Vehicle collision early warning system and method based on multi-sensor fusion

Abstract

The invention discloses a vehicle collision early warning system and method based on multi-sensor fusion, and relates to the technical field of electric vehicles. According to the vehicle collision early warning system based on multi-sensor fusion, firstly, storage battery data and a power supply loop fluctuation state are collected through a storage battery and an electrical safety verification module, and electrical safety levels are divided, so that the influence degree of the power supply state on a multi-source sensor is comprehensively evaluated; then, through a collision early warning grade division module, target priority ranking and collision risk grade division are dynamically optimized according to different electrical safety grades, it is ensured that high-risk targets are processed preferentially, and priority ranking deviation caused by the power supply problem is effectively avoided; and finally, through a time drift correction and degradation operation verification module, carrying out drift correction on a sensor acquisition timestamp, and verifying the effectiveness of a simplified fusion strategy when part of sensor data is unreliable, thereby effectively improving the reliability of the collision early warning system.

Description

Technical Field

[0001] This invention relates to the field of electric vehicle technology, and in particular to a vehicle collision warning system and method based on multi-sensor fusion. Background Technology

[0002] As vehicles become increasingly electrified, batteries serve as the core energy source, powering the sensors, electronic control units (ECUs), and warning devices in collision warning systems. If a battery becomes depleted, short-circuited, or experiences unstable power supply, it can cause sensor data interruptions, processing delays, or even warning failures, significantly increasing the risk of collisions. For example, at low temperatures, the capacity of traditional lead-acid batteries drops sharply. Without an intelligent power management system, this could cause the radar to suddenly lose power while driving at high speeds, missing the opportunity to predict a collision.

[0003] This problem is even more pronounced in electric vehicles. Their high-voltage (300V or higher) battery packs provide the driving energy, while low-voltage auxiliary batteries (such as 12V) continuously power the sensors and control units of the collision warning system. High-voltage batteries are more sensitive to temperature; below -20℃, the usable capacity of lithium-ion batteries can drop to less than 50% of their rated value. If the battery management system fails to properly regulate the distribution of low-voltage power, voltage fluctuations can cause ranging errors in the millimeter-wave radar's radio frequency module, leading to misjudgments of the distance to vehicles ahead and creating potential collision hazards.

[0004] To address the aforementioned issues, existing technologies achieve fusion-based early warning by optimizing power management and multi-sensor collaborative mechanisms. A Battery Management System (BMS) dynamically adjusts the output of high and low voltage batteries, combined with a dual-power supply redundancy design to ensure stable operation of sensors and the ECU. Simultaneously, complementary data from multiple sensors, including millimeter-wave radar, cameras, and lidar, is utilized. Kalman filtering and deep learning algorithms are used to fuse the data, eliminating abnormal signals caused by voltage fluctuations and accurately identifying obstacle distance and speed. Time-series analysis is then combined to predict collision risks, ultimately providing timely alerts through early warning devices, achieving reliable collision warnings.

[0005] For example, Chinese invention patent CN118082819A discloses a vehicle collision warning method based on multi-sensor fusion, which includes: obtaining a target data set by preprocessing raw data collected by millimeter-wave radar; obtaining image data by acquiring image data by a camera and enhancing it in the image preprocessing stage based on an image enhancement algorithm to obtain an image data set; performing data fusion processing based on the obtained target data set and image data set; and making a decision based on a collision warning control strategy when an obstacle target is detected in real time, assisting the driver in identifying the danger ahead in real time through appropriate means.

[0006] In existing technologies, when a battery experiences a sudden voltage drop or abnormal temperature, fluctuations in the power supply circuit directly increase the ranging error of millimeter-wave radar for moving obstacles (such as pedestrians and other vehicles), delay the image processing of cameras for non-moving obstacles (such as guardrails and roadblocks), and distort the point clouds of both types of obstacles for lidar. Furthermore, the lack of a time-stamp drift correction mechanism for power supply fluctuations reduces the accuracy of multi-source sensor acquisition, further exacerbating abnormalities in battery voltage and temperature data and power supply fluctuations. This leads to a bias in the system's prioritization of moving and non-moving obstacles, an over-reliance on single-sensor data for risk level assessment, and a tendency to issue delayed warnings for moving obstacles and misjudge non-moving obstacles during voltage fluctuations, ultimately affecting the reliability of collision warnings for electric vehicles during operation. Summary of the Invention

[0007] To address the technical problems in the prior art, embodiments of the present invention provide a vehicle collision warning system and method based on multi-sensor fusion. The technical solution is as follows:

[0008] On one hand, a vehicle collision warning system based on multi-sensor fusion is provided. This system includes: a battery and electrical safety verification module, a collision warning level classification module, and a time drift correction and degraded operation verification module. The battery and electrical safety verification module is used to determine the electrical safety status and classify the electrical safety level by collecting battery data and combining it with the power supply circuit fluctuation status of multi-source sensors. The battery data includes the voltage, current, temperature, and SOC value of the high-voltage power battery collected in real time by the BMS, as well as the circuit voltage data of the low-voltage auxiliary battery. The circuit voltage data includes the low-voltage circuit voltage value and voltage fluctuation peak value. The power supply circuit fluctuation status is used to reflect the voltage stability and electrical safety of the multi-source sensor power supply circuit. The system includes a flow fluctuation amplitude assessment module to evaluate its impact on the accuracy of multi-source sensor acquisition; a collision warning level classification module to prioritize targets based on marked electrical safety levels, and a collision risk level classification module to determine the monitoring target types for electric vehicles in different driving scenarios, ensuring that high-collision-risk targets are processed first; and a time drift correction and degraded operation verification module to estimate and correct the drift of the multi-source sensor acquisition timestamps based on the collision risk level classification results, thereby calibrating the time synchronization error of the multi-source sensor data and determining whether to perform degraded operation verification. Degraded operation verification is used to assess the accuracy and timeliness of warnings when collision risk levels change in different driving scenarios.

[0009] On the other hand, a vehicle collision warning method based on multi-sensor fusion is provided. This method is applied to a vehicle collision warning system based on multi-sensor fusion. The method includes: Step 1, determining the electrical safety status and classifying the electrical safety level by collecting battery data and combining it with the power supply circuit fluctuation status of multi-source sensors; Step 2, classifying the target priority by marking the electrical safety level, and simultaneously classifying the collision risk level; Step 3, estimating and correcting the drift of the timestamps collected by the multi-source sensors based on the collision risk level classification results to calibrate the time synchronization error of the multi-source sensor data, so as to determine whether to perform degraded operation verification.

[0010] The beneficial effects of the technical solutions provided in the embodiments of the present invention include at least the following:

[0011] 1. This invention collects data such as voltage, current, and temperature of the high-voltage power battery and the fluctuation status of the low-voltage circuit through a battery and electrical safety verification module, classifying electrical safety levels. This solves the problem of implicit decrease in sensor accuracy caused by neglecting power supply status assessment in existing technologies, avoids increased ranging errors of millimeter-wave radar and processing delays of cameras, and provides reliable power supply status information for subsequent warnings. The collision warning level classification module dynamically optimizes target priority and collision risk level. By clarifying the priority of monitoring types such as targets directly ahead in the same lane, it ensures that high-risk targets are processed first, reducing warning delays or misjudgments caused by power supply fluctuations and improving the accuracy of risk assessment. The time drift correction and degraded operation verification module addresses the lack of a timestamp drift dynamic correction mechanism in existing technologies. By calibrating sensor time synchronization errors, it solves the problem of multi-source data asynchrony and verifies the degraded operation effect. Even when some sensors are unreliable due to abnormal power supply, it can still ensure the accuracy and timeliness of warnings through a simplified fusion strategy, avoiding warning failures caused by data asynchrony or sensor failure in existing technologies, and significantly improving system reliability.

