's reaction needs, ensuring the timeliness and effectiveness of information delivery.

[0028] The timing of warning messages is not fixed but adjusted based on the vehicle's real-time status. For example, when the vehicle speed is high, the warning message will be triggered earlier to allow the driver more reaction time; when the vehicle speed is low or the vehicle is close to a critical decision point, the warning message may be triggered at a closer distance to avoid premature information that could cause driver fatigue or confusion. This dynamic calculation ensures the timeliness and effectiveness of the warning messages.

[0029] The vehicle navigation service resource management method proposed in this application works by employing a series of collaborative mechanisms, including intelligent perception, adaptive adjustment, phased planning, priority allocation, and predictive triggering, to effectively address the resource challenges faced by vehicle navigation systems in complex environments. When a vehicle navigation system detects poor signal quality from the Global Navigation Satellite System and saturation of its internal central processing unit (CPU), traditional methods may further exacerbate the CPU burden due to the excessive computational load of the inertial measurement unit (INS) data processing algorithm, leading to severe delays in path planning and voice guidance. This application, by adaptively adjusting the computational complexity of the INS data processing algorithm, effectively reduces the CPU's computational pressure while ensuring basic positioning requirements, freeing up valuable computing resources for the smooth execution of subsequent tasks.

[0030] With CPU resources alleviated, this application employs a phased path planning strategy, prioritizing the generation of coarse-grained paths. This strategy avoids computationally intensive fine-grained path planning under resource constraints, thus quickly providing drivers with a general navigation direction and resolving the problem of slow path planning updates in traditional methods. Simultaneously, allocating high-priority resources to voice guidance tasks and pre-loading voice data ensures that critical voice guidance can be delivered promptly and smoothly even under resource-constrained conditions, avoiding guidance interruptions or delays caused by data loading latency.

[0031] Finally, by combining vehicle speed and positioning module confidence levels, and based on coarse-grained path mapping, pre-loaded voice data is used to predictively trigger voice guidance. This mechanism makes the timing of voice guidance triggering more intelligent and user-friendly, enabling guidance to be issued to the driver at the optimal time based on the vehicle's actual driving status and positioning reliability. This effectively corrects the problem of traditional navigation systems where the navigation still displays the path before the intersection even after the vehicle has passed it, significantly improving the real-time performance and accuracy of navigation. Through the above collaborative efforts, the overall technical solution of this application ensures that in-vehicle navigation services can provide a stable, efficient, and accurate navigation experience in various complex environments.

[0032] The method also includes the following steps: The reliability of vehicle positioning information is evaluated, including assessing whether there is a persistent inconsistency between inertial measurement unit data and vehicle motion state, which includes wheel speed sensor data and steering angle sensor data acquired from the vehicle bus. When there is a persistent inconsistency between the inertial measurement unit data and the vehicle's motion state, the flexible path planning mode is activated. Delineate the search area centered on the estimated vehicle location, and query and extract predefined semantic anchor points; A comprehensive matching score is calculated by combining the vehicle's heading, visual features within the current field of view identified by the onboard camera, and the distance between the estimated vehicle position and the center point of the semantic anchor point. Select semantic anchors with a comprehensive matching score higher than a preset matching threshold as the initial logical location of the vehicle; Based on wheel speed sensor data and steering angle sensor data, the vehicle's short-term historical driving trajectory is reconstructed; The reconstructed vehicle history trajectory is compared with the road topology that associates the vehicle's initial logical location; If the reconstructed vehicle history trajectory is highly consistent with the road topology associated with the vehicle's initial logical location, the vehicle's semantic location is confirmed. Based on the confirmed semantic location of the vehicle, update the navigation instructions to correct erroneous instructions generated due to errors; Based on the updated navigation instructions, the vehicle icon is displayed, the driving route is highlighted, and voice guidance for correction is given.

[0033] Specifically, assessing the reliability of vehicle positioning information aims to determine whether the current vehicle positioning data is accurate and reliable. Inertial measurement unit (IMU) data, typically provided by accelerometers and gyroscopes, reflects the vehicle's attitude and motion changes. The vehicle's motion state is characterized by wheel speed sensor data and steering angle sensor data acquired from the vehicle bus, directly reflecting the vehicle's actual speed and steering behavior. By evaluating whether there are persistent inconsistencies between IMU data and the vehicle's motion state, potential drift, errors, or external interference in the IMU data can be effectively detected, thus determining the reliability of the positioning information. For example, if the IMU data shows the vehicle is accelerating, but the wheel speed sensor data shows the vehicle speed is stable or decelerating, then a persistent inconsistency exists.

[0034] When there is a persistent inconsistency between the inertial measurement unit (IMU) data and the vehicle's motion state, it indicates that the current positioning information may have a significant deviation. At this point, the flexible path planning mode is activated. The flexible path planning mode can be understood as a more robust and fault-tolerant path planning strategy. It no longer relies entirely on a single positioning source that may have errors, but instead combines multiple pieces of information for comprehensive judgment and correction.

[0035] Furthermore, to provide accurate navigation even when location information is unreliable, the system defines a search area centered on the vehicle's estimated location. Within this search area, the system queries and extracts predefined semantic anchors. Semantic anchors are feature points with clear geographical locations and semantic information, such as intersections, buildings, and traffic signs; these anchors are typically pre-stored in high-definition maps.

[0036] Subsequently, a comprehensive matching score is calculated by combining the vehicle's heading, visual features within the current field of view identified by the onboard camera, and the distance between the vehicle's estimated position and the center point of the semantic anchor. The vehicle's heading provides information about its direction of travel; visual features identified by the onboard camera, such as lane lines, traffic lights, and road signs, provide real-time perception information of the vehicle's surrounding environment; and the distance between the vehicle's estimated position and the center point of the semantic anchor reflects the spatial proximity between the vehicle's current position and potential semantic anchors. The comprehensive matching score quantifies the degree of matching between the vehicle's current state and each semantic anchor by weighted fusion of this information.

[0037] Semantic anchor points with a comprehensive matching score higher than a preset matching threshold are selected as the vehicle's initial logical location. The preset matching threshold is an empirical value used to filter out semantic anchor points with a high degree of matching with the vehicle's current state. The initial logical location is a preliminary, semantically meaningful inference about the vehicle's current location based on multi-source information fusion.

[0038] Based on this, the vehicle's short-term historical driving trajectory is reconstructed using wheel speed sensor data and steering angle sensor data. Wheel speed sensor data and steering angle sensor data have high real-time performance and accuracy, and through integration and other methods, the vehicle's movement path over a short period can be accurately reconstructed.

[0039] The reconstructed vehicle historical trajectory is compared with the road topology associated with the vehicle's initial logical location. Road topology refers to road connections, lane information, etc., which is typically stored in high-definition maps. This comparison verifies whether the vehicle's actual driving path matches the road structure of its initial logical location.

