Blind area monitoring and lane changing auxiliary intelligent safety detection system

Abstract

The invention discloses a blind area monitoring and lane changing auxiliary intelligent safety detection system. The system comprises a data acquisition subsystem, a risk calculation subsystem and a lane changing early warning subsystem. The data acquisition subsystem acquires speeds and collision distances of parallel vehicles in dead zones right behind and on two sides of a vehicle through millimeter-wave radars, the camera module acquires front image information, the data preprocessing module fuses multi-source data, and the environment sensing module adjusts a collision threshold according to illumination, rainfall and the like. And the risk calculation subsystem calculates a collision risk value by using a dynamic threat integral and a space threat potential field model, predicts a lane changing trajectory through an LSTM neural network and optimizes risk pre-judgment in combination with a driver intention. And the lane changing early warning subsystem triggers early warning in a grading manner through a multi-modal output unit based on the collision risk value and the lane changing conflict probability. According to the system, through multi-sensor fusion and an intelligent algorithm, blind area monitoring accuracy, environmental adaptability and driver response efficiency are improved, and lane changing safety is guaranteed.

Description

Technical Field

[0001] The present invention relates to the field of vehicle driving detection, and in particular to a blind spot monitoring and lane change assistance intelligent safety detection system. Background Art

[0002] With the increasing number of cars on the road, traffic accidents caused by insufficient blind spot monitoring during lane changes are becoming more frequent. Existing blind spot monitoring systems primarily rely on single sensors (such as cameras or millimeter-wave radars), which suffer from poor environmental adaptability and insufficient data fusion accuracy, making it difficult to accurately predict collision risks in complex traffic scenarios. For example, camera imaging quality significantly degrades in rainy and foggy weather or when illumination changes dramatically. Traditional radar algorithms assess risk based solely on distance and speed, failing to incorporate key factors such as the road friction coefficient and the driver's lane change intentions, resulting in delayed warnings or false alarms. Furthermore, existing systems employ a single warning method (mostly visual or auditory warnings) and lack a multimodal collaborative feedback mechanism. This makes it impossible to dynamically adjust warning intensity based on risk levels, leaving drivers with room for improvement in response efficiency.

[0003] Therefore, in order to solve the above problems, the present invention proposes a blind spot monitoring and lane change assistance intelligent safety detection system. Summary of the Invention

[0004] In view of the shortcomings of the existing technology, the purpose of the present invention is to provide a blind spot monitoring and lane change assistance intelligent safety detection system.

[0005] To achieve the above object, the present invention provides the following technical solutions: A blind spot monitoring and lane change assistance intelligent safety detection system, comprising: A data acquisition subsystem, including a millimeter-wave radar, is used to collect information about parallel vehicles in the direction directly behind the monitoring vehicle and in the blind spots on both sides, the parallel vehicle information including vehicle speed and collision distance; a risk calculation subsystem, comprising a risk calculation module for calculating a collision risk value based on the parallel vehicle information; and outputting a corresponding warning signal based on the direction of the parallel vehicle and entering a lane change warning step when the collision risk value exceeds a set collision threshold; The lane change warning subsystem includes a warning judgment module and a warning output module. The warning judgment module receives the warning signal and determines and outputs the warning type; the warning output module outputs the corresponding warning according to the judgment result of the warning judgment module.

[0006] As a further improvement of the present invention, the data acquisition subsystem further includes a camera module for acquiring image information in front of the monitoring vehicle, wherein the image information includes lane line information and traffic sign information.

[0007] As a further improvement of the present invention, the data acquisition subsystem also includes a data preprocessing module for performing spatiotemporal synchronization, noise filtering and data fusion processing on the raw data collected by the millimeter wave radar and camera module.

[0008] As a further improvement of the present invention, the risk calculation subsystem is further configured with a traffic flow prediction model for predicting the lane change trajectories of parallel vehicles based on the parallel vehicle information and calculating the lane change conflict probability value.