[0012] 2. By integrating multiple data points, including high-voltage battery voltage, temperature, SOC value, and low-voltage circuit fluctuations, the system first determines whether all data are within safe limits to ensure a stable power supply. Then, by calculating and averaging the deviations of each data point from the safe limits, a quantified electrical state safety score is obtained, enabling a refined assessment of the power supply status. Simultaneously, based on the score, three safety levels are defined from high to low, accurately distinguishing the different degrees of impact of the power supply status on collision warnings, thus solving the problem of risk misjudgment caused by coarse assessments in existing technologies. This multi-parameter fusion-based quantitative grading method can identify the potential impact of power supply fluctuations on sensor accuracy in advance, providing a more reliable electrical safety basis for subsequent collision warnings and reducing warning deviations caused by insufficient power supply assessment.

[0013] 3. The system first triggers an assessment based on relative distance, introducing parameters such as relative acceleration and trajectory for dynamic obstacles to achieve refined classification. Simultaneously, a first-priority assessment sets a monitoring benchmark for long-distance targets, while a second-priority assessment sets a dynamic grading standard for short-distance targets, forming a progressive logic from initial sorting to precise grading. This solves the problems of ambiguous priority division and over-reliance on a single parameter in existing technologies. Furthermore, combined with target priority response verification and dynamic optimization of sorting based on power supply status, it ensures that even when sensors are affected by power fluctuations, they can still prioritize identifying high-risk objects such as accelerating targets directly ahead and vehicles cutting in from adjacent lanes. This reduces warning delays or misjudgments caused by priority misalignment, improving the targeting and reliability of collision warnings.

[0014] 4. By monitoring the duty cycle of the signal waveform, the relative distance monitoring amplitude is dynamically adjusted. Specifically, when the duty cycle is abnormal, the monitoring distance is either shortened to improve data reliability, or the system focuses on monitoring directly in front, resolving the risks associated with delayed judgment of distant targets and circuit overheating. Simultaneously, power supply response deviation assessment and transient suppression verification are introduced. This addresses dynamic voltage differences, suppresses transient interference, and verifies the recovery effect. This compensates for the lack of a power supply transient processing mechanism in existing technologies, preventing sensor data distortion due to voltage fluctuations, ensuring accurate collision risk classification even in cases of power supply anomalies, reducing false alarms or delays, and significantly improving the system's stability and reliability in complex power supply environments.

[0015] 5. For static obstacles, the ratio of driving speed to distance is calculated to quantify the urgency of a collision; for dynamic obstacles, the ratio of relative acceleration to distance is calculated to quantify the rate of risk change, enabling differentiated assessment based on obstacle type. Simultaneously, a tiered response mechanism is established based on different thresholds corresponding to the assessment values, forming a precisely matched risk-response chain from emergency braking to early warning prompts to minor alerts. This solves the problems of ambiguous risk levels and over- or under-response issues in existing technologies, ensuring that even in complex scenarios such as power fluctuations, adaptive measures can still be dynamically triggered based on real-time risks, reducing misoperations and missed responses, and significantly improving the accuracy and safety of collision warnings.

[0016] 6. A reference timeline is established using a high-precision vehicle-mounted clock, unifying the time reference standard and eliminating time inconsistency caused by sensor clock deviations at its root. For timestamps of different risk levels, a sliding window method is used to calculate the drift of the first collision risk timestamp, and a linear interpolation method is used to estimate the drift trend of low-priority timestamps, achieving differentiated and precise correction. Simultaneously, synchronization error and drift trend deviation rate are used as dual indicators to determine whether to trigger degraded operation verification, ensuring the time alignment accuracy of high-priority target data and evaluating the time stability of low-priority data. This solves the problems of insufficient time synchronization accuracy and lack of dynamic verification mechanisms in existing technologies. Even when sensors are affected by power supply fluctuations, data time consistency can still be guaranteed, reducing warning failures caused by time deviations and improving the reliability of the collision warning system under complex operating conditions. Attached Figure Description

[0017] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A schematic diagram of the structure of a vehicle collision warning system based on multi-sensor fusion provided in an embodiment of the present invention;

[0019] Figure 2 A flowchart for time drift correction and degradation operation verification provided in this embodiment of the invention;

[0020] Figure 3 A flowchart illustrating a vehicle collision warning method based on multi-sensor fusion provided in an embodiment of the present invention;

[0021] Figure 4 This is the homepage interface of the vehicle collision warning and monitoring platform provided in this embodiment of the invention;

[0022] Figure 5 This is a battery and electrical safety verification page for a vehicle collision warning monitoring platform provided in an embodiment of the present invention. Detailed Implementation

[0023] The technical solution of the present invention will now be described with reference to the accompanying drawings.

[0024] In embodiments of the present invention, words such as "exemplarily," "for example," etc., are used to indicate that something is an example, illustration, or description. Any embodiment or design described as "exemplary" in the present invention should not be construed as being more preferred or advantageous than other embodiments or designs. Specifically, the use of the word "exemplary" is intended to present the concept in a concrete manner. Furthermore, in embodiments of the present invention, the meaning expressed by "and / or" can be both, or either one.

[0025] To make the technical problems, technical solutions and advantages of the present invention clearer, a detailed description will be given below in conjunction with the accompanying drawings and specific embodiments.

[0026] This invention provides a vehicle collision warning system based on multi-sensor fusion, which is a key support for collision risk warning in electric vehicle simulation testing and can accurately address various risks in testing. Figure 1 The diagram shown illustrates the structure of a vehicle collision warning system based on multi-sensor fusion. The system's processing flow can include: a battery and electrical safety verification module, a collision warning level classification module, and a time drift correction and degraded operation verification module, which work together to ensure accurate and efficient warnings.

[0027] The battery and electrical safety verification module is used to determine the electrical safety status and classify the electrical safety level by collecting battery data and combining it with the power supply circuit fluctuation status of multi-source sensors. The battery data includes the voltage, current, temperature, and SOC value of the high-voltage power battery collected in real time by the BMS, as well as the circuit voltage data of the low-voltage auxiliary battery. The circuit voltage data includes the low-voltage circuit voltage value and voltage fluctuation peak. The power supply circuit fluctuation status is used to reflect the voltage stability and current fluctuation amplitude of the power supply circuit of the multi-source sensors to assess the impact on the acquisition accuracy of the multi-source sensors. The SOC value refers to the state of charge of the battery, which is a parameter that measures the percentage of the battery's current remaining charge relative to its rated capacity. It is used to intuitively reflect the battery's remaining power supply capacity, and its value range is usually 0%-100% (0% represents fully discharged, 100% represents fully charged).