[0040] If the reconstructed vehicle historical trajectory is highly consistent with the road topology associated with the vehicle's initial logical location, the vehicle's semantic location is confirmed. High consistency indicates that the vehicle's actual trajectory matches the road structure inferred from semantic anchors, significantly improving the confidence level of localization. The semantic location is a highly reliable, specific position of the vehicle within the road network, determined after multiple verifications.

[0041] Based on the confirmed semantic location of the vehicle, navigation instructions are updated to correct erroneous instructions generated due to errors. Once the semantic location of the vehicle is confirmed, the system can use this highly reliable positioning information to recalculate and adjust the navigation path and instructions, thereby correcting previous erroneous guidance that may have been caused by positioning errors.

[0042] Finally, based on the updated navigation instructions, the vehicle icon is displayed, the driving route is highlighted, and corrective voice guidance is provided. This dual visual and auditory feedback ensures the driver clearly receives the corrected navigation information, thus preventing misoperation and improving driving safety.

[0043] Methods to improve the stability of real-time perception and risk identification processes include the following steps: The clarity and contrast of the images from the vehicle-mounted camera are monitored, as well as the signal-to-noise ratio and echo intensity of the vehicle-mounted millimeter-wave radar signal, to obtain sensor data quality assessment results. When the sensor data quality assessment results indicate a decline in sensor data quality or the presence of interference, the risk identification strategy is dynamically adjusted. The steps for dynamically adjusting the risk identification strategy include: In low-visibility environments, reduce the weight of visual recognition results from vehicle cameras and increase the weight of judging the distance, speed, and density of obstacles detected by vehicle millimeter-wave radar. Activate the auxiliary judgment module based on preset risk areas on high-precision maps; Increase the sensitivity of risk identification when vehicles enter high-visibility risk areas or areas with high electromagnetic interference; By comparing the current sensor data with historical sensor data of the road segment corresponding to the key decision point under similar time and weather conditions, data deviation information is obtained. When data deviation information indicates a deviation between current sensor data and historical sensor data, the current sensor data is marked as potential sensor interference or anomaly, and the confidence level of the risk assessment is adjusted.

[0044] Among these, the sharpness and contrast of images from vehicle-mounted cameras are crucial indicators of image quality. Sharpness reflects the richness of image details, while contrast reflects the difference between bright and dark areas; both directly affect the accuracy of visual recognition algorithms. The signal-to-noise ratio (SNR) and echo intensity of vehicle-mounted millimeter-wave radar signals are key parameters for evaluating radar performance. A high SNR indicates less noise interference, and a strong echo intensity indicates sufficient energy reflected from the target; these parameters directly affect the radar's ability to detect obstacles. Continuously monitoring these indicators allows for real-time understanding of the sensor's operating status and data reliability. Specifically, in low-visibility environments, such as fog, rain, snow, or nighttime conditions, the accuracy of visual recognition by vehicle-mounted cameras significantly decreases. In such situations, reducing the weight of the vehicle-mounted camera's visual recognition results means reducing reliance on visual information during risk assessment. Simultaneously, increasing the weight of the distance, speed, and density of obstacles detected by the vehicle-mounted millimeter-wave radar is because millimeter-wave radar has better penetration in adverse weather and low-visibility conditions, making its detection results relatively more reliable. This weighting adjustment effectively compensates for the shortcomings caused by the decline in camera performance.

[0045] Furthermore, high-precision maps typically include pre-defined high-risk areas such as sharp bends, accident-prone sections, and construction zones. When vehicles enter these areas, even if sensor data quality deteriorates, the auxiliary judgment module can be activated to combine pre-defined risk information for supplementary assessment, improving the robustness of risk identification. Additionally, high-visibility risk areas may refer to road sections with good visual conditions but high potential danger (such as accident-prone areas on highways), while areas with high electromagnetic interference may affect the normal operation of sensors such as radar. In these cases, increasing the sensitivity of risk identification means the system will be more vigilant, adjusting the threshold for potential risks; even weak abnormal signals may trigger risk warnings to ensure driving safety. In practical applications, by establishing a historical sensor data baseline, current data can be compared horizontally and vertically. For example, typical sensor data patterns for specific road sections under different conditions such as sunny days, rainy days, daytime, and nighttime can be stored. If there are significant differences between current and historical data, it may indicate that the sensor has been interfered with, malfunctioned, or that the environment has undergone abnormal changes. At this point, the system will reduce its trust in the sensor data and may combine it with other sensor data or auxiliary information to make a comprehensive judgment in order to avoid erroneous risk identification due to a single sensor malfunction.

[0046] In some preferred embodiments, this application is implemented as follows: Suppose a vehicle is traveling on a mountain road and suddenly encounters dense fog, causing a sharp decrease in the clarity and contrast of the onboard camera images. At this time, the system will detect that the onboard camera image quality assessment results indicate a decline in data quality. Simultaneously, the signal-to-noise ratio and echo intensity of the onboard millimeter-wave radar may remain normal or fluctuate slightly. Based on the sensor data quality assessment results, the system will dynamically adjust its risk identification strategy. Specifically, the system will reduce the weight of the onboard camera's visual recognition results, for example, adjusting the weight from 0.7 to 0.3, while increasing the weight of judging the distance, speed, and density of obstacles detected by the onboard millimeter-wave radar, for example, adjusting the weight from 0.3 to 0.7. This means that when determining whether there are physical obstacles or lane changes ahead, the system will rely more heavily on millimeter-wave radar data.

[0047] Furthermore, if the mountainous road section is pre-defined as a high-risk area with frequent accidents or low visibility in the high-precision map, the system will activate the auxiliary judgment module based on the pre-defined risk area in the high-precision map, and combine map information to assist in the judgment of potential risks. For example, even if the camera cannot clearly identify the guardrail at the bend ahead, but the high-precision map indicates that it is a sharp bend and there is a guardrail, the system will still combine radar data and map information to identify potential collision risks earlier.

[0048] Furthermore, the system continuously compares current sensor data with historical sensor data for the road segments corresponding to key decision points under conditions similar to heavy fog. If the system detects that the current millimeter-wave radar echo intensity is significantly lower than the historical average, it will flag this as potential sensor interference or anomaly and adjust the confidence level of the risk assessment accordingly. For example, it might reduce the overall risk assessment confidence level from 90% to 70%, potentially triggering a driver warning and advising the driver to be more vigilant. Through these dynamic adjustments, the system maintains high accuracy and stability in risk identification even in adverse weather conditions, providing drivers with reliable navigation and warning information.

