A motorcycle helmet multi-source ranging fusion collision warning method

Abstract

This invention relates to the technical field of traffic safety. Specifically, it relates to a multi-source ranging fusion collision warning method for motorcycle helmets. The method includes determining the motorcycle's driving condition based on attitude information from synchronously detected information; adjusting the scanning frequency and range of long-distance ranging information in the outer direction and suppressing echoes in the inner direction when the driving condition is cornering; maintaining continuous scanning of long-distance ranging information and directionally correcting the scanning weight of short-distance ranging information when the driving condition is straight; obtaining attitude-compensated multi-source detection information; determining the environmental visibility state based on the multi-source detection information; obtaining fused target position information and fused target speed information; calculating the target collision time and target distance based on the target position information and target speed information; generating warning trigger information; and generating a helmet warning command based on the warning trigger information. This invention improves motorcycle driving safety.

Description

Technical Field

[0001] This invention relates to the technical field of traffic safety, and in particular to a multi-source ranging fusion collision warning method for motorcycle helmets. Background Technology

[0002] In the past, several methods were commonly used to implement collision warning for motorcycles. One method relied solely on a single ranging sensor, such as an ultrasonic sensor, to determine the risk of collision by detecting the distance to obstacles ahead. These sensors were relatively inexpensive and easy to install. Another method used cameras, employing image recognition technology to identify objects ahead and analyze the likelihood of a collision. Cameras can acquire rich visual information and can identify different types of obstacles. Some systems also used millimeter-wave radar, which can accurately measure the distance and speed of targets. These methods, to some extent, provided collision warnings for motorcycle riders.

[0003] However, existing technologies have significant drawbacks. Single ranging sensors have limitations in measurement range and accuracy, making them ill-suited for complex driving conditions, such as accurately monitoring obstacles in different directions while cornering. Cameras experience a significant drop in image recognition accuracy in low-visibility environments, such as rain or fog. While millimeter-wave radar offers high precision, it is poorly adaptable to changes in motorcycle posture and cannot flexibly adjust its detection range and frequency according to the motorcycle's driving conditions, resulting in delayed and inaccurate collision warnings in certain situations. Summary of the Invention

[0004] To improve motorcycle driving safety, this application provides a multi-source ranging fusion collision warning method for motorcycle helmets.

[0005] The above-mentioned objective of this application is achieved through the following technical solution:

[0006] A multi-source ranging fusion collision warning method for motorcycle helmets, comprising:

[0007] Acquire synchronous detection information, which includes near-range ranging information, far-range ranging information, and attitude information;

[0008] The motorcycle's driving condition is determined based on the attitude information in the synchronous detection information. When the driving condition is cornering, the scanning frequency and range of the long-distance ranging information in the outer direction are adjusted and the echo in the inner direction is suppressed. When the driving condition is straight, the continuous scanning of the long-distance ranging information is maintained, and the scanning weight of the short-distance ranging information is directionally corrected to obtain multi-source detection information with attitude compensation.

[0009] The environmental visibility status is determined based on multi-source detection information. When detecting low visibility conditions, the weights of near-range and long-range ranging information in the multi-source detection information are updated in the data fusion process based on the echo intensity distribution, point cloud sparsity, and stability parameters in the multi-source detection information to obtain the fused target location information and fused target velocity information.

[0010] Calculate the target collision time and target distance based on target location and speed information. Compare the target collision time with a preset collision time threshold and the target distance with a preset distance threshold. When either condition is met, generate a warning trigger message.

[0011] Based on the warning trigger information, generate helmet prompt instructions.

[0012] By adopting the above technical solution, and by acquiring synchronous detection information and combining it with the motorcycle's posture information to determine the driving conditions, the scanning frequency and range of the outer side of the long-distance ranging can be dynamically adjusted when cornering, while suppressing echo interference from the inner side. This adapts to complex turning paths and solves the problem of traditional ranging methods being unable to take into account directional flexibility. Maintaining continuous long-distance scanning and correcting the directional weights of near distances while driving straight helps improve the detection accuracy of targets directly in front. Furthermore, by combining echo intensity distribution, point cloud sparsity, and stability parameters in low-visibility environments, the fusion weights of near and far-distance ranging information are dynamically corrected, overcoming the problem of decreased recognition accuracy of image recognition under adverse weather conditions. Finally, based on the fused target position and speed information, the collision time and target distance are accurately determined, and helmet warning instructions are generated, achieving more realistic warning control and effectively improving riding safety.

[0013] Preferably, the motorcycle's driving condition is determined based on the attitude information in the synchronous detection information. When the driving condition is cornering, the scanning frequency and range of the long-distance ranging information in the outer direction are adjusted and the echo in the inner direction is suppressed, including:

[0014] Obtain the tilt angle and yaw rate from the attitude information;

[0015] When the tilt angle is greater than the preset tilt angle threshold and the yaw rate is greater than the preset yaw rate threshold, the motorcycle is determined to be in a cornering condition.

[0016] In cornering conditions, the scanning frequency of the outer direction in the long-distance ranging information is increased to the upper limit of the preset scanning frequency, and the scanning range of the outer direction is expanded to the preset expansion range. At the same time, the scanning frequency of the inner direction is reduced to the lower limit of the preset scanning frequency, and the scanning range of the inner direction is shrunk to the preset shrinkage range, so as to reduce the interference of the inner direction echo.

[0017] By employing the aforementioned technical solution, the tilt angle and yaw rate obtained from the attitude information can accurately reflect the dynamic state of the motorcycle in space. When the tilt angle exceeds a preset tilt angle threshold and the yaw rate exceeds a preset yaw rate threshold, the motorcycle can be determined to be in a cornering condition, thus distinguishing it from a straight-line driving condition. In this condition, increasing the scanning frequency of the outer direction in the long-distance ranging information to the upper limit of the preset scanning frequency and expanding the scanning range of the outer direction can enhance the detection coverage of potential targets on the outside of the curve. At the same time, reducing the scanning frequency of the inner direction to the lower limit of the preset scanning frequency and narrowing the scanning range of the inner direction helps to suppress interference echoes from the inside of the curve and avoid misjudgments caused by inner curbs, guardrails, etc., thereby improving the reliability and accuracy of target detection during cornering.

[0018] Preferably, when the driving condition is a straight line, continuous scanning of long-range ranging information is maintained, and the scanning weight of short-range ranging information is directionally corrected to obtain attitude-compensated multi-source detection information, including:

[0019] When the tilt angle value is less than or equal to the preset tilt angle threshold and the yaw rate is less than or equal to the preset yaw rate threshold, the motorcycle is determined to be in a straight-line working condition.

[0020] Under straight-line conditions, the omnidirectional continuous scanning mode of long-range ranging information is kept unchanged. Based on the vehicle orientation parameters in the attitude information and the target occurrence frequency distribution of each scanning sector within the preset time window, the target density prediction results of each scanning sector in the next monitoring cycle are predicted. The directional scanning weight of the near-range ranging information in each scanning sector is dynamically adjusted according to the target density prediction results to obtain attitude-compensated multi-source detection information.

[0021] By adopting the above technical solution, when the tilt angle value is less than or equal to a preset tilt angle threshold and the yaw rate is less than or equal to a preset yaw rate threshold, the motorcycle is determined to be in a straight-line condition. This clarifies that the vehicle is currently in a non-steering state, facilitating the stable execution of subsequent ranging strategies. Under this condition, maintaining the omnidirectional continuous scanning mode of long-range ranging information ensures continuous coverage of targets in all directions in the environment. Based on the vehicle orientation parameters in the attitude information and the target occurrence frequency distribution in each scanning sector within a preset time window, the target density prediction results for each scanning sector in the next monitoring cycle are predicted. This helps to identify spatial areas with potential target concentrations. Dynamically adjusting the directional scanning weight of short-range ranging information in each scanning sector based on the target density prediction results allows for the tilting allocation of ranging resources to high-density directions, ultimately obtaining attitude-compensated multi-source detection information, thereby improving the concentration and coverage efficiency of target recognition.