[0028] The collision warning level classification module is used to classify the priority of targets by marking the electrical safety level, and to classify the collision risk level based on the result of the target priority classification. The target priority classification is used to determine the monitoring target type of electric vehicle in different driving scenarios to ensure that high collision risk targets are dealt with first. The monitoring target types include targets in the current lane directly in front, adjacent lane cut-in prediction targets, and static fault targets.

[0029] The time drift correction and degraded operation verification module is used to estimate and correct the drift of the timestamps collected by multi-source sensors based on the collision risk level classification results, so as to calibrate the time synchronization error of the multi-source sensor data and determine whether to perform degraded operation verification. Degraded operation verification is used to verify the accuracy of the warning and the timeliness of the response when the collision risk level changes under different driving scenarios, when the synchronization accuracy of multi-source sensor data is insufficient or the data of some sensors is unreliable due to abnormal power supply.

[0030] In a specific embodiment, for example, on urban roads, the battery and electrical safety verification module continuously monitors the power supply status to prevent sensors from misjudging pedestrians crossing the road due to voltage fluctuations, ensuring the accuracy of warnings under complex road conditions. On highways, the collision warning level classification module prioritizes locking onto vehicles braking suddenly in front, while the time drift correction function calibrates the time synchronization error of multiple sensors to ensure that when a vehicle is traveling at high speed, it can promptly identify targets suddenly cutting in from adjacent lanes and issue warnings, giving the driver reaction time. On rural roads, if some sensors have unreliable data due to abnormal power supply, the time drift correction and degraded operation verification module will initiate degraded verification to ensure accurate judgment of level changes for targets such as statically faulty vehicles on the roadside, avoiding the problem of warning failure when sensors malfunction in existing technologies, and comprehensively improving driving safety in different scenarios.

[0031] In addition to the aforementioned applicable scenarios, this system is also suitable for tunnels and nighttime driving. The battery module ensures stable power supply to the sensors in low-light environments, the collision warning module accurately identifies stationary vehicles in tunnels, and the time correction and degradation module ensures data synchronization, improving the reliability of warnings in complex environments and enhancing driving safety. These three modules work in a progressive and collaborative manner: the battery module provides a solid foundation for power supply security, offering reliable data for the latter two; the collision warning module dynamically classifies vehicles according to electrical levels; and the time correction module calibrates synchronization errors, ensuring accurate warnings. Together, they enhance system efficiency, effectively reducing collision risks in various driving scenarios and enabling accurate warnings even under complex conditions. The three modules work together to form a closed-loop protection system, comprehensively improving the vehicle's active safety capabilities.

[0032] Furthermore, the determination of electrical safety status and the classification of electrical safety levels are carried out as follows: it is determined whether the acquired battery data are all within the safety limits, which include the high-voltage battery voltage range, the temperature safety range, the minimum SOC limit, the low-voltage circuit voltage standard value, and the upper limit of voltage fluctuation peak value; if so, it indicates that the power supply status of the battery and multi-source sensors is stable and the electrical safety level is classified; otherwise, an electrical abnormality prompt is sent and temporary power supply adjustment is initiated.

[0033] The electrical safety level classification is as follows: The deviation scores of each battery data point from its corresponding safety limit are calculated and harmonicly averaged to obtain an electrical state safety score, which quantifies the safety and stability of the battery and multi-source sensor power supply status. When the electrical state safety score is greater than or equal to the set upper limit, the corresponding electrical safety level is determined to be a safety level where the battery and multi-source sensor power supply status has no impact on the accuracy of the collision warning. When the electrical state safety score is not less than the reference electrical state safety score and not greater than the set upper limit, the corresponding electrical safety level is classified as a safety level that may have an impact. When the electrical state safety score is less than the reference electrical state safety score and not less than the set lower limit, the corresponding electrical safety level is classified as a safety level that may cause deviations in the collision warning. When the electrical state safety score is less than the set lower limit, an electrical safety warning is issued.

[0034] Specifically, if the electrical safety score is not less than the set upper limit of the electrical safety score (usually set to 90), it indicates that the power supply status of the battery and multi-source sensors has no impact on the accuracy of the collision warning, and it is judged as Level 1 safety. This is usually manifested as: the high-voltage battery voltage is within ±5% of the midpoint of the range, the temperature is in the optimal range of 25-35℃, the SOC value is ≥80%, the peak voltage fluctuation of the low-voltage circuit is ≤2% of the rated value, and the current fluctuation amplitude is ≤1%.

[0035] If the electrical condition safety score is not less than the reference electrical condition safety score (usually set to 70) and is less than the set upper limit of the electrical condition safety score, it indicates that the power supply status of the battery and multi-source sensors may affect the accuracy of the collision warning, and is judged as Level 2 safety. This is usually manifested as: high voltage battery voltage deviating from the midpoint of the range by 6%-10%, temperature between -10-24℃ or 36-50℃, SOC value of 60%-79%, low voltage circuit voltage fluctuation peak of 3%-5% of the rated value, and current fluctuation amplitude of 2%-3%.

[0036] If the electrical safety score is not less than the set lower limit of the electrical safety score (usually set to 60) and is less than the reference electrical safety score, it indicates that the power supply status of the battery and multi-source sensors may deviate from the collision warning and is judged as Level 3 safety. This is usually manifested as: high voltage battery voltage deviating from the midpoint of the range by 11%-15%, temperature between -20-11℃ or 51-60℃, SOC value of 50%-59%, low voltage circuit voltage fluctuation peak of 6%-8% of the rated value, and current fluctuation amplitude of 4%-5%.

[0037] Among them, Level 1 safety, Level 2 safety, and Level 3 safety reflect the gradual change in the power supply status of the battery and multi-source sensors from stable and reliable to increasingly volatile, and the degree of impact on the accuracy of collision warning also increases accordingly.

[0038] It's important to understand that the aforementioned Level 1, Level 2, and Level 3 safety classifications are merely a basic framework, not the only grading method. In practical applications, safety levels can be further refined based on specific scenario requirements and system accuracy needs. For example, a "Level 1.5 safety" can be added between Level 1 and Level 2, corresponding to an electrical state safety score of 80-89 points. At this level, the high-voltage battery voltage deviates from the midpoint of the range by 5%-6%, and the temperature is close to the optimal range boundary (e.g., 24℃ or 36℃), to more accurately reflect subtle changes in the power supply status. Alternatively, a "Level 4 safety" can be added below Level 3 safety, specifically for situations with a score below 60 points, triggering forced power reduction or a stop warning in extreme conditions such as a high-voltage battery voltage deviation exceeding 15% or a temperature exceeding -20℃ or 60℃. This flexible grading adjustment can better adapt to different vehicle models, road conditions, and sensor performance, improving the adaptability and reliability of the collision warning system.

[0039] In this embodiment, the deviation fraction specifically includes: high-voltage power battery voltage deviation fraction, which is the ratio of the actual voltage to the median of the high-voltage battery voltage range; high-voltage power battery temperature deviation fraction, which is the ratio of the actual temperature to the median of the safe temperature range; SOC value deviation fraction, which is the ratio of the actual SOC value to the minimum SOC limit; low-voltage circuit voltage deviation fraction, which is the ratio of the actual low-voltage circuit voltage value to the standard low-voltage circuit voltage value; and voltage fluctuation peak deviation fraction, which is the ratio of the actual voltage fluctuation peak value to the upper limit of the voltage fluctuation peak value.