[0049] This application further proposes that, in the event that a risk is determined to exist at a critical decision point, the method also includes the following steps: Risk levels are dynamically assessed and assigned based on the severity, duration, scope of impact of the conflict, and the density and speed of vehicles in key lanes. The severity of visual conflict is quantified based on the size and location of physical obstacles, their impact on critical lanes, the type of lane line changes, and the degree of change in traffic direction. The severity of traffic disruption is quantified based on the number of stationary or very low-speed vehicles in the critical lane, the proportion of vehicles occupying the critical lane, and the duration of the disruption. Assess the urgency of the risk based on the vehicle's current speed and distance from the critical decision point; By combining the risk level, the severity of visual conflict, the severity of traffic obstruction, the urgency of the risk, and the impact of the risk on different driving behaviors, warnings and suggestions with clear priorities are generated. The priority of warnings and suggestions can be distinguished by the graphic warnings on the navigation display screen and the changes in tone and speed of the voice broadcast.

[0050] Specifically, when a risk is identified at a key decision point, the method also includes the following steps: The severity of the conflict can refer to the potential consequences of the danger, such as whether it may lead to a collision or property damage; the duration refers to the expected length of time the risk event will last, such as whether it is a momentary event or a long-term one; the scope of impact refers to the area affected by the risk event on lanes, intersections, or the entire traffic flow; and the density and speed of vehicles in key lanes reflect the degree of congestion and mobility of traffic conditions. By comprehensively considering these factors, risks can be assessed in a multi-dimensional way and assigned corresponding risk levels, such as low, medium, and high risk levels.

[0051] The larger the physical obstacle, the closer it is to the vehicle's path, and the greater its impact on critical lanes (e.g., complete blockage or severe restriction of traffic), the more severe the visual conflict. The type of lane marking change, such as from dashed to solid lines, lane reduction, or lane merging, and the degree to which it alters the direction of travel, are also quantified to reflect their impact on driver decision-making and driving safety.

[0052] For example, when there are many stationary vehicles in a critical lane, occupying a large proportion of the lane and for a long period of time, the severity of traffic obstruction is judged as high, indicating that the lane is almost impossible to pass or has extremely low traffic efficiency.

[0053] In practical applications, the urgency of a risk is determined based on the vehicle's current speed and its distance from the critical decision point. For example, when a vehicle approaches a critical decision point at high speed and is relatively close, the driver needs to react immediately, so the urgency of the risk is judged as high; conversely, if the vehicle is traveling at a lower speed and is farther away, the urgency is relatively low, and the driver has more time to make a decision.

[0054] Therefore, by combining the risk level, the severity of visual conflict, the severity of traffic obstruction, the urgency of the risk, and the impact of the risk on different driving behaviors, the system generates warnings and suggestions with clear priorities. This means that the system will integrate all quantitative assessment results, not only informing the driver of the existence of a risk, but also providing specific driving suggestions with clear priorities based on the comprehensive risk assessment results, such as "immediately slow down and change lanes to the right" or "beware of roadwork ahead and maintain a safe distance."

[0055] This application further proposes steps for generating prioritized warnings and recommendations, including: Continuously monitor the vehicle's current driving status, including vehicle speed, acceleration, steering angle, and relative distance and speed to surrounding vehicles; Real-time assessment of the driver's cognitive load status, which includes the driver's eye movement trajectory, head posture, and interaction behavior with navigation information; When multiple alerts and recommendations with similar priorities are detected, identify whether there are potential conflicts among the multiple alerts and recommendations with similar priorities; When a potential conflict is identified, the priority of warnings and suggestions is dynamically adjusted based on the driver's cognitive load. If the driver's cognitive load is high, prioritize presenting suggestions that have the greatest impact on driving safety and are the simplest to operate; If the driver's cognitive load is low, a warning and suggestion regarding potential conflicts will be presented, along with explanatory information. Based on the vehicle's current driving status, predict the potential risks and benefits of implementing different suggestions, and select the suggestion with the lowest potential risk and the highest benefit as the final recommendation. The final recommended warnings and suggestions are presented in a multimodal manner, including graphic indicators on the navigation display and voice announcements.

[0056] Specifically, continuous monitoring of the vehicle's current driving status refers to acquiring the vehicle's dynamic information in real time through onboard sensors (e.g., vehicle bus data, millimeter-wave radar, cameras, etc.), including but not limited to the vehicle's speed, acceleration, steering angle, and relative distance and speed with other surrounding vehicles. This data provides the basis for the system to assess the current driving environment and predict the consequences of driving behaviors. The vehicle's current driving status can be understood as the kinematic and dynamic characteristics of the vehicle at a specific point in time and spatial location, with the aim of providing precise contextual information for subsequent risk assessment and recommendation generation.

[0057] Real-time assessment of a driver's cognitive load involves analyzing the driver's eye movements, head posture, and interaction with navigation information (e.g., whether they frequently check the navigation screen or input voice commands) through a driver monitoring system (e.g., facial and eye recognition using in-vehicle cameras). The driver's cognitive load reflects their current level of attention and information processing capacity, aiming to ensure that the system adapts to the driver's actual state when providing warnings and suggestions, avoiding information overload or omissions.

[0058] When the system detects multiple warnings and suggestions with similar priorities, it further identifies potential conflicts between these warnings and suggestions. A potential conflict refers to two or more warnings or suggestions that contradict each other in execution, are difficult to implement simultaneously, or whose simultaneous implementation may lead to undesirable consequences. For example, one suggestion is "change lanes to the left," while another suggests "change lanes to the right," or one suggestion is "slow down," while another suggests "speed up." The purpose of identifying potential conflicts is to avoid providing drivers with ambiguous or contradictory instructions, thereby reducing the difficulty of driver decision-making.

[0059] When a potential conflict is detected, the system dynamically adjusts the priority of warnings and suggestions based on the driver's cognitive load. Specifically, if the driver's cognitive load is high, such as at high speeds or in complex road conditions, the system prioritizes presenting suggestions that have the greatest impact on driving safety and are the simplest to implement. This is because under high cognitive load, the driver's ability to process complex information decreases, requiring the most direct and critical guidance to ensure safety. Conversely, if the driver's cognitive load is low, such as at low speeds or in simple road conditions, the system will present both warnings and suggestions regarding potential conflicts, along with explanatory information. This allows the driver to gain a more comprehensive understanding of the situation and make a choice based on their own judgment.

[0060] Furthermore, the system predicts the potential risks and benefits of implementing different suggestions based on the vehicle's current driving status. For example, for conflicting suggestions like "change lanes left" and "change lanes right," the system combines information such as current vehicle speed, surrounding traffic density, and whether there are obstacles in adjacent lanes to predict the potential collision risk, traffic efficiency improvement, or decrease associated with each suggestion. By quantifying risks and benefits, the system can select the suggestion with the lowest potential risk and highest benefit as the final recommendation, thus providing the driver with optimal decision support. The final recommended warnings and suggestions are presented in a multimodal manner, including graphical indicators on the navigation display and voice announcements, to enhance information delivery and improve the driver's perception efficiency.