[0022] Preferably, based on the vehicle orientation parameters in the attitude information and the target occurrence frequency distribution of each scanning sector within a preset time window, the target density prediction results for each scanning sector in the next monitoring cycle are predicted, including:

[0023] Obtain real-time speed information and road curvature information for motorcycles;

[0024] The window length of the preset time window is determined based on real-time vehicle speed information and road curvature information, and the sampling time interval is determined based on the window length;

[0025] Based on the sampling time interval, time normalization is performed on the number of target detections in each scanning sector, and the weighted historical target occurrence frequency is calculated in combination with the echo signal quality factor;

[0026] Based on the correlation between the weighted historical target occurrence frequency and vehicle orientation parameters, the target occurrence trend value of each scanning sector is calculated;

[0027] Based on the target occurrence trend value, the probability of target occurrence in each scanning sector in the next monitoring period is estimated, and the potential target density prediction result of each scanning sector is obtained.

[0028] By adopting the above technical solution, real-time vehicle speed and road curvature information of motorcycles can be acquired. This information, combined with the vehicle's current motion state and path shape, provides basic parameter support for subsequent sampling strategy adjustments. The window length of the preset time window is determined based on the real-time vehicle speed and road curvature information, and the sampling time interval is determined based on the window length. This helps to dynamically adapt to the target monitoring rhythm under different driving conditions. Based on the sampling time interval, time normalization processing is performed on the number of target detections in each scanning sector, and the weighted historical target occurrence frequency is calculated in combination with the echo signal quality factor. This improves the stability and representativeness of frequency statistics under different driving environments. Based on the correlation between the weighted historical target occurrence frequency and the vehicle orientation parameter, the target occurrence trend value of each scanning sector is calculated, making the subsequent target prediction process directionally sensitive. The probability of target occurrence in the next monitoring cycle is estimated based on the target occurrence trend value, and the potential target density prediction result of each scanning sector is obtained. This enables dynamic prediction of the degree of target aggregation in the forward space and improves the directional adaptability of the perception strategy.

[0029] Preferably, the directional scanning weights of near-range ranging information in each scanning sector are dynamically adjusted based on the target density prediction results to obtain attitude-compensated multi-source detection information, including:

[0030] Obtain the target density value of each scanned sector from the target density prediction results;

[0031] The target density value is compared with a preset target density threshold to determine the priority of each scan sector;

[0032] The directional scanning weights of near-range ranging information in each scanning sector are adjusted by increasing or decreasing according to priority to obtain the adjusted directional scanning weights.

[0033] The near-range ranging information is weighted according to the adjusted directional scanning weights to obtain attitude-compensated multi-source detection information.

[0034] By adopting the above technical solution, the target density values ​​of each scanning sector can be obtained from the target density prediction results, which can clarify the distribution trend of potential targets in each spatial direction. The target density values ​​are compared with the preset target density threshold to determine the priority of each scanning sector, which helps to screen out the monitoring areas that need to be given priority. According to the priority, the directional scanning weight of the near-range information in each scanning sector is increased or decreased to obtain the adjusted directional scanning weight. This enables the dynamic redistribution of sensing resources and enhances the detection capability in high-density directions. The near-range information is weighted according to the adjusted directional scanning weight to obtain attitude-compensated multi-source detection information, thereby forming a detection result that integrates the current spatial target distribution characteristics and the vehicle's own attitude state, enhancing the directional adaptability and accuracy stability of target perception.

[0035] Preferably, the environmental visibility state is determined based on multi-source detection information. When detecting low visibility conditions, the weights of near-range and long-range ranging information in the multi-source detection information are updated during the data fusion process based on the echo intensity distribution, point cloud sparsity, and stability parameters in the multi-source detection information, to obtain the fused target location information and fused target velocity information, including:

[0036] Initial weight parameters are assigned to near-range and long-range ranging information based on echo intensity distribution;

[0037] The initial weight parameters are corrected based on the sparsity of the point cloud to obtain the corrected weight parameters;

[0038] The corrected weight parameters are normalized by the stability parameters to obtain the updated weight parameters.

[0039] During the data fusion process, updated weight parameters are applied to perform weighted fusion of near-range and long-range ranging information to obtain fused target location information and fused target velocity information.

[0040] By adopting the above technical solution, and assigning initial weight parameters to near-range and far-range ranging information based on echo intensity distribution, an initial evaluation model for measurement reliability can be established by combining the echo energy characteristics of each ranging information, thus providing a parameter basis for the subsequent fusion process. Correcting the initial weight parameters according to point cloud sparsity reflects the distribution density of spatial measurement data in different directions, and dynamically adjusts the reliability of the ranging information accordingly, ensuring a reasonable fusion contribution ratio in target edge areas or low-reflection areas. Normalizing the corrected weight parameters using stability parameters eliminates the impact of instantaneous fluctuations caused by environmental disturbances or equipment jitter from each ranging source, ensuring statistical consistency and temporal continuity of the final weights. Applying the updated weight parameters during data fusion, and weightedly fusing near-range and far-range ranging information, ensures both the accuracy of near-range targets and the continuous perception of far-range targets, ultimately obtaining fused target position and velocity information, thereby improving the overall integrity and stability of target detection.

[0041] Preferably, during the data fusion process, updated weight parameters are applied to perform weighted fusion of near-range and long-range ranging information to obtain fused target location information and fused target velocity information, including:

[0042] By multiplying the near-range and far-range ranging information by the corresponding updated weight parameters, we obtain the near-range weighted ranging information and the far-range weighted ranging information.

[0043] Consistency verification is performed on the near-range weighted ranging information and the far-range weighted ranging information. Data points with ranging deviations exceeding a preset threshold are removed to obtain the near-range weighted ranging information and the far-range weighted ranging information with consistency verification.

[0044] The near-range weighted ranging information from the consistency check is fused with the far-range weighted ranging information from the consistency check to obtain the fused ranging result;

[0045] The target's three-dimensional spatial position is calculated by combining the fused ranging results with attitude information, thus obtaining the fused target position information;

[0046] The displacement change rate of the target position information to be fused is calculated within a preset fusion period to obtain the target velocity information to be fused.

[0047] By adopting the above technical solution, dynamic weighted fusion of ranging information can be achieved even when ranging errors and confidence differences exist in multi-source ranging data. By updating the weight parameters and combining them with the actual confidence distribution of near-range and long-range ranging information, the accuracy and stability of the fusion results can be effectively improved. After weighting, outlier data points are removed through consistency checks to ensure relative consistency of the data participating in the fusion calculation, effectively suppressing error interference caused by factors such as temporary occlusion and reflection interference. In the fusion calculation stage, the advantages of each data source are combined to further enhance ranging accuracy, and three-dimensional position conversion is completed through attitude information, thereby obtaining more accurate spatial positioning results. Finally, velocity estimation is performed based on the rate of change of target position information within the fusion cycle, improving the real-time and continuous nature of target motion state recognition. The overall process achieves closed-loop optimization from data acquisition, weighting processing, consistency checks to fusion calculation, significantly improving the accuracy and stability of target detection around motorcycles in complex road environments, providing a more reliable perception basis for subsequent path planning and obstacle avoidance control.