[0040] Harmonic averaging balances the influence weights of various battery data points, preventing any single indicator from dominating the result and accurately quantifying the power supply safety status. In practical applications, specific values ​​such as safety limits and reference electrical condition safety scores can be adaptively adjusted based on vehicle type (e.g., passenger cars vs. trucks), sensor configuration (e.g., lidar vs. ultrasonic radar), and regional climate differences.

[0041] The innovation of this example lies not in the data value setting itself, but in the borrowing and averaging of power supply safety. By combining hierarchical division and flexible adaptation, a closed-loop mechanism is constructed from the power supply head to the early warning response, which improves the system's adaptability to complex working conditions and the accuracy of safety protection. This effectively enhances the system's stability and accuracy under complex working conditions and strengthens the vehicle's active safety protection capabilities.

[0042] Furthermore, the target priority classification process is as follows: when the relative distance between the electric vehicle and the target obstacle is less than the set safe distance, the relative distance between the electric vehicle and the target obstacle is obtained; otherwise, it indicates that there is no risk of collision between the electric vehicle and the target obstacle. The relative distance represents the driving path distance between the current position of the electric vehicle and the position of the target obstacle. This distance is not simply a straight-line distance, but is calculated based on the vehicle's current driving trajectory (such as lane lines and planned paths), representing the shortest projected distance along the vehicle's direction of travel and the obstacle on the path. This more closely reflects the real-world scenarios where vehicles and obstacles may come into contact during actual driving, and can more accurately reflect the collision risk; if the target... If the obstacle is static, the vehicle's speed is obtained. If the target obstacle is static, the vehicle's speed is obtained. If the target obstacle is dynamic, the relative acceleration between the vehicle and the target obstacle is obtained. At the same time, the trajectory of the dynamic obstacle (such as whether it tends to deviate from the lane, whether its direction of movement intersects with the vehicle's direction of travel) is combined to perform priority evaluation. Priority evaluation includes first priority evaluation and second priority evaluation. Relative acceleration is used to reflect the degree of speed change between the vehicle and the static obstacle as they approach or move away from each other. After priority evaluation, target priority response verification is also included to dynamically verify and optimize the priority evaluation results.

[0043] The first priority assessment and the second priority assessment form a dynamic and progressive priority change relationship, transitioning from high collision risk warning to tiered response execution priority. That is, the first priority focuses on determining the urgency of the collision risk and clarifies the priority monitoring right of high-risk targets; the second priority is based on the risk level to convert into the execution priority of specific response measures, ensuring the orderly connection from risk identification to response actions such as warning and braking, and achieving a precise match between the degree of risk and the intensity of response.

[0044] In this embodiment, the set safe distance is a safe braking distance calculated based on parameters such as the current driving speed of the electric vehicle and the road surface adhesion coefficient, ensuring that the vehicle has sufficient reaction and braking space. This example filters potential risk targets by using relative distance, avoiding meaningless assessments. Static obstacles are assessed in conjunction with driving path evaluation, while dynamic obstacles are assessed in conjunction with relative acceleration and trajectory trend evaluation, achieving differentiated priority determination for static and dynamic targets, ensuring that high-risk targets receive priority responses. Furthermore, dynamic optimization of priority response verification can accurately identify key risks such as vehicles braking suddenly in the same lane and targets cutting in from adjacent lanes, reducing false positives and false negatives, improving the targeting and timeliness of collision warnings, giving drivers valuable reaction time, and reducing the probability of collisions.

[0045] Furthermore, the priority assessment process is as follows: First, an initial priority ranking of targets corresponding to each electrical safety level is obtained. This initial ranking represents a preliminary arrangement of various target obstacles based on their collision risk under different electrical safety conditions. The preliminary ranking includes setting a first priority assessment and a second priority assessment. The first priority assessment includes setting the priority of the first target, the second target, and the third target, with the collision risk decreasing sequentially to form a hierarchical monitoring system for distant targets, ensuring that high-risk targets directly ahead are prioritized for monitoring. The second priority assessment includes setting the priority of the first sub-target, the second sub-target, and the third sub-target, with the priority increasing progressively based on the urgency of close-range dynamic risks, focusing on the accurate identification and response to sudden dangerous targets. The setting of the first and second priority assessments achieves a progression from preliminary ranking to precise grading, improving the accuracy of priority assessment.

[0046] Specifically, the first target priority setting represents a target in the lane directly ahead with a relative distance of not less than the first relative distance upper limit (usually set to 100m) and a driving speed greater than 0, in order to identify high-collision-risk targets approaching from a distance in advance and allow sufficient reaction time for evasive maneuvers; the second target priority setting represents a predicted target cutting into the adjacent lane with a relative distance of not less than the second relative distance upper limit (usually set to 80m), in order to reduce the risk of lateral collisions; the third target priority setting represents a static fault target with a relative distance of not less than the third relative distance upper limit (usually set to 50m), in order to improve the timeliness of early warning of collision risks with fixed road obstacles and ensure sufficient distance for lane changes or deceleration to avoid collisions;

[0047] The first sub-target priority setting indicates that dynamic targets in the lane directly ahead with a relative distance less than the first relative distance upper limit and a relative acceleration greater than 0 will be prioritized for monitoring high-risk targets that are accelerating close to the vehicle, thus triggering an emergency avoidance warning in a timely manner. The second sub-target priority setting indicates that adjacent lane cutting-in predicted targets with a relative distance less than the second relative distance upper limit and a cutting-in angle greater than the reference cutting-in angle (e.g., 10°) will be prioritized to enhance the monitoring of adjacent lane vehicles with obvious cutting-in trends and improve the response speed to side collision risks. The third sub-target priority setting indicates that dynamic interference targets with a relative distance less than the third relative distance upper limit and a trajectory intersection probability greater than the reference trajectory intersection probability (e.g., 60%) will be prioritized to handle sudden moving targets that may intersect with the vehicle's trajectory, thereby reducing the risk of collision.

[0048] In this embodiment, the initial priority ranking of targets is set based on the distance and risk-based ranking logic in the prior art. The cutting angle refers to the angle formed between the driving direction of the adjacent target and the lane line of the current driving lane of the vehicle. When the angle is greater than the reference cutting angle (e.g., 10°, usually less than 90° is used to determine the cutting angle), it indicates that the trend of the target cutting into the lane of the vehicle has reached the level that requires close attention. The larger the angle, the more obvious the cutting intention and the higher the risk of side collision. When the angle is less than or equal to the corresponding reference value, it is temporarily not included in the monitoring focus of the sub-target priority. The trajectory intersection probability refers to the probability of the two overlapping paths in the future by analyzing the real-time movement trajectory (such as speed, direction, steering angle, etc.) of the dynamic interference target and the planned driving trajectory of the vehicle. The higher the probability, the greater the potential risk of the target colliding with the vehicle, and it should be given priority in the monitoring and early warning scope.