[0061] In some preferred embodiments, suppose the vehicle is traveling on a multi-lane highway and is approaching a complex merging area. According to the aforementioned vehicle navigation service resource management method, the system may simultaneously generate two warnings and suggestions with similar priorities: one is "The lane ahead is narrowing, please change lanes to the right," and the other is "There is an accident ahead, it is recommended to detour via the left lane." At this time, the system will first continuously monitor the vehicle's current driving status, for example, a speed of 100 km / h, dense surrounding traffic, and a close distance to the vehicle ahead. Simultaneously, the system will assess the driver's cognitive load in real time. Through the driver monitoring system, it will detect that the driver's eye movements frequently scan the rearview mirror, their head posture appears slightly tense, and they are not interacting with the navigation information, indicating a high cognitive load for the driver.

[0062] When detecting warnings and suggestions with similar priorities, such as "change lanes to the right" and "detour from the left lane," the system recognizes a potential conflict between them, as performing both actions simultaneously in a short period could lead to confusion or danger. Given the driver's high cognitive load, the system dynamically adjusts the priority of warnings and suggestions. In this case, the system prioritizes the suggestion with the greatest impact on driving safety and the simplest operation. For example, based on prediction, the system determines that "changing lanes to the right" carries a higher risk of collision with vehicles on the right, while "detour from the left lane," although slightly more complex, more effectively avoids the accident area under current traffic conditions and has a lower overall risk. Therefore, the system selects "detour from the left lane" as the final recommendation. Finally, the system presents this recommendation in a multimodal manner, such as highlighting the left detour route on the navigation display and issuing a clear voice announcement: "Accident ahead, please immediately change lanes to the left and detour safely." This approach provides clear and safe guidance to the driver under high cognitive load, avoiding confusion and potential dangers caused by conflicting information.

[0063] This application further proposes steps for real-time assessment of a driver's cognitive load, including: Based on the vehicle's current driving status, identify whether the vehicle is traveling at high speed or in a complex traffic scenario; When the vehicle is traveling at high speed or in complex traffic scenarios, it can identify whether the driver's eye movement trajectory, head posture, and interaction with navigation information are related to the current driving task. If relevant, it is judged as a normal driving reaction by the driver; If they are irrelevant, it is judged that the driver has a high cognitive load; When the vehicle is not traveling at high speed or in complex traffic scenarios, the system identifies whether the driver's eye movement trajectory, head posture, and interaction with navigation information are related to the current driving task. If relevant, it is determined that the driver's cognitive load is low; If it is irrelevant, it is judged that the driver has a high cognitive load.

[0064] Specifically, the vehicle's current driving state can be understood as the vehicle's operating status at a specific point in time, encompassing information such as vehicle speed, acceleration, steering angle, road type (e.g., highway, urban road, rural road), and traffic flow density. Its purpose is to provide crucial contextual information for assessing driver cognitive load, ensuring the accuracy of the assessment results. High-speed driving or complex traffic scenarios specifically refer to situations where the vehicle speed exceeds a preset threshold (e.g., highway speed limits) or the surrounding traffic environment is complex (e.g., multiple lane merging, frequent lane changes, dense traffic lights at intersections, mixed pedestrian and non-motorized vehicle traffic). These scenarios typically require drivers to exert greater attention, thus naturally resulting in a higher cognitive load.

[0065] Furthermore, the driver's eye movement, head posture, and interaction with navigation information are key indicators for assessing the driver's cognitive load. Eye movement can be acquired using in-vehicle cameras combined with eye-tracking technology, including fixation point, fixation duration, and scan rate, to determine the driver's focus of attention. Head posture can be acquired using in-vehicle cameras combined with posture recognition algorithms, including head orientation and rotation amplitude, to determine the driver's observation direction and attention allocation. Interaction with navigation information includes whether the driver actively checks the navigation screen, inputs voice commands, or adjusts navigation settings; these behaviors reflect the driver's level of attention to the navigation system and their willingness to operate it. Relevance to the current driving task means that the driver's above-mentioned behaviors (eye movement, head posture, interaction) are directly related to the current vehicle movement, road conditions, traffic rules, and navigation guidance. Examples include checking the rearview mirror before changing lanes, checking the navigation screen at intersections, and adjusting direction according to voice guidance. Normal driving response means that the driver's behavior conforms to the reasonable expectations of the current driving situation, indicating that the driver can effectively handle the driving task and that the cognitive load is at a normal level. High cognitive load refers to a driver's attention being distracted by things unrelated to driving, or a failure to demonstrate corresponding focus in scenarios requiring high cognitive input, indicating that the driver's information processing capacity may be nearing saturation or overload. Low cognitive load means that the driver can easily handle the current driving task and has spare capacity to process other information, indicating sufficient cognitive resources.

[0066] As a specific implementation, suppose a vehicle is traveling at 100 km / h on a highway. The onboard camera detects that the driver's eye movements are consistently focused on the road ahead and the rearview mirror, with a stable head posture and no interaction unrelated to navigation information. In this high-speed driving scenario, the system recognizes that the driver's behavior is highly relevant to the current driving task, thus determining that the driver's cognitive load is normal. Further, suppose the vehicle enters a complex urban intersection with heavy traffic and frequent lane changes. In this case, the system detects that the driver's eye movements frequently scan the navigation screen to confirm lane information, and the head posture also turns to observe vehicles on both sides, occasionally interacting with the navigation system via voice commands. Since these behaviors are all related to the current complex driving task, the system determines that the driver's cognitive load is at a normal level, although the cognitive input is high. However, if, in the aforementioned complex intersection scenario, the system detects that the driver's eye movements deviate from the road ahead for an extended period, the head posture turns towards the passenger seat, and the driver is operating the in-vehicle entertainment system, the system will determine that the driver's behavior is unrelated to the current driving task, thus judging the driver's cognitive load to be high. For example, when a vehicle is traveling at a low speed on an open rural road, the system detects that the driver's eye movements occasionally glance at the navigation screen, and their head posture is relatively relaxed, but overall they are still focused on the road conditions ahead. Since the vehicle is not traveling at high speed or in complex traffic scenarios, and the driver's behavior is related to the driving task, the system will judge the driver's cognitive load to be low. However, if the driver looks down at their phone for an extended period of time, it will be judged as having a high cognitive load. In this way, the system can perform a refined and intelligent assessment of the driver's cognitive load based on different driving situations.

[0067] The steps to identify potential conflicts among multiple warnings and recommendations with similar priorities include: Continuously monitor the vehicle's current driving status; Based on the vehicle's current driving status, determine whether the vehicle is traveling at high speed or in a complex traffic scenario; When a vehicle is traveling at high speed or in a complex traffic scenario, multiple warnings and suggestions with similar priorities that occur frequently within a short period of time are timestamped and serialized to obtain the event serialization result. Based on the event serialization results, calculate the time interval between adjacent warnings and recommendations; When the time interval is less than a preset time threshold, multiple warnings and suggestions with similar priorities are identified as a potential conflict event group. For each potential conflict event group, analyze the types of driving behaviors involved in each warning and recommendation; When there are contradictory types of driving behavior in a potential conflict event group, it is determined that there is a potential conflict. Assign real-time processing priorities to identified potential conflicts, with real-time processing priorities exceeding those of regular warnings and recommendations; The conflict resolution process is triggered based on the real-time processing priority.