[0009] As a further improvement of the present invention, the risk calculation module also includes a module for calculating a dynamic threat integral by combining the parallel vehicle speed, the monitoring vehicle speed, the speed difference sensitivity coefficient, and the time-varying relative distance function, and calculating a spatial threat potential field by combining the road surface information and the relative distance, and performing a comprehensive risk decision calculation by combining the dynamic threat integral and the spatial threat potential field to obtain a collision risk value.

[0010] As a further improvement of the present invention, the warning judgment module is used to issue a warning signal based on the lane change trajectory of the parallel vehicle combined with the lane change conflict probability. When the collision risk value is greater than the set collision threshold, if the lane change conflict probability value of the parallel vehicle is lower than the preset conflict threshold, a visual warning signal is issued; if the lane change conflict probability value of the parallel vehicle is higher than the preset conflict threshold, an auditory warning signal and a tactile warning signal in the same direction as the lane change trajectory are issued.

[0011] As a further improvement of the present invention, the data acquisition subsystem also includes an environmental perception module for collecting ambient light intensity, rainfall intensity and road surface slipperiness parameters, and adjusting the collision threshold according to the ambient light intensity, rainfall intensity and road surface slipperiness parameters.

[0012] As a further improvement of the present invention, the risk calculation subsystem also includes an intention recognition module, which is used to judge the lane-changing intention of the monitored vehicle based on the steering wheel angle, turn signal and accelerator pedal travel combined with image information. When the monitored vehicle has the intention to change lanes, the risk calculation module calculates the collision risk value.

[0013] As a further improvement of the present invention, the warning output module also includes a multimodal output unit, which selects to output a visual warning signal or simultaneously output an auditory warning signal and a tactile warning signal according to the output result of the warning judgment module.

[0014] The beneficial effects of the present invention are as follows: through the multi-sensor fusion of millimeter-wave radar, camera, and environmental perception module, combined with the dynamic threat integral fusion of speed difference, relative distance and spatial threat potential field fusion of road friction coefficient and vehicle overlap, accurate calculation of collision risk in the time and space dimensions is achieved. At the same time, the LSTM neural network is introduced to predict the lane change trajectory of parallel vehicles, and the steering wheel angle, turn signal signal, etc. are identified in combination with the driver's lane change intention, thereby improving the real-time and accuracy of risk prediction. Through the fusion of multiple technologies and the optimization of intelligent algorithms, the accuracy, environmental adaptability and driver response efficiency of blind spot monitoring are significantly improved, providing technical support for vehicle lane change safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 This is a block diagram of the blind spot monitoring and lane change assistance intelligent safety detection system of the present invention; DETAILED DESCRIPTION The present invention will be described in further detail below with reference to the accompanying drawings and embodiments. Identical components are denoted by the same reference numerals. It should be noted that the terms "front," "rear," "left," "right," "upper," and "lower" used in the following description refer to directions in the accompanying drawings, and the terms "bottom," "top," "inner," and "outer" refer to directions toward or away from the geometric center of a particular component, respectively.

[0016] A blind spot monitoring and lane change assistance intelligent safety detection system, such as Figure 1 Shown, including: The data acquisition subsystem includes millimeter-wave radars, which collect information about vehicles traveling directly behind the monitoring vehicle and in its blind spots. This information includes vehicle speed and collision distance. Millimeter-wave radars are installed on both sides and in the center of the vehicle's rear bumper to cover the rear and left and right blind spots. This information includes the vehicle's speed and collision distance from the monitoring vehicle. This data is transmitted to the risk calculation subsystem via the CAN bus at a sampling frequency of 20 Hz to ensure real-time performance.

[0017] The data acquisition subsystem also includes a camera module for capturing and monitoring image information in front of the vehicle, including lane markings and traffic signs. The camera is mounted inside the front windshield, with a field of view covering the vehicle's front and both lanes. Image processing algorithms are used to extract lane markings and traffic sign information, such as lane departure warning lines and speed limit signs. Data from the camera module and the millimeter-wave radar are synchronized using timestamps to ensure data consistency.

[0018] Furthermore, the data acquisition subsystem also includes a data preprocessing module for performing spatiotemporal synchronization, noise filtering and data fusion processing on the raw data collected by the millimeter wave radar and camera module.