[0049] The innovation of this example lies in incorporating electrical safety levels into the consideration and refining the hierarchical dimensions of dynamic and static targets, forming a complete system through multi-level priority settings. The values ​​of parameters such as the upper limit of the first / second / third relative distance, the reference entry angle, and the probability of reference trajectory intersection can be flexibly fine-tuned according to specific scenarios such as highways, urban areas, and rural areas.

[0050] This example combines electrical safety levels to make the priority ranking more closely reflect the actual impact of power supply on sensor accuracy; the refined multi-level settings take into account both near and far targets, enabling early identification of high-risk targets at a distance while accurately focusing on sudden dangers at close range; the progression from initial ranking to precise classification significantly improves the accuracy of priority assessment, reduces missed and false judgments, allows high-risk targets to receive priority responses, buys drivers more time to avoid collisions, and effectively reduces the risk of collisions.

[0051] Furthermore, the specific process for target priority response verification is as follows: Obtain the duty cycle of the signal fluctuations from the multi-source sensors. The signal waveform duty cycle represents the proportion of the effective voltage output time in the low-voltage power supply circuit within one cycle, reflecting the fluctuation state of the power supply. If the signal waveform duty cycle is within the set range (usually set to 45%-55%), it indicates that the distance and speed monitoring deviations of each target obstacle are within an acceptable range, and the upper limit of the relative distance between each target is maintained. If the signal waveform duty cycle is lower than the minimum value of the set range, it indicates that the collision risk of distant target obstacles is high. If there is a risk of lag, the relative distance monitoring amplitude is reduced by the preset amplitude. The relative distance monitoring amplitude includes the first, second and third relative distance monitoring amplitudes, which represent the graded monitoring sensitivity set for target obstacles in different distance ranges. The amplitude of the reduction of the three amplitudes decreases in sequence to improve data reliability by shortening the monitoring distance and prioritize the timeliness of early warning for targets with high collision risk at close range. If the duty cycle of the signal waveform is higher than the maximum value of the set signal waveform duty cycle range, it indicates that there is a risk of overheating in the low voltage power supply circuit. In this case, only the first relative distance monitoring amplitude of the preset amplitude is reduced to focus on the accurate monitoring of obstacles directly in front.

[0052] Specifically, the target priority response verification also includes: acquiring the power supply response deviation of the low-voltage power supply circuit at various power supply moments. The power supply response deviation is used to quantify the dynamic difference between the actual power supply voltage and the rated voltage of the low-voltage power supply circuit, typically representing the instantaneous difference between the actual power supply voltage and the rated voltage. The average value of this difference at various power supply moments is then calculated to further quantify the power supply continuity of the low-voltage power supply circuit to the multi-source sensors. Based on the acquired power supply response deviation, transient suppression effect verification is performed. Specifically, if the acquired power supply response deviation is not greater than the set power supply response deviation, it indicates that the transient characteristics of the low-voltage power supply circuit meet the working requirements of the multi-source sensors, the power supply continuity is good, and the validity of the sensor data is not affected by power supply fluctuations. Collision detection is then performed. Risk level classification; if the acquired power supply response deviation is greater than the set power supply response deviation, corresponding transient suppression measures are applied, including but not limited to connecting a high-frequency filter capacitor (such as a solid capacitor, which quickly absorbs transient spikes) and activating a TVS diode (transient voltage suppression diode, which clamps overvoltage signals). After transient suppression, if the reacquired power supply response deviation is not greater than the set power supply response deviation, and the deviation recovery time is not greater than the set deviation recovery time, it indicates that the transient suppression effect is qualified and a collision risk level classification is performed; otherwise, a transient suppression warning is issued. The deviation recovery time refers to the length of time from the moment when the power supply response deviation exceeds the set power supply response deviation to the moment when the reacquired power supply response deviation is not greater than the set power supply response deviation.

[0053] It is important to understand that the collision warning database is a database established before the design of a multi-sensor fusion vehicle collision warning system to store various set data. The database includes, but is not limited to, set signal waveform duty cycle ranges, set power supply response deviations, set deviation recovery times, various relative distance upper limits, reference entry angles, reference trajectory intersection probabilities, etc. Various values ​​are directly set by technicians. The set power supply response deviation and set deviation recovery time can be determined based on the actual application scenario of the vehicle collision warning system. For example, the set power supply response deviation is represented by the sum and average of historical power supply response deviations during the historical transient suppression process in the collision warning database, and the set deviation recovery time is represented by the sum and average of historical deviation recovery times after the historical transient suppression ends in the collision warning database. In addition, various values ​​in the database can be set and fine-tuned by technicians according to actual debugging.

[0054] In this embodiment, the preset amplitude is set in conjunction with the sensor accuracy decay curve and the deviation recovery time. Since the deviation recovery time reflects the system recovery speed after a power supply anomaly, the amplitude needs to be adjusted accordingly to ensure data reliability during the recovery period. This example adapts to power supply fluctuations through a dynamic adjustment mechanism, and sets the amplitude in conjunction with the deviation recovery time to ensure that high-risk targets at close range can still be accurately monitored during power instability or recovery phases. Transient suppression measures, combined with deviation recovery time verification, quickly stabilize the power supply, reduce the data failure window, improve the real-time performance and accuracy of priority verification, and reduce the risk of early warnings caused by power supply anomalies.

[0055] Furthermore, the collision risk level classification process is as follows: A collision risk assessment value is obtained based on the target priority classification results. If the target obstacle is static, the obtained collision risk assessment value is recorded as the static collision risk assessment value. The static collision risk assessment value represents the ratio of the driving speed obtained after target priority classification to the corresponding relative distance, used to quantify the urgency of the collision. If the target obstacle is dynamic, the obtained collision risk assessment value is recorded as the dynamic collision risk assessment value. The dynamic collision risk assessment value represents the ratio of the relative acceleration obtained after target priority classification to the corresponding relative distance, used to quantify the rate of change of collision risk. The larger this value, the greater the relative acceleration between the dynamic obstacle and itself, and the closer the relative distance, the faster the collision risk deteriorates, and the more drastically the possibility of a collision in a short period of time increases. Emergency avoidance or deceleration measures should be prioritized to mitigate the risk.

[0056] If the obtained collision risk assessment value does not meet the preset risk level range, it indicates that the current driving state of the electric vehicle has no collision risk. Otherwise, the corresponding collision risk level is divided according to the following rules: When the collision risk assessment value is greater than or equal to the set first collision risk level threshold (usually set to 0.5), it is determined to be a first collision risk, triggering a level one emergency response, including immediately issuing a sharp audible and visual alarm, automatically applying emergency braking (dynamically adjusting the braking force according to the distance), and simultaneously shutting down non-essential on-board electrical equipment to ensure power supply to the braking system, ensuring that the probability of collision is reduced in the shortest possible time; when the collision risk assessment value is not less than the set second collision risk level threshold (usually set to 0.3) and less than the first collision risk level threshold, it is determined to be a second collision risk. The system detects a collision risk of 1.5 times, triggering a Level 2 warning response. This response provides continuous audible and visual alerts, activates the pre-pressurization function of the braking system, and displays the collision prediction time and avoidance suggestions on the instrument panel, reminding the driver to take proactive measures to slow down or avoid the collision. When the collision risk assessment value is not less than the set threshold for the third collision risk level (usually set to 0.1) and is less than the threshold for the second collision risk level, it is determined to be a Level 3 collision risk, triggering a Level 3 warning response. This response only displays a yellow warning icon and target distance information on the instrument panel, and alerts the driver to the target ahead through slight steering wheel or seat vibrations. The system continuously monitors the target's dynamics and updates the risk assessment value in real time. The first, second, and third collision risks exhibit a gradient relationship of decreasing risk level and progressively weakening response intensity.