[0068] Continuous monitoring of the vehicle's current driving status refers to the system acquiring information such as the vehicle's speed, acceleration, steering angle, and relative distance and speed to surrounding vehicles in real time to comprehensively understand the vehicle's dynamic behavior. Determining whether a vehicle is traveling at high speed or in a complex traffic scenario can be achieved through preset speed thresholds, road type (e.g., highways, congested urban areas, complex intersections), or traffic flow density. For example, when the vehicle speed exceeds a certain threshold (e.g., 80 km / h) or the system detects that the vehicle is located in a multi-lane merging or diverging area, it can be determined to be traveling at high speed or in a complex traffic scenario.

[0069] When vehicles are traveling at high speeds or in complex traffic scenarios, multiple warnings and suggestions with similar priorities that occur frequently within a short period of time are timestamped and serialized into events, resulting in an event serialization result. Timestamping involves attaching precise time information to the generation or trigger time of each warning and suggestion. Event serialization arranges these timestamped warnings and suggestions according to their chronological order of occurrence, forming an ordered event stream. For example, if the system issues two warnings with similar priorities within one second: "Construction ahead, please change lanes to the left" and "Slow vehicle in the right lane, please keep a safe distance," these two warnings will be timestamped separately and recorded sequentially.

[0070] Based on the event serialization results, the time interval between adjacent warnings and suggestions is calculated. This time interval refers to the time difference between two immediately adjacent warnings or suggestions in the sequence. When the time interval is less than a preset time threshold, such as less than 2 seconds, these warnings and suggestions that occur densely within a short period of time and have similar priorities are identified as a potential conflict event group. This preset time threshold can be adjusted according to the urgency of the driving scenario and the system response speed.

[0071] For potential conflict event groups, the system further analyzes the driving behavior types involved in each warning and suggestion. Driving behavior types can include, but are not limited to, acceleration, deceleration, left turn, right turn, lane change, and lane keeping. For example, "Please change lanes to the left" involves the behavior of "changing lanes to the left," and "Please maintain a safe distance" involves the behavior of "decelerating" or "keeping the current lane." When there are contradictory driving behavior types in a potential conflict event group, such as one suggestion requiring "changing lanes to the left" while another requires "keeping the current lane," or one suggestion requiring "acceleration" while another requires "deceleration," then a potential conflict is identified.

[0072] Identified potential conflicts are assigned a real-time processing priority, which is higher than the processing priority of regular warnings and suggestions. This means that once a potential conflict is identified, the system will immediately elevate it to a higher processing level to ensure priority handling. Based on the real-time processing priority, the system will trigger a conflict resolution process, which may include reassessing all relevant information, making decisions in conjunction with the driver's cognitive load, or requesting driver confirmation.

[0073] Furthermore, for identified groups of potential conflict events, the system conducts in-depth analysis of the driving behavior types involved in each warning and suggestion. For example, one warning might suggest "changing lanes to the left," while another might suggest "keeping the current lane" or "changing lanes to the right." By identifying the contradictions between these behavior types, the system can accurately determine whether a genuine potential conflict exists. This behavior-type analysis-based approach moves conflict identification beyond mere prioritization; it delves into the logical conflict judgment of the driving operation itself, thereby improving the accuracy and reliability of conflict identification. Once a potential conflict is identified, the system assigns it a higher real-time processing priority than regular warnings and suggestions, ensuring that these critical conflicts are immediately addressed and the corresponding conflict resolution process is triggered, avoiding safety risks caused by untimely or improper conflict handling.

[0074] In some preferred embodiments, assuming a car is traveling on a highway at a speed of 100 km / h, the system determines that the vehicle is traveling at high speed. Within a short period, the navigation system issues two warnings and suggestions of similar priority: First warning (timestamp T1): "Accident ahead 200 meters, please change lanes to the left." Second warning (timestamp T2): "An emergency vehicle is approaching from the right lane. Please remain in your current lane." The system continuously monitors the vehicle's current driving status and confirms that the vehicle is in a high-speed driving scenario.

[0075] The system timestamps and serializes the two warnings to obtain the event sequence: [Warning 1 (T1), Warning 2 (T2)].

[0076] The system calculates the time interval between T2 and T1, assuming it is 1.5 seconds. This time interval is less than the preset time threshold (e.g., 2 seconds), so the two warnings are identified as a potential conflict event group.

[0077] For this event group, the system analyzes the types of driving behaviors involved: Warning 1 pertains to the driving behavior type of "changing lanes to the left".

[0078] Warning 2 pertains to the driving behavior type of "keeping the current lane".

[0079] The system determined that "changing lanes to the left" and "keeping the current lane" are contradictory driving behavior types, and therefore identified a potential conflict.

[0080] The system then assigns a real-time processing priority to the identified potential conflicts, which is higher than the processing priority of regular warnings and suggestions. Based on this real-time processing priority, the system triggers a conflict resolution process. For example, it may further combine the driver's cognitive load status and surrounding environmental perception data (such as the precise distance and speed of the emergency vehicle on the right detected by millimeter-wave radar) to determine the final suggestion, or prompt the driver to pay attention and wait for further instructions, thereby avoiding confusion or incorrect operation by the driver receiving conflicting instructions in a short period of time.

[0081] This application further proposes that, once a group of potential conflict events has been identified, the method also includes the following steps: Continuously monitor images from vehicle-mounted cameras, millimeter-wave radar, and ultrasonic sensor data to acquire ambient perception data. Based on the vehicle's current driving status, determine whether the vehicle is at a complex intersection or in an area with dense traffic flow; When vehicles are at complex intersections or areas with high traffic flow, for potential conflict event groups, the impact of the types of driving behaviors involved on the vehicle's future driving path and interactions with surrounding traffic participants is analyzed. By combining surrounding environmental perception data, the impact is assessed in multiple dimensions, including: the impact of driving behavior type on adjacent lanes, the impact on vehicles behind, and the impact on right-of-way at intersections. When the multi-dimensional assessment results indicate that there are interactions between different types of driving behaviors that cannot be performed simultaneously or that would trigger a chain of dangers after being performed, it is determined that there is a multi-dimensional, non-linear behavioral conflict.

[0082] Specifically, continuous monitoring of images from vehicle-mounted cameras, millimeter-wave radar, and ultrasonic sensor data aims to acquire detailed information about the vehicle's surrounding environment in real time, such as the location and speed of obstacles, lane markings, traffic signs, and the dynamics of other road users. This sensor data is fused and processed to build a comprehensive environmental perception data model. Vehicle-mounted camera images provide rich visual information, millimeter-wave radar accurately measures distance and speed, and ultrasonic sensors are suitable for detecting obstacles at close range.