[0048] Preferably, the calculation of target collision time and target distance based on target location information and target velocity information includes:

[0049] Based on the projection of the target location information and the motorcycle's current location information onto the direction of travel, the relative distance in the direction of travel is determined, and the relative distance information is obtained.

[0050] Based on the projection of the target speed information and the motorcycle's current speed information onto the direction of travel, the relative speed in the direction of travel is determined, and the relative speed information is obtained.

[0051] Time-series prediction is performed on relative distance and relative velocity information to obtain the predicted relative distance curve;

[0052] Under the constraint of the predicted relative distance curve, when the relative speed information indicates that the target is approaching the motorcycle, the moment when the predicted relative distance curve first reaches zero distance is determined as the target collision time; when the relative speed information does not indicate that the target is approaching the motorcycle, the target collision time is set to an invalid value.

[0053] The lateral offset is determined based on the difference in the projection of the target location information and the motorcycle's current location information in the lateral direction.

[0054] When the lateral offset does not exceed the preset lateral threshold, the relative distance information at the current moment is determined as the target distance; when the lateral offset exceeds the preset lateral threshold, the target distance is set to an invalid value.

[0055] By employing the aforementioned technical solution, the relative distance and relative speed between the target and the motorcycle can be accurately calculated based on the projection of the target's position and speed information onto the motorcycle's travel direction. Combined with time-series prediction methods, a predicted relative distance curve is generated, effectively assessing the approach trend in the future. When an approach trend is detected, the collision time can be clearly given, enhancing the forward-looking judgment capability of collision risk identification. If the target does not have an approach trend, it can be promptly determined that the target will not pose a collision threat under current conditions, improving the targeting of subsequent processing and the efficiency of resource allocation. Simultaneously, by calculating the offset in the lateral direction and setting effective distance conditions, it ensures that only targets truly located in front of the motorcycle are included in the distance judgment, eliminating misjudgments caused by interfering targets on the side. This helps improve the overall collision warning accuracy and enhances the motorcycle's active safety perception capability in dynamic road environments.

[0056] Preferably, the lateral offset is determined based on the difference in the projection of the target location information and the motorcycle's current location information in the lateral direction, including:

[0057] Project the target location information and the current location information of the motorcycle onto the horizontal coordinate axis perpendicular to the motorcycle's direction of travel to obtain the target's horizontal coordinate value and the motorcycle's horizontal coordinate value.

[0058] The initial lateral offset is obtained by calculating the difference between the lateral coordinates of the information target and the lateral coordinates of the information motorcycle.

[0059] The initial lateral offset of the information is corrected according to the attitude correction angle of the motorcycle's driving direction to obtain the lateral offset.

[0060] By employing the above technical solution, the target location information and the motorcycle's current location information can be accurately projected onto a lateral coordinate axis perpendicular to the motorcycle's direction of travel. This yields the target's lateral coordinate values ​​and the motorcycle's lateral coordinate values, respectively. An initial lateral offset is then obtained through difference calculations. Based on this, the initial lateral offset is corrected using the motorcycle's current driving posture, effectively eliminating coordinate offset errors caused by vehicle deflection angles and ensuring the accuracy of the lateral offset. The corrected lateral offset accurately reflects the actual positional difference of the target in the lateral direction of the road, enabling subsequent target selection, path determination, and risk identification processing logic to have higher spatial resolution capabilities, thereby improving the reliability and stability of the overall perception results.

[0061] Preferably, based on the warning trigger information, a helmet prompt instruction is generated, including:

[0062] The warning trigger information is matched with the preset prompt type mapping rules to determine the prompt category information;

[0063] Based on the prompt category information, the corresponding prompt content is retrieved from the prompt content library, and the prompt content sequence is converted into prompt signal parameters, which include prompt method, prompt intensity, and prompt duration.

[0064] The prompt method, prompt intensity, and prompt duration are uniformly encapsulated to obtain helmet prompt instructions.

[0065] By adopting the above technical solution, after the warning trigger information is generated, the corresponding warning category information can be quickly matched according to the preset warning type mapping rules, and the warning content corresponding to the warning category can be accurately called from the warning content library. It is then converted into warning signal parameters that include warning method, warning intensity, and warning duration, making the warning content executable and adjustable. On this basis, the above warning signal parameters are uniformly encapsulated to form a structured helmet warning instruction, enabling the warning content to be efficiently transmitted to the wearer through vibration, sound, or light signals, thereby achieving timely warning and risk communication for the driver and enhancing information response speed and execution efficiency.

[0066] In summary, this application includes at least one of the following beneficial technical effects:

[0067] 1. By acquiring synchronous detection information and combining it with the motorcycle's posture information to determine the driving conditions, the system can dynamically adjust the outer scanning frequency and range of long-distance ranging when cornering, while suppressing echo interference from the inner direction. This adapts to complex turning paths and solves the problem of traditional ranging methods being unable to balance directional flexibility. Maintaining continuous long-distance scanning and correcting the near-distance directional weights during straight-line driving helps improve the detection accuracy of targets directly in front. Furthermore, by combining echo intensity distribution, point cloud sparsity, and stability parameters in low-visibility environments, the system dynamically corrects the fusion weights of near and far-distance ranging information, overcoming the problem of decreased image recognition accuracy under adverse weather conditions. Finally, based on the fused target position and speed information, the system accurately determines the collision time and target distance and generates helmet warning commands, achieving more realistic warning control and effectively improving riding safety.

[0068] 2. By obtaining the target density values ​​of each scanning sector from the target density prediction results, the distribution trend of potential targets in each spatial direction can be clarified. By comparing the target density values ​​with the preset target density threshold, the priority of each scanning sector can be determined, which helps to screen out the monitoring areas that need to be given priority. According to the priority, the directional scanning weight of the near-range information in each scanning sector is increased or decreased to obtain the adjusted directional scanning weight. This enables the dynamic redistribution of sensing resources and strengthens the detection capability in high-density directions. The near-range information is weighted according to the adjusted directional scanning weight to obtain attitude-compensated multi-source detection information, thereby forming a detection result that integrates the current spatial target distribution characteristics and the vehicle's own attitude state, enhancing the directional adaptability and accuracy stability of target perception.

[0069] 3. After the warning is triggered, the system can quickly match the corresponding warning category information according to the preset warning type mapping rules, and accurately call the warning content corresponding to the warning category from the warning content library. It then converts the warning content into warning signal parameters that include the warning method, warning intensity, and warning duration, making the warning content executable and adjustable. On this basis, the above warning signal parameters are uniformly encapsulated to form a structured helmet warning instruction, enabling the warning content to be efficiently transmitted to the wearer through vibration, sound, or light signals. This achieves timely warning and risk communication for the driver, enhancing information response speed and execution efficiency. Attached Figure Description

[0070] Figure 1 This is a flowchart of a multi-source ranging fusion collision warning method for motorcycle helmets according to one embodiment of this application.

[0071] Figure 2 This is a flowchart illustrating the implementation of step S20 in a multi-source ranging fusion collision warning method for a motorcycle helmet according to an embodiment of this application.

[0072] Figure 3 This is another implementation flowchart of step S20 in a multi-source ranging fusion collision warning method for motorcycle helmets according to one embodiment of this application;

[0073] Figure 4 This is a flowchart illustrating the implementation of step S205 in a multi-source ranging fusion collision warning method for a motorcycle helmet according to an embodiment of this application.

[0074] Figure 5 This is another implementation flowchart of step S205 in a multi-source ranging fusion collision warning method for motorcycle helmets according to an embodiment of this application. Detailed Implementation

[0075] The present application will be further described in detail below with reference to the accompanying drawings.