[0057] It's important to understand that because the collision risk mechanisms of static and dynamic obstacles differ, the logic behind setting the critical values ​​for corresponding collision risk levels also differs. The risk of static obstacles stems from the urgency of decreasing distance, while that of dynamic obstacles arises from the rate of risk deterioration caused by relative acceleration. The physical meanings of these assessment values ​​differ. Therefore, in practical applications, differentiated thresholds can be set for the first, second, and third collision risk levels based on the risk evolution patterns of static and dynamic obstacles. This ensures that the corresponding critical values ​​better reflect the risk characteristics of different types of obstacles, improving the accuracy of the risk classification.

[0058] In this embodiment, the set third collision risk level threshold values ​​are all normalized results, and the corresponding specific values ​​can be fine-tuned according to vehicle type (e.g., small cars and large trucks), driving scenario (e.g., urban roads and highways), and sensor performance to adapt to risk perception needs under different conditions, further enhancing the system's practicality and reliability, and providing more accurate protection for vehicle driving safety. This example accurately quantifies the collision risk of different types of targets through differentiated calculation of static and dynamic collision risk assessment values, achieving a scientific classification of risk levels. The gradient response mechanism, from Level 1 emergency braking to Level 3 mild warning, avoids excessive warnings that interfere with the driver, while enabling rapid action in high-risk situations, significantly improving collision avoidance efficiency.

[0059] like Figure 2 The flowchart shown illustrates the time drift correction and degraded operation verification process. Its design logic is as follows: First, obtain the timestamps of the baseline time axis and data collected from multiple sensors, and classify them into different collision risk timestamps. For the first collision risk timestamp, calculate the drift amount using the sliding window method; for the second / third, estimate the drift trend using linear interpolation, and obtain the synchronization error and drift trend deviation rate respectively. Then, determine whether the set values ​​are met. If they are, the data synchronization accuracy is considered qualified; otherwise, trigger degraded operation verification to ensure data reliability and system safety.

[0060] Further understanding is needed regarding the drift estimation and correction process, which involves: acquiring a reference time axis established by a high-precision onboard clock; simultaneously acquiring timestamps of data collected by multiple sensors. The reference time axis provides a unified time reference standard for the multi-source sensor data, eliminating the inconsistency in time references caused by clock deviations between different sensors. The timestamps include a first collision risk timestamp and second and third collision risk timestamps. The first collision risk timestamp represents the time stamp when the multi-source sensors collect data from the first collision risk target, while the second and third collision risk timestamps represent the time stamps when the multi-source sensors collect data from the second and third collision risk targets, respectively. The drift amount corresponding to the first collision risk timestamp is calculated using the sliding window method, while the drift trend corresponding to the second and third collision risk timestamps is estimated using linear interpolation. The drift amount is used to quantify the instantaneous deviation between the first collision risk timestamp and the reference time axis, and the drift trend is used to predict the offset pattern of the second and third collision risk timestamps over time.

[0061] It's important to understand that the specific process for determining whether to perform a downgraded operation verification is as follows: Obtain the synchronization error of the first collision risk timestamp after drift estimation and correction, and the drift trend deviation rate of the second and third collision risk timestamps. The synchronization error represents the remaining time difference between the first collision risk timestamp after drift estimation and correction and the baseline time axis, used to measure the time alignment accuracy of high-priority target data. The drift trend deviation rate reflects the difference between the actual drift amount and the predicted drift trend of the second and third collision risk timestamps, used to assess the time stability of low-priority target data. If the obtained synchronization error is not greater than the set synchronization error value (usually set to 10ms), the assessment is successful. If the obtained drift trend deviation rate is not greater than the set drift trend deviation rate (usually set to 2%), the data synchronization accuracy is deemed qualified and no downgrade verification is required. If the obtained synchronization error is greater than the set synchronization error value, or the obtained drift trend deviation rate is greater than the set drift trend deviation rate, downgrade operation verification is triggered to verify the reliability of collision warning under the simplified strategy. For example, a simplified fusion strategy is enabled (such as retaining only radar and camera data fusion), and the warning accuracy (≥85%) and response latency (≤300ms) of electric vehicles corresponding to the collision risk level change in the current driving scenario are tested. If either indicator fails to meet the standard, sensor redundancy warning is performed.

[0062] In this embodiment, the sliding window method selects a window of a certain time length and calculates the timestamp data of the first collision risk within the window to obtain its drift amount relative to the reference time axis, which can dynamically capture instantaneous deviations. The set synchronization error value, drift trend deviation rate, and other parameters can be adaptively adjusted according to the sensor type (such as lidar and millimeter-wave radar), data transmission rate, and scene real-time requirements.

[0063] The innovation of this example lies not in the data itself, but in linking time synchronization accuracy with collision risk levels. It employs differentiated estimation and correction methods for different levels and a mechanism to dynamically trigger degradation verification based on synchronization effectiveness. By using the sliding window method and linear interpolation, time deviations are precisely corrected to ensure data time alignment. Degradation is determined by combining synchronization error and drift trend deviation rate, initiating verification when data is unreliable. This ensures accurate warnings and fast response under simplified strategies, improves the system's robustness and reliability in complex situations, and provides a precise time benchmark for collision warnings.

[0064] This invention provides a vehicle collision warning method based on multi-sensor fusion, such as... Figure 3The flowchart shown is for a vehicle collision warning method based on multi-sensor fusion. The processing flow of this method may include the following steps: Step 1, by collecting battery data and combining it with the power supply circuit fluctuation status of multi-source sensors, the electrical safety status is determined and the electrical safety level is classified; Step 2, the target priority is classified according to the marked electrical safety level, and the collision risk level is classified simultaneously; Step 3, the drift estimation and correction of the timestamps collected by the multi-source sensors are performed based on the collision risk level classification results to calibrate the time synchronization error of the multi-source sensor data, so as to determine whether to perform degraded operation verification.

[0065] In this embodiment, early warning effectiveness is enhanced through a three-step collaborative approach: Step 1 determines electrical safety from the power supply head to ensure the basic reliability of sensor data; Step 2 prioritizes targets and categorizes collision risks based on electrical levels to ensure high-risk targets receive priority responses; Step 3 calibrates time synchronization errors and dynamically triggers degradation verification to enhance adaptability to complex operating conditions. These three steps form a closed loop, reducing misjudgments caused by power supply anomalies and data asynchrony, improving early warning accuracy and system robustness, and effectively reducing collision risks.