[0083] Determining whether a vehicle is at a complex intersection or in a high-traffic area based on its current driving status is crucial for contextualizing conflict analysis. Complex intersections typically refer to areas with multiple lanes, multiple turning directions, complex traffic lights, or areas where pedestrians and non-motorized vehicles share the road. High-traffic areas refer to road sections with high vehicle density, frequent speed changes, and complex traffic interactions. These determinations can be based on high-precision map information, real-time traffic data, and the vehicle's own location and speed information.

[0084] When vehicles are in the aforementioned complex or densely populated areas, for the identified groups of potential conflict events, a thorough analysis is needed to determine the impact of the types of driving behaviors involved on the vehicle's future path and interactions with surrounding traffic participants. For example, one suggestion is to change lanes, while another is to slow down. At complex intersections, changing lanes may affect vehicles in adjacent lanes, while slowing down may affect the passage of vehicles behind.

[0085] By combining surrounding environmental perception data, a multi-dimensional assessment of the impact is conducted. This assessment considers not only the impact of driving behavior on the vehicle itself, but also extends to the impact on other road users. Specifically, it assesses the impact of different driving behavior types on adjacent lanes, such as whether lane changes will cause vehicles in adjacent lanes to brake suddenly or swerve to avoid them; it assesses the impact on vehicles behind, such as whether sudden braking will cause a rear-end collision risk; and it assesses the impact on right-of-way at intersections, such as whether a suggestion at an unsignalized intersection will infringe on the right-of-way of vehicles traveling in other directions.

[0086] A multi-dimensional, non-linear behavioral conflict exists when multi-dimensional assessments indicate interactions between different driving behavior types that cannot be performed simultaneously or whose execution would trigger a chain reaction of hazards. For example, when approaching a congested and complex intersection, a navigation system might simultaneously suggest "change lanes to the left" and "stay in your current lane and slow down." If the multi-dimensional assessment finds that changing lanes to the left carries a risk of collision with a vehicle approaching at high speed from the left, while staying in the current lane and slowing down carries a risk of rear-end collisions, and neither of these risks can be effectively avoided in the current environment, then this constitutes a multi-dimensional, non-linear behavioral conflict.

[0087] In some preferred embodiments, suppose a vehicle is driving at a busy, multi-lane city intersection, and the navigation system, based on coarse-grained path planning, identifies a critical decision point approximately 100 meters ahead. At this point, the system may simultaneously detect two warnings and suggestions of similar priority: 1. Warning A: There is construction ahead in the right lane. It is recommended to change lanes to the left.

[0088] 2. Warning B: The right-turn lane at the intersection ahead is congested. It is recommended to stay in your current lane and go straight.

[0089] Based on the aforementioned basic scheme, the system may identify a potential conflict between two driving behaviors: "changing lanes to the left" and "maintaining the current lane and continuing straight." However, the scheme in this application further refines this identification process.

[0090] First, the system continuously monitors images from the vehicle's camera, millimeter-wave radar, and ultrasonic sensor data to acquire real-time environmental perception data around the intersection, including the speed and distance of vehicles in the left lane, the degree of congestion in the right-turn lane, and the following distance of vehicles behind. Simultaneously, the system determines that the vehicle is currently at a complex intersection and in an area with dense traffic flow.

[0091] Next, for the two potentially conflicting driving behavior types of "changing lanes to the left" and "staying in the current lane and going straight", the system will analyze their impact on the vehicle's future driving path and the interaction with surrounding traffic participants.

[0092] For "changing lanes to the left": The system assesses its impact on the adjacent left lane. If a vehicle is approaching at high speed from the left lane, changing lanes may cause the left vehicle to brake suddenly or cause a side collision. Simultaneously, the impact on vehicles behind is assessed; if the lane change is too abrupt, it may disrupt the driving rhythm of vehicles behind.

[0093] For "Maintaining straight in the current lane": The system assesses its impact on vehicles behind. If going straight in the current lane would lead to severe congestion, vehicles may need to brake suddenly or stop for an extended period, potentially causing rear-end collisions. Simultaneously, the system assesses the impact on right-of-way at the intersection. If the straight lane has a traffic light at the intersection, while the right-turn lane has a green light, then going straight might obstruct the passage of right-turning vehicles.

[0094] Through multi-dimensional evaluation, the system found that: While changing lanes to the left can avoid construction, the current traffic flow in the left lane is dense and the vehicles are traveling at high speeds, making the lane change extremely risky and potentially causing a collision with vehicles on the left.

[0095] While "staying in the current lane and going straight" can avoid right-turn congestion, there is also severe congestion ahead in the straight lane, and vehicles behind are following closely, so sudden braking may cause a rear-end collision.

[0096] In this scenario, the multi-dimensional assessment results indicate an interaction between these two driving behavior types: they cannot be performed simultaneously (because both involve high risks) or their execution would trigger a chain reaction of dangers. Therefore, the system determines that a multi-dimensional, non-linear behavioral conflict exists. Based on this deeper conflict identification, the system can trigger a higher-level conflict resolution process. For example, it might suggest that the driver gradually slow down in the current lane and look for the next intersection to detour, or, if safe, attempt a more cautious lane change, while simultaneously providing the driver with a detailed explanation of the risks, thus avoiding simply choosing one of the high-risk options.

[0097] refer to Figure 2 This application proposes an in-vehicle navigation service resource management system, which is applied to an in-vehicle navigation service resource management method, and includes: The perception module senses the signal quality of the global navigation satellite system in the environment in which the vehicle navigation system is located, as well as the load status of the central processing unit inside the system. The adjustment module adaptively adjusts the computational complexity of the inertial measurement unit (IMU) data processing algorithm when the global navigation satellite system signal is poor and the internal central processing unit is saturated, thus obtaining the adjusted IMU data processing algorithm. The planning module, based on the adjusted inertial measurement unit data processing algorithm, performs path planning in stages, prioritizing the generation of coarse-grained paths; The allocation module allocates high-priority resources to voice guidance tasks and preloads voice data; The triggering module, combining vehicle speed and positioning module confidence, predictively triggers voice guidance based on coarse-grained path and using pre-loaded voice data; The processing module initiates a real-time perception and risk identification process for key decision points on the coarse-grained path. It uses onboard cameras to identify physical obstacles, lane changes, and traffic signs at these key decision points, generating visual perception results. It then assesses the traffic conditions at these key decision points using onboard millimeter-wave radar and ultrasonic sensors, obtaining a trafficability assessment result. This assessment includes whether there are a large number of stationary or extremely slow-moving vehicles in the key lanes. When the visual perception results conflict with the static map information used for coarse-grained path planning, or when there are a large number of stationary or extremely slow-moving vehicles in the key lanes, the module determines that the key decision point poses a risk. If a risk is identified, customized warning information is generated, including the location of the conflict point and suggested driving behavior. The module assigns the highest priority to the generation, display, and voice broadcast of warning information. The warning information is displayed graphically on the navigation screen, highlighting the adjacent lanes to which the driver should change lanes. Voice guidance synchronized with actual road conditions is issued via the voice broadcast module. Finally, the module dynamically calculates the triggering timing of the warning information based on the vehicle's current speed and distance from the key decision point.