[0076] In one embodiment, such as Figure 1 As shown, this application discloses a multi-source ranging fusion collision warning method for motorcycle helmets, which specifically includes the following steps:

[0077] S10: Obtain synchronization detection information, which includes near-range ranging information, far-range ranging information, and attitude information.

[0078] Specifically, when acquiring synchronous detection information, near-range ranging information refers to the time-series numerical set of the straight-line distance between the target and the motorcycle reference point within a near-range area. The near-range area refers to a distance interval not exceeding a preset near-range threshold, such as 0.2 meters to 5 meters. The target refers to an object located around the motorcycle that can be identified during the ranging process, such as a vehicle, pedestrian, or fixed obstacle. The motorcycle reference point refers to the geometric reference position of the vehicle body used for distance calculation, such as the geometric center of the vehicle body or the center of the front axle. The process of acquiring near-range ranging information is as follows: a detection signal is emitted within the near-range area and the emission timestamp is recorded; the echo signal from the target is received and the reception timestamp is recorded; the propagation time difference between emission and reception is calculated and converted into a distance value based on the propagation speed of the propagation medium; and the continuously sampled distance values ​​are arranged in chronological order to form near-range ranging information. Long-range ranging information refers to the time-series numerical set of the straight-line distance between the target and the motorcycle reference point within a long-range area. The long-range area refers to a distance interval greater than the near-range threshold but not exceeding a preset long-range threshold, such as 5 meters to 60 meters. The process of acquiring long-range ranging information is as follows: A detection signal is emitted within a long distance range, or the long-range parallax is acquired and the corresponding timestamp is recorded. The target distance is obtained by calculating the phase change caused by the round-trip propagation of the signal, or by calculating the trigonometric geometric quantities corresponding to the parallax angle and baseline length. The continuously sampled distance values ​​are arranged in chronological order to form long-range ranging information. Attitude information refers to a time-series set describing the motorcycle's attitude angles and angular velocities in space, including pitch angle, roll angle, and yaw angle, and their corresponding angular velocities. The process of acquiring attitude information is as follows: Triaxial angular velocities and triaxial accelerations are collected and timestamps are recorded in each sampling period. The angle increment is obtained by integrating the angular velocity within the sampling period. Drift correction is performed using the direction of the gravitational component in the acceleration. The pitch angle, roll angle, and yaw angle are calculated. The angle and angular velocity quantities are arranged in chronological order to form attitude information. To generate synchronous detection information, timestamps are added to the near-range, long-range, and attitude information respectively and unified to the same time reference. When the sampling rates of the three types of information are inconsistent, linear interpolation or nearest neighbor interpolation is used to register them on a common time axis, so that a set of records containing near-range, long-range, and attitude information under the same time index is formed. The sequence of records arranged in ascending order of time is the synchronous detection information.

[0079] S20: Determine the motorcycle's driving condition based on the attitude information in the synchronous detection information. When the driving condition is cornering, adjust the scanning frequency and range of the long-distance ranging information in the outer direction and suppress the echo in the inner direction. When the driving condition is straight, maintain the continuous scanning of the long-distance ranging information and perform directional correction on the scanning weight of the short-distance ranging information to obtain multi-source detection information with attitude compensation.

[0080] Specifically, when determining the motorcycle's driving condition based on the attitude information in the synchronous detection information, the pitch angle, roll angle, and yaw angle in the attitude information are used as the basis for judgment. The pitch angle is the angle between the motorcycle's longitudinal axis and the horizontal plane, the roll angle is the angle between the motorcycle's lateral axis and the horizontal plane, and the yaw angle is the pointing angle of the motorcycle's longitudinal axis on the horizontal plane. First, the absolute value of the roll angle is calculated and compared with a preset roll angle threshold. At the same time, the yaw angle change rate is calculated. The yaw angle change rate is the amount of change of the yaw angle per unit time. When the absolute value of the roll angle is greater than the preset roll angle threshold and the yaw angle change rate is not equal to zero, it is judged as a cornering condition. When the absolute value of the roll angle is not greater than the preset roll angle threshold and the yaw angle change rate is close to zero, it is judged as a straight-line condition. In cornering conditions, the outer and inner directions are determined by the sign of the rate of change of the heading angle. When the rate of change of the heading angle is positive, the outer direction is the left side of the motorcycle; when the rate of change of the heading angle is negative, the outer direction is the right side of the motorcycle. The scanning frequency of the outer direction in the long-distance ranging information is adjusted to the upper limit of the preset outer scanning frequency, expanding the scanning angle coverage of the outer direction to the preset outer scanning angle range. At the same time, the scanning frequency of the inner direction is reduced to the lower limit of the preset inner scanning frequency, shrinking the scanning angle coverage of the inner direction to the preset inner scanning angle range. An echo suppression threshold is set in the inner direction to block low-amplitude echoes and multipath echoes. In straight-line conditions, the scanning frequency of the long-distance ranging information is maintained at the preset straight-line scanning frequency, and the scanning angle coverage is maintained at the preset straight-line scanning angle range to maintain continuous scanning in all directions. When performing directional correction on near-range ranging information, the angle between the target and the longitudinal axis of the motorcycle is used as the azimuth angle. Near-range ranging information with an azimuth angle close to zero is assigned a higher scanning weight value, while near-range ranging information with an azimuth angle close to the side is assigned a lower scanning weight value. The assigned scanning weight values ​​are normalized under each time index to ensure that the weight and value range are consistent. After the outer and inner adjustments of the far-range ranging information and the directional correction of the near-range ranging information, they are combined with the attitude information under the corresponding time index to form attitude-compensated multi-source detection information.

[0081] S30: Determine the environmental visibility status based on multi-source detection information. When detecting low visibility conditions, update the weights of near-range and long-range ranging information in the multi-source detection information during the data fusion process based on the echo intensity distribution, point cloud sparsity, and stability parameters in the multi-source detection information to obtain the fused target location information and fused target velocity information.

[0082] Specifically, when judging the environmental visibility state based on multi-source detection information, the attenuation degree of optical or radar signals during transmission is determined by analyzing the amplitude variation characteristics of the echo intensity distribution in the multi-source detection information. Echo intensity distribution refers to the statistical distribution of echo signal intensity in each ranging direction for near-range and far-range ranging information. The spatial distribution density of point cloud sparsity in the multi-source detection information reflects the sparseness or density of target reflection points within the ranging range. Point cloud sparsity refers to the ratio of the number of effective reflection points generated by near-range and far-range ranging information to the number of theoretical sampling points within a unit ranging range. Stability is extracted from the multi-source detection information. The time-varying curves of parameters are used to evaluate the continuity of multi-frame detection results after attitude compensation. The stability parameter is a stability coefficient calculated from the echo intensity and position coordinate fluctuation range of near-range and far-range ranging information in the same ranging direction of multiple consecutive frames. When detecting low visibility conditions, the weight ratio of near-range and far-range ranging information in the data fusion process is adjusted according to the echo intensity distribution, point cloud sparsity and stability parameter. The fused target position information and fused target velocity information are obtained through weighted fusion calculation. The target position information refers to the spatial position of the fused target in the three-dimensional coordinate system, and the target velocity information refers to the displacement change rate of the fused target per unit time.

[0083] S40: Calculate the target collision time and target distance based on the target position information and target speed information, compare the target collision time with a preset collision time threshold, and compare the target distance with a preset distance threshold. When either condition is met, generate a warning trigger information.