[0019] The preprocessing steps include: spatiotemporal synchronization, which aligns radar and camera data through timestamps to eliminate time differences between sensors; noise filtering, which uses a Kalman filter algorithm to remove noise from radar data and improve data accuracy; and data fusion, which combines the radar's distance and speed information with the camera's image information to generate a more comprehensive environmental model.

[0020] Furthermore, the data acquisition subsystem also includes an environmental perception module, which collects ambient light intensity, rainfall intensity, and road slipperiness parameters and adjusts the collision threshold based on these parameters. This environmental perception module is implemented using a light sensor, a rain sensor, and a road friction coefficient sensor. The collision threshold is dynamically adjusted based on these environmental parameters, for example, lowering the collision threshold by 20% on rainy days to improve system sensitivity.

[0021] a risk calculation subsystem, comprising a risk calculation module for calculating a collision risk value based on the parallel vehicle information; and outputting a corresponding warning signal based on the direction of the parallel vehicle and entering a lane change warning step when the collision risk value exceeds a set collision threshold; Furthermore, the risk calculation module also includes a function for calculating a dynamic threat score by combining the parallel vehicle speed, the monitoring vehicle speed, the speed difference sensitivity coefficient, and the time-varying relative distance function, and calculating a spatial threat potential field by using road surface information and relative distance, and combining the dynamic threat score and the spatial threat potential field to perform a comprehensive risk decision calculation to obtain a collision risk value.

[0022] Specifically, the risk calculation module uses the collected data to measure the temporal and spatial likelihood of a collision between the rear parallel vehicle and the monitored vehicle, and ultimately generates a final collision risk calculation function. The dynamic threat score function is configured as follows: ; T dyn is a dynamic threat integral function, which represents the collision threat caused by the speed between the parallel vehicle behind and the monitoring vehicle. v2 is the speed of the parallel vehicle, and v1 is the speed of the monitoring vehicle. The secondary amplification of the speed difference, i.e., the relative speed, makes the nonlinear growth threat in the high-speed difference scenario more intuitive. is the speed difference sensitivity coefficient, which represents the preset attenuation of the speed difference effect on the threat integral function. It adjusts the threat weight of the speed difference to compensate for the differences in speed tolerance of different road types. It is obtained through training with collected historical accident data. dτ is the time-varying relative distance function, reflecting the interaction between the two vehicles in the kinematic equation. is the distance perception attenuation coefficient, which is used to simulate the nonlinear sensitivity of human drivers to distance. It is determined by combining the driver reaction time experiment summarized by big data. c is the critical time window, which is used to define the time required to avoid collision in the worst case, and τ is the integration time variable.

[0023] At this time, when the two vehicles are moving at the same speed, only the Gaussian kernel function is retained. At this time, the threat value is determined only by the distance. When d> , the exponential term decays to less than 10% of the initial value, which is consistent with the driver's intuitive judgment of the safe distance.

[0024] The spatial threat potential field function configuration is: ; Among them, T spa is the spatial threat potential field function, μ represents the road friction coefficient, and is used to integrate tire-road friction characteristics into the spatial threat field, correcting the variation of potential field strength with road conditions. d0 is the initial relative distance, ρ is the vehicle projection overlap, and when ρ > 1, it indicates physical space overlap, indicating the possibility of collision. k! is the series expansion order, and n represents the maximum convergence order. Modeling is performed by fully coupling threats based on road friction characteristics, vehicle geometric parameters, and kinematic state.

[0025] The comprehensive risk decision function configuration is: ; Among them, R is the collision risk value of the comprehensive risk decision function, β is the risk amplification factor, which is used to adjust the system sensitivity and compensate for the influence of sensor noise, and ε is the gradient suppression coefficient, which is used to suppress virtual alarms in high gradient areas. is the environmental characteristic gradient, which is preset based on the fusion of road curvature, visibility and obstacle density, a2 is the acceleration of the rear parallel vehicle, and a safe is the safety acceleration threshold, It is a nonlinear adjustment parameter used to control the judgment sharpness of acceleration difference and prevent small fluctuations in acceleration from causing state jumps. is the Hadamard product.