[0098] Specifically, the perception module is configured to continuously monitor the external environment of the in-vehicle navigation system, such as the signal quality of the Global Navigation Satellite System, as well as the system's own operating status, such as the load status of the internal central processing unit. This module aims to provide real-time and accurate basic data for subsequent resource management decisions.

[0099] The adjustment module is designed to adaptively adjust the computational complexity of the inertial measurement unit (IMU) data processing algorithm under specific operating conditions, namely when the signal quality of the global navigation satellite system is poor and the load on the internal central processing unit is saturated. This dynamic adjustment ensures that IMU data can still be effectively processed even in resource-constrained environments, thereby generating the adjusted IMU data processing algorithm.

[0100] In practical applications, the planning module is responsible for performing phased path planning based on the adjusted inertial measurement unit (IMU) data processing algorithm. This module prioritizes generating coarse-grained paths to quickly provide navigation directions, and then gradually refines the path as needed.

[0101] Furthermore, the allocation module allocates high-priority system resources to voice guidance tasks and preloads the necessary voice data. Its purpose is to ensure the timeliness and fluency of voice guidance, thereby improving the user's navigation experience.

[0102] Furthermore, the triggering module combines the vehicle's current speed and the positioning module's confidence level, based on a pre-generated coarse-grained path, and utilizes pre-loaded voice data to predictively trigger voice guidance. This predictive triggering mechanism aims to provide guidance before the driver needs information, thereby improving the predictability and safety of navigation.

[0103] The content disclosed above is only a preferred and feasible embodiment of the present invention, and is not intended to limit the scope of protection of the present invention. Therefore, all equivalent technical changes made based on the content of the present invention specification and drawings are included within the scope of protection of the present invention. Furthermore, the elements therein can be updated as technology develops.

Claims

1. A method for managing vehicle navigation service resources, characterized in that, The method includes the following steps: The system senses the signal quality of the global navigation satellite system in the environment in which the vehicle navigation system is located, as well as the load status of the central processing unit inside the system. When the signal of the global navigation satellite system is poor and the central processing unit inside the system is saturated, the computational complexity of the inertial measurement unit data processing algorithm is adaptively adjusted to obtain the adjusted inertial measurement unit data processing algorithm. Based on the adjusted inertial measurement unit data processing algorithm, path planning is performed in stages, with priority given to generating coarse-grained paths; Allocate high-priority resources to voice guidance tasks and preload voice data; By combining vehicle speed and positioning module confidence, and based on coarse-grained path, voice guidance is predictively triggered using pre-loaded voice data. A real-time perception and risk identification process is initiated for key decision points on the coarse-grained path. The system uses onboard cameras to identify physical obstacles, lane changes, and traffic signs at these key decision points, generating visual perception results. Onboard millimeter-wave radar and ultrasonic sensors assess the traffic conditions at these key decision points, providing a trafficability assessment result. This assessment includes the presence of a large number of stationary or extremely slow-moving vehicles in the key lanes. If the visual perception results conflict with the static map information used for coarse-grained path planning, or if a large number of stationary or extremely slow-moving vehicles are present in the key lanes, a risk is identified at the key decision point. If a risk is identified, customized warning information is generated, including the location of the conflict point and suggested driving behavior. The generation, display, and voice broadcast of warning information are assigned the highest priority. The warning information is displayed graphically on the navigation screen, highlighting the adjacent lanes to which the driver should change lanes. Voice guidance synchronized with actual road conditions is issued via the voice broadcast module. The triggering timing of the warning information is dynamically calculated based on the vehicle's current speed and distance from the key decision point.

2. The in-vehicle navigation service resource management method as described in claim 1, characterized in that, The method also includes the following steps: The reliability of vehicle positioning information is evaluated, including assessing whether there is a persistent inconsistency between inertial measurement unit data and vehicle motion state, which includes wheel speed sensor data and steering angle sensor data acquired from the vehicle bus. When there is a persistent inconsistency between the inertial measurement unit data and the vehicle's motion state, the flexible path planning mode is activated. Delineate the search area centered on the estimated vehicle location, and query and extract predefined semantic anchor points; A comprehensive matching score is calculated by combining the vehicle's heading, visual features within the current field of view identified by the onboard camera, and the distance between the estimated vehicle position and the center point of the semantic anchor point. Select semantic anchors with a comprehensive matching score higher than a preset matching threshold as the initial logical location of the vehicle; Based on wheel speed sensor data and steering angle sensor data, the vehicle's short-term historical driving trajectory is reconstructed; The reconstructed vehicle history trajectory is compared with the road topology that associates the vehicle's initial logical location; If the reconstructed vehicle history trajectory is highly consistent with the road topology associated with the vehicle's initial logical location, the vehicle's semantic location is confirmed. Based on the confirmed semantic location of the vehicle, update the navigation instructions to correct erroneous instructions generated due to errors; Based on the updated navigation instructions, the vehicle icon is displayed, the driving route is highlighted, and voice guidance for correction is given.

3. The in-vehicle navigation service resource management method as described in claim 1, characterized in that, To improve the stability of the real-time perception and risk identification process, the method also includes the following steps: The clarity and contrast of the images from the vehicle-mounted camera are monitored, as well as the signal-to-noise ratio and echo intensity of the vehicle-mounted millimeter-wave radar signal, to obtain sensor data quality assessment results. When the sensor data quality assessment results indicate a decline in sensor data quality or the presence of interference, the risk identification strategy is dynamically adjusted. The steps for dynamically adjusting the risk identification strategy include: In low-visibility environments, reduce the weight of visual recognition results from vehicle cameras and increase the weight of judging the distance, speed, and density of obstacles detected by vehicle millimeter-wave radar. Activate the auxiliary judgment module based on preset risk areas on high-precision maps; Increase the sensitivity of risk identification when vehicles enter high-visibility risk areas or areas with high electromagnetic interference; By comparing the current sensor data with historical sensor data of the road segment corresponding to the key decision point under similar time and weather conditions, data deviation information is obtained. When data deviation information indicates a deviation between current sensor data and historical sensor data, the current sensor data is marked as potential sensor interference or anomaly, and the confidence level of the risk assessment is adjusted.