[0084] Specifically, when determining the environmental visibility status based on multi-source detection information, the echo intensity distribution in the multi-source detection information is analyzed. The echo intensity distribution refers to the statistical distribution of echo signal intensity in each ranging direction of near-range and far-range ranging information. By comparing the amplitude changes of echo intensity in different directions, the attenuation characteristics of the signal during transmission are determined. The sparsity of the point cloud in the multi-source detection information is analyzed. The point cloud sparsity refers to the ratio of the number of effective reflection points generated by near-range and far-range ranging information to the number of theoretical sampling points in a unit ranging space. This ratio reflects the density or sparseness of target reflection points within the ranging range. The stability parameter in the multi-source detection information is extracted. The stability parameter is the stability coefficient calculated by considering the fluctuation range of near-range and far-range ranging information in the same ranging direction in multiple consecutive time indices in terms of echo intensity and position coordinates. By analyzing the changing trend of the stability coefficient, the continuity of multi-frame detection results after attitude compensation is evaluated. When the environment is determined to be under low visibility conditions, initial weight parameters are first assigned to near-range and far-range ranging information based on the echo intensity distribution. Then, the initial weight parameters are corrected according to the point cloud sparsity to obtain corrected weight parameters. Subsequently, the corrected weight parameters are normalized using stability parameters to obtain updated weight parameters. During the data fusion process, the near-range and far-range ranging information are multiplied by the corresponding updated weight parameters and weighted. The weighted fusion result is then calculated after being aligned with the time index. Combined with attitude information, the result is converted into the target position information for fusion. Within a preset fusion period, the displacement change rate is calculated based on the fused target position information to obtain the fused target velocity information.

[0085] S50: Generate helmet prompt instructions based on the warning trigger information.

[0086] Specifically, the warning trigger information is matched with preset prompt type mapping rules. The prompt type mapping rules refer to a set of rules established in advance to determine the corresponding prompt type information according to the category, urgency and triggering conditions of different warning trigger information. After matching, the prompt type information is determined. The prompt type information refers to the parameter identifier used to identify the specific type of helmet prompt, including voice prompts, light prompts, vibration prompts, etc. According to the prompt type information, the corresponding prompt content is called from the prompt content library. The prompt content library refers to the data set of prompt voice files, light flashing patterns or vibration patterns, etc., which are stored according to the prompt type information. The prompt content sequence is converted into prompt signal parameters. The prompt signal parameters refer to the control parameters used to drive the helmet to output prompt content, including prompt mode, prompt intensity and prompt duration. The prompt mode, prompt intensity and prompt duration are uniformly encapsulated to obtain the helmet prompt command.

[0087] In one embodiment, such as Figure 2 As shown, in step S20, the motorcycle's driving condition is determined based on the attitude information in the synchronous detection information. When the driving condition is cornering, the scanning frequency and range of the long-distance ranging information in the outer direction are adjusted and the echo in the inner direction is suppressed. This includes:

[0088] S201: Obtain the tilt angle and yaw rate from the attitude information.

[0089] In this embodiment, the roll angle value refers to the instantaneous angular value of the roll angle, and the yaw rate refers to the angular rate of the heading angle about the vertical axis.

[0090] Specifically, firstly, three-axis angular velocity data and three-axis acceleration data are read on the time axis corresponding to the attitude information. A vehicle coordinate system is established, with the forward direction defined as the x-axis, the left side of the vehicle as the y-axis, and the vertical upward direction as the z-axis. Then, zero-bias calibration and denoising are performed in each sampling period. Zero-bias calibration refers to estimating the static bias of the angular velocity channel during low-dynamic periods and subtracting the bias value in subsequent sampling. Denoising refers to applying a first-order low-pass filter to the angular velocity data and acceleration data respectively to suppress high-frequency jitter. After preprocessing, the gravity direction vector is estimated using the acceleration data, and the gravity direction is... The angle between the y-axis and z-axis in the plane is calculated to obtain the roll angle, thus yielding the pitch angle value. In continuous sampling, a complementary approach is used to fuse the roll angle increment obtained by integrating the angular velocity with the roll angle obtained by calculating the direction of gravity to reduce drift. Simultaneously, within the same sampling period, the component around the z-axis is read from the angular velocity data as the heading angle angular velocity. After removing the zero bias and aligning the coordinate system, this component is output as the yaw rate angular velocity. Finally, the pitch angle value and yaw rate angular velocity are aligned to the unified time base of the synchronous detection information according to the timestamp of the attitude information for use in subsequent steps.

[0091] S202: When the tilt angle value is greater than the preset tilt angle threshold and the yaw rate is greater than the preset yaw rate threshold, the motorcycle is determined to be in a cornering condition.

[0092] Specifically, after obtaining the tilt angle and yaw rate, the tilt angle is compared with a preset tilt angle threshold. The preset tilt angle threshold is a numerical boundary set based on the stability analysis of the motorcycle at different tilt angles to determine whether it has entered a significant tilt state. At the same time, the yaw rate is compared with a preset yaw rate threshold. The preset yaw rate threshold is a numerical boundary set based on the lateral rotation rate change characteristics of the motorcycle during turning to determine whether it has entered a significant lateral rotation state. When both the tilt angle and yaw rate values ​​are greater than the preset tilt angle threshold, the motorcycle is confirmed to be in a cornering condition based on the simultaneous satisfaction of the above two judgment conditions, thereby completing the identification of the current motorcycle driving state.

[0093] S203: In the case of cornering, the scanning frequency of the outer direction in the long-distance ranging information is increased to the upper limit of the preset scanning frequency, and the scanning range of the outer direction is expanded to the preset expansion range. At the same time, the scanning frequency of the inner direction is reduced to the lower limit of the preset scanning frequency, and the scanning range of the inner direction is shrunk to the preset shrinkage range, so as to reduce the interference of the inner direction echo.

[0094] Specifically, when the motorcycle is determined to be in a cornering condition, a dynamic parameter allocation operation is performed based on the directional differences in long-distance ranging information. The outer direction refers to the ranging direction relative to the center point of the motorcycle's curve on the outer edge of the vehicle body, while the inner direction refers to the ranging direction relative to the center point of the motorcycle's curve on the inner edge of the vehicle body. The preset upper limit of the scanning frequency is the highest scanning frequency threshold set to improve the accuracy of long-distance obstacle detection, the preset lower limit of the scanning frequency is the lowest scanning frequency threshold set to reduce redundant invalid directional echoes, the preset extended range refers to the maximum angular range covered by the outer direction ranging signal, and the preset contraction range is... The minimum angular range after the inward direction ranging signal is scanned and compressed is used. Based on the tilt angle and yaw rate obtained from the attitude information, the cornering condition is determined. The scanning frequency of the corresponding outward direction in the long-distance ranging information is increased to the upper limit of the preset scanning frequency, and the scanning range of the outward direction is expanded to the preset expansion range. At the same time, the scanning frequency of the corresponding inward direction is reduced to the lower limit of the preset scanning frequency, and the scanning range of the inward direction is shrunk to the preset shrinkage range. Finally, the sampling center of gravity of the long-distance ranging information is shifted to the outward direction to avoid the influence of echo interference in the inward direction. Echo interference refers to the ranging error signal generated by non-target obstacles or nearby reflection sources.

[0095] In one embodiment, such as Figure 3 As shown, in step S20, when the driving condition is a straight line, continuous scanning of long-range ranging information is maintained, and the scanning weight of short-range ranging information is corrected directionally to obtain attitude-compensated multi-source detection information, including:

[0096] S204: When the tilt angle value is less than or equal to the preset tilt angle threshold and the yaw rate is less than or equal to the preset yaw rate threshold, the motorcycle is determined to be in a straight-line condition.