[0026] Furthermore, the risk calculation subsystem is equipped with a traffic flow prediction model, which is used to predict the lane change trajectories of parallel vehicles based on parallel vehicle information and calculate the probability of lane change conflicts. Based on historical data and real-time information, the model uses an LSTM neural network to predict vehicle trajectories within the next three seconds and calculate the probability of lane change conflicts.

[0027] Furthermore, the risk calculation subsystem also includes an intention recognition module, which is used to judge the lane-changing intention of the monitored vehicle based on the steering wheel angle, turn signal signal and accelerator pedal travel combined with image information. When the monitored vehicle has the intention to change lanes, the risk calculation module calculates the collision risk value. The specific judgment logic is: if the steering wheel angle is greater than 5° and the turn signal is on, it is judged as an active lane-changing intention; if the accelerator pedal travel suddenly increases and the lane line deviates, it is judged as a potential lane-changing intention; when there is an intention to change lanes, the risk calculation module gives priority to calculating the collision risk value.

[0028] The lane change warning subsystem includes a warning judgment module and a warning output module. The warning judgment module receives the warning signal and determines and outputs the warning type; the warning output module outputs the corresponding warning according to the judgment result of the warning judgment module.

[0029] Furthermore, the warning judgment module is used to issue a warning signal based on the lane change trajectory of the parallel vehicle and the lane change conflict probability. When the collision risk value is greater than the set collision threshold, if the lane change conflict probability value of the parallel vehicle is lower than the preset conflict threshold, a visual warning signal is issued; if the lane change conflict probability value of the parallel vehicle is higher than the preset conflict threshold, an auditory warning signal and a tactile warning signal in the same direction as the lane change trajectory are issued. Its multi-level warning trigger mechanism is as follows: Level 1 warning (low risk): When the collision risk value exceeds 80% of the set threshold but is lower than 100%, and the lane change conflict probability value is lower than 50% of the preset conflict threshold, the system only displays a yellow warning icon on the instrument panel without triggering audio or tactile feedback.

[0030] Level 2 warning (medium risk): When the collision risk value exceeds the set threshold and the lane change conflict probability value is between 50% and 80% of the preset conflict threshold, the system triggers a visual warning signal (HUD red flashing icon) and a low-frequency auditory prompt (single beep).

[0031] Level 3 warning (high risk): When the collision risk value significantly exceeds the set threshold (e.g., exceeds 120%), and the lane change conflict probability value is higher than 80% of the preset conflict threshold, the system simultaneously triggers a high-intensity visual warning (full-screen flashing red), a high-frequency auditory alarm (continuous beeping), and a tactile warning (steering wheel or seat vibration) in the same direction of the lane change trajectory.

[0032] The system dynamically updates the lane change conflict probability based on real-time traffic flow data, combined with the acceleration, turn signal status, and lane departure trends of parallel vehicles. If a parallel vehicle suddenly accelerates and activates its turn signal, the conflict probability will instantly increase by 20%.

[0033] Furthermore, the warning output module also includes a multimodal output unit that selects to output a visual warning signal or simultaneously output an auditory warning signal and a tactile warning signal based on the output results of the warning judgment module. The visual warning signal is displayed on the windshield via a heads-up display. The warning icon is projected on the windshield, with the position corresponding to the risk direction, with the left risk icon on the left and the right risk icon on the right. Dynamic effects: The icon color gradually changes from yellow to red, and the flashing frequency increases as the risk level increases.

[0034] Auditory warning signal: graded volume and frequency, level 2 warning, 500Hz single tone, volume 60dB; level 3 warning, 800Hz intermittent tone, volume 75dB.

[0035] The car's microphone monitors ambient noise and dynamically adjusts the volume. When the noise level is greater than 70dB, the volume is increased by 10dB.

[0036] Tactile warning signals: Steering wheel vibration, using a built-in eccentric motor with three levels of vibration intensity; seat vibration, for high-risk scenarios, the corresponding side of the driver's seat back vibrates synchronously.

[0037] Upon activation, the warning output module automatically detects the status of each feedback unit. If the haptic motor fails, the auditory warning volume is automatically increased. If the head-up display fails, the full-screen instrument panel warning is used instead. If the multimodal output unit fails completely, an emergency signal is sent to the vehicle ECU via the CAN bus, forcing the activation of the hazard warning lights.