[0066] It should be added that, such as Figure 4 The homepage interface of the vehicle collision warning monitoring platform shown is an interactive presentation of a multi-sensor fusion-based vehicle collision warning system. The left side is a function navigation bar, covering access points for the homepage, alarms, and history. The right side displays core information: a system status overview showing the operating status (normal operation), electrical safety level (Level 1 safety), highest collision risk (Level 3 risk), and risk update time, allowing for quick understanding of the system's basic situation; core sensor status display showing that millimeter-wave radar, cameras, and lidar are all connected, ensuring stable data acquisition sources; warning status details breaking down the distance and angle of three types of targets directly ahead—those in the current lane, those cutting in from adjacent lanes, and those with static faults—accurately presenting collision risks; system information includes communication status (synchronizing with the cloud), hardware load rate, and an emergency contact entry to assist in handling anomalies, providing comprehensive support for vehicle collision warning operations.

[0067] like Figure 5 The battery and electrical safety verification page of the vehicle collision warning monitoring platform shown is an important component of the multi-sensor fusion-based vehicle collision warning system. This module can be accessed via the left navigation bar. The page displays key monitoring data for the battery and electrical system: voltage, current, temperature, low-voltage circuit voltage, SOC value, current fluctuation amplitude, and other parameters, presented through visual charts and numerical values. The safety score reaches 92 points (Level 1 safety), with all key parameters within acceptable ranges. For example, voltage deviation of 1.3% is within ±5%, and the low-voltage circuit voltage of 12.2V is normal. These reasonable parameters ensure the stability of the electrical system, providing reliable power to sensors and other equipment, and supporting the accurate operation of the collision warning system from its electrical source.

[0068] The above embodiments can be implemented, in whole or in part, by software, hardware (such as circuits), firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the flow or function according to the embodiments of the present invention is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired (e.g., infrared, wireless, microwave, etc.) means. A computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. Available media can be magnetic media (e.g., floppy disks, hard disks, magnetic tapes), optical media (e.g., DVDs), or semiconductor media. Semiconductor media can be solid-state drives.

[0069] It should be understood that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. A and B can be singular or plural. Additionally, the character " / " in this article generally indicates an "or" relationship between the preceding and following related objects, but it can also represent an "and / or" relationship. Please refer to the context for a more accurate understanding.

[0070] In various embodiments of the present invention, the order of the above-mentioned process numbers does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0071] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementations should not be considered beyond the scope of this invention.

[0072] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A vehicle collision warning system based on multi-sensor fusion, characterized in that, include: Battery and electrical safety verification module, collision warning level classification module, time drift correction and degraded operation verification module; The battery and electrical safety verification module is used to determine the electrical safety status and classify the electrical safety level by collecting battery data and combining it with the power supply circuit fluctuation status of multi-source sensors. The battery data includes the voltage, current, temperature, and SOC value of the high-voltage power battery and the circuit voltage data of the low-voltage auxiliary battery collected in real time by the BMS. The circuit voltage data includes the low-voltage circuit voltage value and the voltage fluctuation peak value. The power supply circuit fluctuation status is used to reflect the voltage stability and current fluctuation amplitude of the power supply circuit of the multi-source sensors to evaluate the impact on the acquisition accuracy of the multi-source sensors. The collision warning level classification module is used to classify target priorities based on the marked electrical safety level, and at the same time classify collision risk levels. The target priority classification is used to determine the monitoring target type of electric vehicle in different driving scenarios, so as to ensure that high collision risk targets are processed first. The time drift correction and degraded operation verification module is used to estimate and correct the drift of the timestamps collected by the multi-source sensors based on the collision risk level classification results, so as to calibrate the time synchronization error of the multi-source sensor data and determine whether to perform degraded operation verification. The degraded operation verification is used to improve the accuracy of early warning and the timeliness of response when the collision risk level changes under different driving scenarios.

2. The vehicle collision warning system based on multi-sensor fusion as described in claim 1, characterized in that, The specific process for determining the electrical safety status and classifying the electrical safety levels is as follows: Determine whether all acquired battery data are within safety limits, including high-voltage battery voltage range, temperature safety range, minimum SOC limit, low-voltage circuit voltage standard value, and upper limit of voltage fluctuation peak value; If yes, it indicates that the power supply status of the battery and multi-source sensors is stable and electrical safety levels are divided; otherwise, an electrical abnormality warning is sent and temporary power supply adjustment is initiated. The electrical safety level classification is as follows: The deviation scores of each battery data point from the corresponding safety limit are calculated separately and then harmonic averaged to obtain the electrical state safety score, which quantifies the safety and stability of the battery and multi-source sensor power supply status. When the electrical safety score is greater than or equal to the set upper limit of the electrical safety score, the corresponding electrical safety level will be determined as a safety level in which the power supply status of the battery and multi-source sensors has no impact on the accuracy of the collision warning. When the electrical safety score is not less than the reference electrical safety score and not greater than the set upper limit of the electrical safety score, the corresponding electrical safety level will be classified as a safety level that may have an impact. When the electrical safety score is less than the reference electrical safety score and not less than the set lower limit of the electrical safety score, the corresponding electrical safety level will be classified as a safety level that may cause the collision warning to deviate. When the electrical safety score is less than the set lower limit of the electrical safety score, an electrical safety warning will be issued.

3. The vehicle collision warning system based on multi-sensor fusion as described in claim 1, characterized in that, The specific process for prioritizing the targets is as follows: When the relative distance between the electric vehicle and the target obstacle is less than the set safe distance, the relative distance between the electric vehicle and the target obstacle is obtained; otherwise, it indicates that there is no risk of collision between the electric vehicle and the target obstacle. The relative distance represents the driving path distance between the current position of the electric vehicle and the position of the target obstacle. If the target obstacle is static, then obtain the electric vehicle's speed; If the target obstacle is dynamic, the relative acceleration between the electric vehicle and the target obstacle is obtained, and a priority assessment is performed in conjunction with the trajectory of the dynamic obstacle. The priority assessment includes a first priority assessment and a second priority assessment. Following the priority assessment, target priority response verification is also included to dynamically verify and optimize the priority assessment results.

4. The vehicle collision warning system based on multi-sensor fusion as described in claim 3, characterized in that, The priority evaluation process is as follows: The initial priority ranking of targets corresponding to electrical safety levels is obtained respectively. The initial priority ranking of targets represents the preliminary ranking of various target obstacles according to the level of collision risk under different electrical safety states. The preliminary ranking settings include the setting of a first priority evaluation and the setting of a second priority evaluation; The first priority assessment includes setting a first target priority, setting a second target priority, and setting a third target priority, with the collision risk corresponding to the three decreasing sequentially; The second priority assessment includes setting the priority of the first sub-target, the priority of the second sub-target, and the priority of the third sub-target, and the three are progressively increased according to the urgency of the close-range dynamic risk; The setting of the first priority assessment and the setting of the second priority assessment realize the progression from preliminary sorting to precise classification, thereby improving the accuracy of priority assessment.