[0097] Specifically, when the tilt angle value is less than or equal to the preset tilt angle threshold and the yaw rate is less than or equal to the preset yaw rate threshold, the working condition identification and judgment operation is performed. If both of the above conditions are met, the current state of the motorcycle is determined to be a straight working condition. A straight working condition means that the motorcycle has not undergone significant directional changes in the current driving state, the vehicle attitude is stable, and it is not in the process of cornering.

[0098] S205: Under straight-line conditions, the omnidirectional continuous scanning mode of long-range ranging information is kept unchanged. Based on the vehicle orientation parameters in the attitude information and the target occurrence frequency distribution of each scanning sector within the preset time window, the target density prediction results of each scanning sector in the next monitoring cycle are predicted. The directional scanning weight of the short-range ranging information in each scanning sector is dynamically adjusted according to the target density prediction results to obtain attitude-compensated multi-source detection information.

[0099] Specifically, under straight-line conditions, the omnidirectional continuous scanning mode for long-distance ranging information is maintained to acquire real-time motorcycle speed and road curvature information. Real-time speed refers to the speed value of the motorcycle in its current motion state, and road curvature refers to the degree of curvature change of the motorcycle's current travel path. The window length of the preset time window is determined based on the real-time speed and road curvature information, and the sampling time interval is determined based on the window length. The sampling time interval is used to perform unified time reference processing on the target detection data. Subsequently, based on the sampling time interval, time normalization processing is performed on the number of target detections in each scanning sector. Time normalization processing refers to converting the number of target detections into a detection frequency per unit time. The target detection frequency is then weighted and corrected by combining the echo signal quality factor corresponding to each scanning sector to obtain the weighted historical target occurrence frequency. The weighted historical target occurrence frequency is used to reflect the confidence level of target occurrence in different scanning sectors. Then, based on the weighted historical target occurrence frequency... The correlation between target frequency and vehicle orientation parameters is established, and the target occurrence trend value for each scanning sector is calculated. The target occurrence trend value refers to the trend index of the weighted target frequency changing with the vehicle orientation angle. Based on the target occurrence trend value, the probability of target occurrence in each scanning sector in the next monitoring cycle is estimated, and the potential target density prediction result for each scanning sector is obtained. The target density prediction result represents the probability distribution of possible target occurrences in each scanning sector in the future monitoring cycle. Finally, the directional scanning weight of near-range information in each scanning sector is dynamically adjusted according to the target density prediction result. The directional scanning weight refers to the numerical parameter that controls the scanning priority of near-range information in different scanning directions. The weighting is adjusted according to the angle relationship between the potential target density prediction result and the vehicle orientation parameter, and the directional scanning weight is embedded in the scanning command of the near-range information. In conjunction with the vehicle orientation parameter, attitude compensation processing is performed to obtain attitude-compensated multi-source detection information.

[0100] In one embodiment, such as Figure 4 As shown, in step S205, based on the vehicle orientation parameters in the attitude information and the target occurrence frequency distribution of each scanning sector within a preset time window, the target density prediction result for each scanning sector in the next monitoring cycle is predicted, including:

[0101] S2051: Obtain real-time speed information of the motorcycle and road curvature information.

[0102] Specifically, the speed of the motorcycle at the current moment is obtained by the speed sensor installed on the motorcycle. This speed value is used as the real-time speed information and recorded in chronological order to form a speed data sequence. At the same time, based on the motorcycle's location and the path geometry information provided by the navigation system, the curvature value of the current road is calculated by using the road centerline data set in the geographic information system and the rate of change of the tangent of the location point on the road centerline. This curvature value is used as the road curvature information, which refers to the measure of directional change within a unit path length and is used to reflect the degree of road curvature.

[0103] S2052: Determine the window length of the preset time window based on real-time vehicle speed information and road curvature information, and determine the sampling time interval based on the window length.

[0104] Specifically, the current vehicle speed value from the real-time vehicle speed information and the current curvature value from the road curvature information are extracted separately. The window length of the preset time window is calculated based on the inverse relationship between the product of the current vehicle speed value and the road curvature value. The window length refers to the data analysis cycle for subsequent continuous acquisition of detection information. After obtaining the window length, the sampling time interval is determined according to the conversion rule between the window length and the data continuity requirement. The sampling time interval refers to the time interval between continuous sampling actions within the time window, which is used to control the uniform distribution of detection data on the time axis.

[0105] S2053: Based on the sampling time interval, perform time normalization processing on the number of target detections in each scanning sector, and calculate the weighted historical target occurrence frequency in combination with the echo signal quality factor.

[0106] Specifically, based on the sampling time interval, the number of times a target is detected in each scanning sector is counted within a preset time window. The number of target detections in each scanning sector is divided by the corresponding sampling time interval to obtain the target detection frequency per unit time for each scanning sector. The target detection frequency is used to reflect the density of target occurrence in the scanning sector. Further, the echo signal quality factor corresponding to the target is extracted. The echo signal quality factor is a numerical quantitative index used to represent the reliability of the target detection result. It represents the ratio of the effective signal strength to the interference formed by the laser or radar signal during the target reflection process. The target detection frequency and the echo signal quality factor are weighted according to the weighted average rule to obtain the weighted historical target occurrence frequency used to describe the historical occurrence intensity of the target in each scanning sector.

[0107] S2054: Calculate the target occurrence trend value for each scanning sector based on the correlation between the weighted historical target occurrence frequency and the vehicle orientation parameter.

[0108] Specifically, the weighted historical target occurrence frequency of each scanning sector is extracted and combined with the vehicle orientation parameter information at the corresponding time. The vehicle orientation parameter refers to the attitude angle value reflecting the current spatial orientation of the motorcycle to quantify the orientation state of the motorcycle during driving. The angle direction of each scanning sector is aligned and matched with the vehicle orientation parameter to calculate the numerical distribution of target occurrence frequency under different vehicle orientations. Based on this distribution relationship, the concentration degree and trend of target occurrence frequency in each scanning sector near the current vehicle orientation are statistically analyzed. Methods such as linear fitting or difference trend analysis are used to obtain the target occurrence trend value used to characterize the target activity tendency of each scanning sector. This target occurrence trend value is used to reflect the strength of the target activity trend in spatial distribution relative to the vehicle orientation.

[0109] S2055: Estimate the probability of target occurrence in each scanning sector in the next monitoring period based on the target occurrence trend value, and obtain the potential target density prediction result for each scanning sector.

[0110] Specifically, the target occurrence trend value refers to the weighted sum of the number of target detections and the echo signal quality factor for each scanning sector within a preset time window, after time normalization processing. This reflects the direction and intensity of target activity frequency changes in each sector. The target occurrence trend values ​​for each scanning sector are arranged according to the scan sequence number to construct a trend sequence. A continuous time period sliding window is set, and the numerical difference between trend values ​​in the same sector within adjacent windows is calculated to obtain the trend change rate. The trend change rate refers to the magnitude of the increase or decrease in the trend value per unit time. A weighted summation is performed on the trend value and the trend change rate to form the trend extension value. The trend extension value is used to characterize the next monitoring cycle. The estimated activity level of a target in the corresponding sector is transformed by substituting the trend extension value into a normalized mapping rule. The normalized mapping rule refers to the function mapping relationship that linearly compresses the trend extension value to the closed interval [0,1] based on the maximum and minimum boundary range of the trend extension value in historical samples. This is used to unify the probabilistic expression scale of target activity. The target occurrence probability is generated based on the mapping result. The target occurrence probability refers to the probability value of detecting a target in the corresponding scanned sector in the next monitoring period. Finally, the target occurrence probabilities corresponding to all scanned sectors are arranged according to spatial distribution to form the potential target density prediction result, which is used to express the spatial distribution of high-risk targets that may gather in the environment.