[0038] The above shows and describes the basic features, principles, and advantages of the present invention. It should be noted that the present invention is not limited to the above embodiments, which are only some embodiments. Without departing from the spirit and scope of the present invention, various improvements and supplements made are considered to be within the scope of protection of the present invention.

Claims

1. A blind spot monitoring and lane change assistance intelligent safety detection system, characterized in that: include, A data acquisition subsystem, including a millimeter-wave radar, is used to collect information about parallel vehicles in the direction directly behind the monitoring vehicle and in the blind spots on both sides, the parallel vehicle information including vehicle speed and collision distance; a risk calculation subsystem, comprising a risk calculation module for calculating a collision risk value based on the parallel vehicle information; and outputting a corresponding warning signal based on the direction of the parallel vehicle and entering a lane change warning step when the collision risk value exceeds a set collision threshold; The lane change warning subsystem includes a warning judgment module and a warning output module. The warning judgment module receives the warning signal and determines and outputs the warning type. The early warning output module performs corresponding early warning output according to the judgment result of the early warning judgment module.

2. The blind spot monitoring and lane change assistance intelligent safety detection system according to claim 1, characterized in that: The data acquisition subsystem further includes a camera module for acquiring image information in front of the monitoring vehicle, wherein the image information includes lane line information and traffic sign information.

3. The blind spot monitoring and lane change assistance intelligent safety detection system according to claim 1, characterized in that: The data acquisition subsystem also includes a data preprocessing module for performing spatiotemporal synchronization, noise filtering, and data fusion processing on the raw data collected by the millimeter wave radar and camera modules.

4. The blind spot monitoring and lane change assistance intelligent safety detection system according to claim 1, characterized in that: The risk calculation subsystem is also equipped with a traffic flow prediction model for predicting the lane change trajectory of the parallel vehicle based on the parallel vehicle information and calculating the lane change conflict probability value, which is used to represent the probability of the parallel vehicle colliding with the monitoring vehicle.

5. The blind spot monitoring and lane change assistance intelligent safety detection system according to claim 1, characterized in that: The risk calculation module also includes a function for calculating a dynamic threat integral by combining the parallel vehicle speed, the monitoring vehicle speed, the speed difference sensitivity coefficient, and the time-varying relative distance function, and calculating a spatial threat potential field by combining the road surface information and the relative distance, and performing a comprehensive risk decision calculation based on the dynamic threat integral and the spatial threat potential field to obtain a collision risk value.

6. The blind spot monitoring and lane change assistance intelligent safety detection system according to claim 1 or 4, characterized in that: The warning judgment module is used to issue a warning signal based on the lane change trajectory of the parallel vehicles in combination with the lane change conflict probability. When the collision risk value is greater than the set collision threshold, if the lane change conflict probability value of the parallel vehicles is lower than the preset conflict threshold, a visual warning signal is issued; if the lane change conflict probability value of the parallel vehicles is higher than the preset conflict threshold, an auditory warning signal and a tactile warning signal in the same direction as the lane change trajectory are issued.

7. The blind spot monitoring and lane change assistance intelligent safety detection system according to claim 1, characterized in that: The data acquisition subsystem further includes an environment perception module for collecting parameters of ambient light intensity, rainfall intensity, and road surface slipperiness, and adjusting the collision threshold according to the parameters of ambient light intensity, rainfall intensity, and road surface slipperiness.

8. The blind spot monitoring and lane change assistance intelligent safety detection system according to claim 1 or 2, characterized in that: The risk calculation subsystem also includes an intention recognition module, which is used to determine the lane-changing intention of the monitored vehicle based on the steering wheel angle, turn signal signal and accelerator pedal travel combined with image information. When the monitored vehicle has the intention to change lanes, the risk calculation module calculates the collision risk value.

9. The blind spot monitoring and lane change assistance intelligent safety detection system according to claim 6, characterized in that: The warning output module also includes a multimodal output unit, which selects to output a visual warning signal or simultaneously output an auditory warning signal and a tactile warning signal according to the output result of the warning judgment module.

Images