4. The in-vehicle navigation service resource management method as described in claim 1, characterized in that, If the method determines that there is a risk at the key decision point, it also includes the following steps: Risk levels are dynamically assessed and assigned based on the severity, duration, scope of impact of the conflict, and the density and speed of vehicles in key lanes. The severity of visual conflict is quantified based on the size and location of physical obstacles, their impact on critical lanes, the type of lane line changes, and the degree of change in traffic direction. The severity of traffic disruption is quantified based on the number of stationary or very low-speed vehicles in the critical lane, the proportion of vehicles occupying the critical lane, and the duration of the disruption. Assess the urgency of the risk based on the vehicle's current speed and distance from the critical decision point; By combining the risk level, the severity of visual conflict, the severity of traffic obstruction, the urgency of the risk, and the impact of the risk on different driving behaviors, warnings and suggestions with clear priorities are generated. The priority of warnings and suggestions can be distinguished by the graphic warnings on the navigation display screen and the changes in tone and speed of the voice broadcast.

5. The in-vehicle navigation service resource management method as described in claim 4, characterized in that, The steps for generating prioritized alerts and recommendations also include: Continuously monitor the vehicle's current driving status, including vehicle speed, acceleration, steering angle, and relative distance and speed to surrounding vehicles; Real-time assessment of the driver's cognitive load status, which includes the driver's eye movement trajectory, head posture, and interaction behavior with navigation information; When multiple alerts and recommendations with similar priorities are detected, identify whether there are potential conflicts among the multiple alerts and recommendations with similar priorities; When a potential conflict is identified, the priority of warnings and suggestions is dynamically adjusted based on the driver's cognitive load. If the driver's cognitive load is high, prioritize presenting suggestions that have the greatest impact on driving safety and are the simplest to operate; If the driver's cognitive load is low, a warning and suggestion regarding potential conflicts will be presented, along with explanatory information. Based on the vehicle's current driving status, predict the potential risks and benefits of implementing different suggestions, and select the suggestion with the lowest potential risk and the highest benefit as the final recommendation. The final recommended warnings and suggestions are presented in a multimodal manner, including graphic indicators on the navigation display and voice announcements.

6. The in-vehicle navigation service resource management method as described in claim 5, characterized in that, The steps for real-time assessment of a driver's cognitive load include: Based on the vehicle's current driving status, identify whether the vehicle is traveling at high speed or in a complex traffic scenario; When the vehicle is traveling at high speed or in complex traffic scenarios, it can identify whether the driver's eye movement trajectory, head posture, and interaction with navigation information are related to the current driving task. If relevant, it is judged as a normal driving reaction by the driver; If they are irrelevant, it is judged that the driver has a high cognitive load; When the vehicle is not traveling at high speed or in complex traffic scenarios, the system identifies whether the driver's eye movement trajectory, head posture, and interaction with navigation information are related to the current driving task. If relevant, it is determined that the driver's cognitive load is low; If it is irrelevant, it is judged that the driver has a high cognitive load.

7. The in-vehicle navigation service resource management method as described in claim 5, characterized in that, The steps to identify potential conflicts among multiple warnings and recommendations with similar priorities include: Continuously monitor the vehicle's current driving status; Based on the vehicle's current driving status, determine whether the vehicle is traveling at high speed or in a complex traffic scenario; When a vehicle is traveling at high speed or in a complex traffic scenario, multiple warnings and suggestions with similar priorities that occur frequently within a short period of time are timestamped and serialized to obtain the event serialization result. Based on the event serialization results, calculate the time interval between adjacent warnings and recommendations; When the time interval is less than a preset time threshold, multiple warnings and suggestions with similar priorities are identified as a potential conflict event group. For each potential conflict event group, analyze the types of driving behaviors involved in each warning and recommendation; When there are contradictory types of driving behavior in a potential conflict event group, it is determined that there is a potential conflict. Assign real-time processing priorities to identified potential conflicts, with real-time processing priorities exceeding those of regular warnings and recommendations; The conflict resolution process is triggered based on the real-time processing priority.

8. The in-vehicle navigation service resource management method as described in claim 7, characterized in that, Once a group of potential conflict events has been identified, the method further includes the following steps: Continuously monitor images from vehicle-mounted cameras, millimeter-wave radar, and ultrasonic sensor data to acquire ambient perception data. Based on the vehicle's current driving status, determine whether the vehicle is at a complex intersection or in an area with dense traffic flow; When vehicles are at complex intersections or areas with high traffic flow, for potential conflict event groups, the impact of the types of driving behaviors involved on the vehicle's future driving path and interactions with surrounding traffic participants is analyzed. By combining surrounding environmental perception data, the impact is assessed in multiple dimensions, including: the impact of driving behavior type on adjacent lanes, the impact on vehicles behind, and the impact on right-of-way at intersections. When the multi-dimensional assessment results indicate that there are interactions between different types of driving behaviors that cannot be performed simultaneously or that would trigger a chain of dangers after being performed, it is determined that there is a multi-dimensional, non-linear behavioral conflict.

9. A vehicle navigation service resource management system, applied to the vehicle navigation service resource management method as described in claim 1, characterized in that, The system includes: The perception module senses the signal quality of the global navigation satellite system in the environment in which the vehicle navigation system is located, as well as the load status of the central processing unit inside the system. The adjustment module adaptively adjusts the computational complexity of the inertial measurement unit (IMU) data processing algorithm when the global navigation satellite system signal is poor and the internal central processing unit is saturated, thus obtaining the adjusted IMU data processing algorithm. The planning module, based on the adjusted inertial measurement unit data processing algorithm, performs path planning in stages, prioritizing the generation of coarse-grained paths; The allocation module allocates high-priority resources to voice guidance tasks and preloads voice data; The triggering module, combining vehicle speed and positioning module confidence, predictively triggers voice guidance based on coarse-grained path and using pre-loaded voice data; The processing module initiates a real-time perception and risk identification process for key decision points on the coarse-grained path. It uses onboard cameras to identify physical obstacles, lane changes, and traffic signs at these key decision points, generating visual perception results. It then assesses the traffic conditions at these key decision points using onboard millimeter-wave radar and ultrasonic sensors, obtaining a trafficability assessment result. This assessment includes whether there are a large number of stationary or extremely slow-moving vehicles in the key lanes. When the visual perception results conflict with the static map information used for coarse-grained path planning, or when there are a large number of stationary or extremely slow-moving vehicles in the key lanes, the module determines that the key decision point poses a risk. If a risk is identified, customized warning information is generated, including the location of the conflict point and suggested driving behavior. The module assigns the highest priority to the generation, display, and voice broadcast of warning information. The warning information is displayed graphically on the navigation screen, highlighting the adjacent lanes to which the driver should change lanes. Voice guidance synchronized with actual road conditions is issued via the voice broadcast module. Finally, the module dynamically calculates the triggering timing of the warning information based on the vehicle's current speed and distance from the key decision point.

Images