[0111] In one embodiment, such as Figure 5 As shown, in step S205, the directional scanning weights of the near-range ranging information in each scanning sector are dynamically adjusted based on the target density prediction results to obtain attitude-compensated multi-source detection information, including:

[0112] S2056: Obtain the target density value of each scanned sector from the target density prediction results.

[0113] Specifically, when obtaining the target density value of each scanning sector from the target density prediction results, the target occurrence frequency of each scanning sector is first calculated based on the number of target detections and the sampling time interval within a preset time window. Then, the target occurrence frequency is weighted by combining the echo signal quality factor of each scanning sector to obtain the weighted target occurrence frequency. Subsequently, the weighted target occurrence frequency is fitted by angle weighting based on the angle relationship between the vehicle orientation parameter and the scanning sector direction to generate a target occurrence trend value that represents the target distribution trend. Then, the target occurrence trend value and the trend slope of each scanning sector on the time axis are used to estimate the target occurrence probability of each scanning sector in the next monitoring period. Finally, the target occurrence probability is used as the target density value.

[0114] S2057: Compare the target density value with the preset target density threshold to determine the priority of each scan sector.

[0115] Specifically, when comparing the target density value with a preset target density threshold to determine the priority of each scanning sector, the target density value of each scanning sector is first extracted. This target density value represents the probability that a target exists in the corresponding scanning sector in the next monitoring cycle. The preset target density threshold is a density judgment benchmark value manually set based on historical monitoring data and task requirements, used to divide the attention priority level of scanning sectors. The target density value of each scanning sector is compared with the target density threshold one by one. If the target density value is higher than the target density threshold, the corresponding scanning sector is marked as a high-priority sector; otherwise, it is marked as a low-priority sector. The priority information of all scanning sectors is arranged in the order of scanning sectors to form a priority judgment sequence. This priority judgment sequence is used for the dynamic adjustment of directional scanning weights in subsequent operations.

[0116] S2058: Adjust the directional scanning weights of near-range ranging information in each scanning sector according to priority to obtain the adjusted directional scanning weights.

[0117] Specifically, when adjusting the directional scanning weights of near-range ranging information in each scanning sector according to priority, a correspondence is first established between each scanning sector in the priority determination sequence and its corresponding current directional scanning weight. The directional scanning weight refers to the parameter weight used to adjust the resource allocation ratio of near-range ranging information in different spatial directions, which is used to highlight the ranging frequency or accuracy in certain directions. For scanning sectors determined to be of high priority, the directional scanning weight value is increased to the maximum extent not exceeding the upper limit of the directional scanning weight, thereby enhancing the near-range ranging frequency and signal accuracy in the corresponding direction. For scanning sectors determined to be of low priority, the directional scanning weight value is decreased to the minimum extent not lower than the lower limit of the directional scanning weight, thereby compressing the near-range ranging resource allocation in the corresponding direction. After the directional scanning weights of all scanning sectors are adjusted, they are arranged in sequence to form an adjusted directional scanning weight sequence. This directional scanning weight sequence, together with the far-range ranging information, constitutes attitude-compensated multi-source detection information.

[0118] S2059: The near-range ranging information is weighted according to the adjusted directional scanning weights to obtain attitude-compensated multi-source detection information.

[0119] Specifically, when weighting the near-range ranging information according to the adjusted directional scanning weights, firstly, a weighting factor is set for each scanning sector based on its directional scanning weight value. The weighting factor is a coefficient used to adjust the numerical influence of the target echo signal in the multi-source fusion process. The target echo intensity, echo stability, and distance confidence are differentiated according to the directional scanning weights. Within each scanning sector, the signal amplitude is adjusted and the confidence level is corrected for the corresponding near-range ranging information using the weighting factor, forming the weighted echo results for each scanning sector. The weighted echo results are then realigned according to the vehicle orientation parameters, and spatial orientation consistency correction is performed. After attitude compensation processing, the panoramic scanning results from the long-range ranging information are combined with the multi-source information fusion based on temporal alignment and spatial reconstruction methods, ultimately obtaining attitude-compensated multi-source detection information.

[0120] In one embodiment, in step S30, i.e., determining the environmental visibility state based on multi-source detection information, when detecting low visibility conditions, the weights of near-range and long-range ranging information in the multi-source detection information during the data fusion process are updated based on the echo intensity distribution, point cloud sparsity, and stability parameters in the multi-source detection information, to obtain the fused target location information and fused target velocity information, including:

[0121] S301: Assign initial weight parameters to near-range and far-range ranging information based on echo intensity distribution.

[0122] Specifically, when assigning initial weight parameters to near-range and far-range ranging information based on echo intensity distribution, the echo intensity data from near-range and far-range ranging information collected by multiple ranging channels within the same monitoring period are first acquired. The echo intensity data is then categorized according to the corresponding scanning sectors, constructing echo intensity distribution curves for each scanning sector. Echo intensity distribution refers to the statistical distribution result reflecting the changing trend of echo signal intensity in each sector within a spatial range, used to measure the target's reflectivity and echo stability level within that sector. Based on the echo intensity distribution trend of each sector, the corresponding normalized intensity factor is calculated using the maximum value normalization method. The normalized intensity factor is a numerical parameter that maps the echo intensity under different ranging methods to a unified standard scale range, used to weigh the information reliability of different ranging methods. After normalization, the initial weight parameters for near-range and far-range ranging information in each scanning sector are assigned according to the distribution ratio of the normalized intensity factor for each ranging method in each scanning sector.

[0123] Furthermore, the maximum value normalization method involves dividing each value in a set of original data by the maximum value in that set, so that the normalized result falls within the range of 0 and 1. This eliminates the influence between different units or numerical scales, facilitating unified comparison and weighted processing of echo intensity between different ranging methods or different sensing channels.

[0124] S302: Correct the initial weight parameters based on the sparsity of the point cloud to obtain the corrected weight parameters.

[0125] Specifically, point cloud data for each scanning sector is extracted from both near-range and long-range ranging information. Within each scanning sector, equal-sized three-dimensional mesh voxel units are constructed, and the number of laser echo points in each voxel is counted. Based on the total number of laser echo points contained in non-empty voxels across all voxel units and the number of voxel units, the average point density per unit voxel is calculated. The reciprocal of this average point density is used as the point cloud sparsity value for the corresponding scanning sector. A larger point cloud sparsity value indicates a sparser ranging result. The point cloud sparsity values ​​for all scanning sectors are normalized to their maximum values ​​to obtain a normalized sparsity factor. The initial weight parameters are then weighted and corrected based on the normalized sparsity factor for each scanning sector. A larger normalized sparsity factor results in a smaller correction magnitude for the corresponding initial weight parameters, thus obtaining a correction weight parameter inversely proportional to the sparsity.

[0126] S303: Normalize the corrected weight parameters using stability parameters to obtain the updated weight parameters.

[0127] Specifically, near-range and far-range ranging information for each scanning sector is collected within a preset time window. Based on the ranging information arranged in chronological order, a sliding window is used to calculate the standard deviation of the ranging values ​​for each scanning sector. The standard deviation is used as the stability parameter for the corresponding scanning sector. The stability parameter is a statistical measure of the fluctuation range of the ranging information within the preset time window. The smaller the value, the more stable the ranging process. Then, the stability parameters of each scanning sector are normalized using the minimum value normalization method. The minimum value normalization method involves subtracting the minimum value among all stability parameters from a certain stability parameter and then dividing by the difference between the maximum and minimum values. Next, the corrected weight parameters of each scanning sector, calculated based on the point cloud sparsity, are multiplied by the corresponding normalized stability parameters to obtain the updated weight parameters for each scanning sector.