5. The vehicle collision warning system based on multi-sensor fusion as described in claim 4, characterized in that, The first target priority setting indicates that the relative distance is not less than the first relative distance upper limit and the driving speed is greater than 0 for the target in the lane directly in front, so as to identify high collision risk targets approaching from a distance in advance; The setting of the second target priority indicates that the adjacent lane cutting-in prediction target is at a relative distance of not less than the upper limit of the second relative distance, so as to reduce the risk of side collision; The setting of the third target priority indicates that the relative distance is not less than the upper limit of the third relative distance for static fault targets, so as to improve the timeliness of early warning of collision risk with fixed obstacles on the road; The setting of the first sub-target priority means that dynamic targets in the lane directly ahead that are less than the upper limit of the first relative distance and have a relative acceleration greater than 0 will be given priority monitoring as high-risk targets that are accelerating close at close range, so as to trigger an emergency avoidance warning in a timely manner. The setting of the second sub-target priority means that adjacent lane cut-in prediction targets with a relative distance less than the upper limit of the second relative distance and a cut-in angle greater than the reference cut-in angle are included to enhance the monitoring of adjacent lane vehicles with obvious cut-in trends and improve the response speed of side collision risk. The setting of the third sub-target priority means that dynamic interference targets whose relative distance is less than the upper limit of the third relative distance and whose trajectory intersection probability is greater than the reference trajectory intersection probability will be given priority in handling sudden moving targets that may intersect with the vehicle's driving trajectory, thereby reducing the risk of collision.

6. The vehicle collision warning system based on multi-sensor fusion as described in claim 3, characterized in that, The specific process for verifying the target priority response is as follows: The duty cycle of signal fluctuations from multiple sensor feedback is obtained to reflect the fluctuation state of the power supply. If the signal waveform duty cycle is within the set signal waveform duty cycle range, it indicates that the distance and speed monitoring deviation of each target obstacle is within the acceptable range. At this time, the current upper limit of the relative distance of each target is maintained. If the duty cycle of the signal waveform is lower than the minimum value of the set signal waveform duty cycle range, it indicates that there is a lag risk in the collision risk assessment of distant target obstacles. At this time, the relative distance monitoring amplitude is reduced by the preset amplitude respectively. The relative distance monitoring amplitude includes the first, second and third relative distance monitoring amplitudes, which represent the graded monitoring sensitivity set for target obstacles in different distance ranges, and the reduction amplitude of the three decreases in sequence. If the duty cycle of the signal waveform is higher than the maximum value of the set signal waveform duty cycle range, it indicates that there is a risk of overheating in the low-voltage power supply circuit. At this time, only the first relative distance monitoring amplitude of the preset amplitude is reduced to focus on the accurate monitoring of obstacles directly in front.

7. The vehicle collision warning system based on multi-sensor fusion as described in claim 6, characterized in that, The target priority response verification also includes: The power supply response deviation of the low-voltage power supply circuit at various power supply times is obtained. The power supply response deviation is used to quantify the dynamic difference between the actual power supply voltage and the rated voltage of the low-voltage power supply circuit. The transient suppression effect was verified based on the obtained power supply response deviation. The specific process is as follows: If the obtained power supply response deviation is not greater than the set power supply response deviation, it indicates that the transient characteristics of the low-voltage power supply circuit meet the working requirements of the multi-source sensor, and the collision risk level is classified. If the obtained power supply response deviation is greater than the set power supply response deviation, the corresponding transient suppression measures are applied. After transient suppression, if the re-obtained power supply response deviation is not greater than the set power supply response deviation and the deviation recovery time is not greater than the set deviation recovery time, it indicates that the transient suppression effect is qualified and a collision risk level is classified; otherwise, a transient suppression warning is issued. The deviation recovery time refers to the length of time from the moment when the power supply response deviation exceeds the set power supply response deviation to the moment when the re-acquired power supply response deviation is no greater than the set power supply response deviation.

8. The vehicle collision warning system based on multi-sensor fusion as described in claim 7, characterized in that, The specific process for classifying collision risk levels is as follows: The collision risk assessment value is obtained based on the result of target priority classification; If the target obstacle is static, the obtained collision risk assessment value will be recorded as the static collision risk assessment value, which will be used to quantify the time urgency of the collision. If the target obstacle is dynamic, the obtained collision risk assessment value will be recorded as the dynamic collision risk assessment value, which is used to quantify the rate of change of collision risk. If the obtained collision risk assessment value does not meet the preset risk level range, it indicates that the current driving state of the electric vehicle has no collision risk; otherwise, the corresponding collision risk level is determined according to the following rules: When the collision risk assessment value is greater than or equal to the set first collision risk level threshold, it is determined to be a first collision risk, thus triggering a first-level emergency response. When the collision risk assessment value is not less than the set threshold value for the second collision risk level and is less than the threshold value for the first collision risk level, it is determined to be a second collision risk, thereby triggering a level two warning response. When the collision risk assessment value is not less than the set third collision risk level threshold and is less than the second collision risk level threshold, it is determined to be a third collision risk, thus triggering a level three warning response. The first collision risk, the second collision risk, and the third collision risk exhibit a gradient relationship of decreasing risk level and progressively weakening response intensity.

9. The vehicle collision warning system based on multi-sensor fusion as described in claim 1, characterized in that, The drift estimation and correction process is as follows: A reference time axis is established by a high-precision vehicle clock, and timestamps of data collected by multiple sources are acquired synchronously. The reference time axis is used to provide a unified time reference standard for the data from multiple sources. The timestamps include a first collision risk timestamp and second and third collision risk timestamps. The first collision risk timestamp represents the time stamp when the multiple sources acquire data of the first collision risk target, and the second and third collision risk timestamps represent the time stamps when the multiple sources acquire data of the second and third collision risk targets. The drift amount corresponding to the first collision risk timestamp is calculated using the sliding window method, while the drift trend corresponding to the second and third collision risk timestamps is estimated using the linear interpolation method. The drift amount is used to quantify the instantaneous deviation between the first collision risk timestamp and the reference time axis, and the drift trend is used to predict the offset pattern of the second and third collision risk timestamps over time. The specific process for determining whether to perform a downgrade operation verification is as follows: The synchronization error of the first collision risk timestamp after drift estimation and correction and the drift trend deviation rate of the second and third collision risk timestamps are obtained. The synchronization error represents the remaining time difference between the first collision risk timestamp and the reference time axis after drift estimation and correction. The drift trend deviation rate is used to reflect the degree of difference between the actual drift amount and the predicted drift trend of the second and third collision risk timestamps. If the obtained synchronization error is not greater than the set synchronization error value, and the obtained drift trend deviation rate is not greater than the set drift trend deviation rate, then the data synchronization accuracy is deemed qualified and no downgrade verification is required. If the obtained synchronization error is greater than the set synchronization error value, or the obtained drift trend deviation rate is greater than the set drift trend deviation rate, a downgraded operation verification will be triggered to verify the reliability of the collision warning under the simplified strategy.

10. A vehicle collision warning method based on multi-sensor fusion, applied to a vehicle collision warning system based on multi-sensor fusion, characterized in that, Includes the following steps: Step 1: By collecting battery data and combining it with the power supply circuit fluctuation status of multi-source sensors, determine the electrical safety status and classify the electrical safety level. Step two: Prioritize targets based on their marked electrical safety levels, and simultaneously classify them according to collision risk levels; Step 3: Based on the collision risk level classification results, the timestamps collected by the multi-source sensors are drifted and corrected to calibrate the time synchronization error of the multi-source sensor data, so as to determine whether to perform degraded operation verification.

Images