[0128] S304: During the data fusion process, the updated weight parameters are applied to perform weighted fusion of near-range and long-range ranging information to obtain fused target location information and fused target velocity information.

[0129] Specifically, spatial coordinate values ​​are extracted from near-range and far-range ranging information at the same timestamp. An arithmetic weighted average is then used to multiply these spatial coordinate values ​​with the updated weight parameters to obtain weighted coordinate values ​​for each data source. The sum of these weighted coordinate values ​​is then divided by the sum of the updated weight parameters to obtain the fused target location information. Similarly, velocity vector information is extracted from near-range and far-range ranging information at the same timestamp. These velocity vectors are then multiplied with the updated weight parameters using the same weighting method. The two weighted velocity values ​​are then summed and divided by the sum of the updated weight parameters to obtain the fused target velocity information.

[0130] In one embodiment, in step S304, i.e., during the data fusion process, the updated weight parameters are applied to perform weighted fusion of near-range ranging information and far-range ranging information to obtain fused target position information and fused target velocity information, including:

[0131] S3041: Multiply the near-range and long-range ranging information by the corresponding updated weight parameters to obtain the near-range weighted ranging information and the long-range weighted ranging information.

[0132] Specifically, at each preset sampling time point, the ranging value corresponding to the target position is extracted from the near-range ranging information, and the ranging value corresponding to the same target position is extracted from the far-range ranging information. These are marked as the near-range original ranging value and the far-range original ranging value, respectively. Combined with the updated weight parameter corresponding to that time point, a weighted operation is performed. The near-range weighted ranging information is obtained by multiplying the near-range original ranging value by the corresponding updated weight parameter, and the far-range weighted ranging information is obtained by multiplying the far-range original ranging value by the corresponding updated weight parameter.

[0133] S3042: Perform consistency verification on the near-range weighted ranging information and the far-range weighted ranging information, remove data points whose ranging deviation exceeds the preset threshold, and obtain the near-range weighted ranging information and the far-range weighted ranging information with consistency verification.

[0134] Specifically, at each preset sampling time point, the corresponding near-range weighted ranging information and far-range weighted ranging information are acquired. The difference between the two is taken as the ranging deviation. The ranging deviation is compared point by point with a preset ranging consistency threshold. When the ranging deviation is greater than the preset ranging consistency threshold, the near-range weighted ranging information and far-range weighted ranging information corresponding to that time point are removed together. The remaining data points that meet the consistency conditions are retained and constitute the near-range weighted ranging information and far-range weighted ranging information for consistency verification, respectively. The ranging consistency threshold refers to the maximum range of ranging differences allowed in the multi-source ranging fusion process, which is used to remove abnormal ranging data caused by environmental interference or reflection anomalies.

[0135] S3043: The near-range weighted ranging information from the consistency check is fused with the far-range weighted ranging information from the consistency check to obtain the fused ranging result.

[0136] Specifically, for the same sampling time point, the near-range weighted ranging information and the far-range weighted ranging information of consistency verification are extracted. Based on their respective updated weight parameters as weighting factors, the near-range weighted ranging information and the far-range weighted ranging information of consistency verification are weighted and summed respectively. The weighted summation result is then normalized and weighted to obtain the fused ranging result for the corresponding time point. The fused ranging result refers to the estimated distance value of a single target determined after fusing different ranging sources.

[0137] S3044: Based on the fused ranging results and attitude information, the three-dimensional spatial position of the target is calculated to obtain the fused target position information.

[0138] Specifically, the fusion ranging results and the corresponding attitude information are collected. The attitude information refers to the directional angle parameters representing the sensor's orientation in three-dimensional space, including pitch angle, yaw angle, and roll angle. A three-dimensional rotation matrix is ​​constructed by parsing the attitude information, and then the distance value corresponding to the fusion ranging result is transformed from the sensor's own coordinate system to a unified geographic coordinate system or track coordinate system. Combined with the real-time sensor installation height and horizontal pitch relationship, the target's position coordinates in three-dimensional space are calculated. The position coordinates include spatial horizontal coordinates, vertical coordinates, and height coordinates. Finally, the fused target position information is formed for subsequent target trajectory tracking and velocity estimation.

[0139] S3045: Calculate the displacement change rate of the target position information to be fused within the preset fusion period, and obtain the target velocity information to be fused.

[0140] Specifically, the preset fusion period refers to two or more consecutive time intervals, which are time windows used for analyzing the trend of target position changes. Within each fusion period, the target position information corresponding to multiple consecutive time points is acquired. The fused target position information includes the target's position coordinate data in three-dimensional space. By calculating the spatial coordinate difference between two adjacent time points and dividing it by the corresponding time interval, the instantaneous displacement change rate within that time period is obtained. The instantaneous displacement change rates of multiple time periods are weighted and averaged to obtain the average displacement change rate within the entire fusion period. This average displacement change rate is used as the fused target velocity information.

[0141] In one embodiment, step S40, which calculates the target collision time and target distance based on the target position information and target velocity information, includes:

[0142] S401: Based on the projection of the target location information and the motorcycle's current location information onto the driving direction, determine the relative distance in the driving direction and obtain the relative distance information.

[0143] Specifically, target location information refers to the fused position coordinates of the target in three-dimensional space, motorcycle current location information refers to the real-time spatial position coordinates of the motorcycle in the same coordinate system obtained through the vehicle positioning device, and driving direction refers to the current velocity vector direction of the motorcycle or the direction of the preset navigation path. The target location information and the motorcycle current location information are projected onto the driving direction respectively, and the distance difference between the two projected coordinates is calculated to obtain the relative distance in the driving direction. This relative distance in the driving direction is used as the relative distance information.

[0144] S402: Based on the projection of the target speed information and the motorcycle's current speed information onto the direction of travel, determine the relative speed in the direction of travel and obtain the relative speed information.

[0145] Specifically, target speed information refers to the speed vector information obtained by calculating the rate of displacement change between consecutive moments based on the fused target position information within a preset fusion period. Motorcycle current speed information refers to the real-time speed vector information of the motorcycle itself measured by an onboard speed sensor. The driving direction refers to the direction vector of the motorcycle's current speed vector. The target speed information and the motorcycle's current speed information are projected onto the driving direction to obtain the projected target speed value and the projected motorcycle current speed value. The difference between the two is calculated to obtain the relative speed in the driving direction, which is then used as the relative speed information.

[0146] S403: Perform time-series prediction on relative distance and relative velocity information to obtain the predicted relative distance curve.

[0147] Specifically, when making time-series predictions of relative distance and relative speed information, a unified prediction time series is first constructed. This time series refers to a series of equally spaced sampling time points extending from the current moment into the future, used to describe the short-term trend of the relative motion state between the motorcycle and the target. Then, the relative distance and relative speed information corresponding to the current moment are selected as initial state values. The relative distance information refers to the real-time distance difference between the motorcycle and the target in the direction of travel, and the relative speed information refers to the real-time speed difference between the motorcycle and the target in the direction of travel. Then, based on the kinematic constant model, a stepwise prediction is made under this initial state. At each prediction time point, the predicted distance value of the previous time point is subtracted from the product of the relative speed information and the time interval to obtain a new predicted distance value. The prediction relative distance values ​​of multiple time points are generated iteratively, and the predicted distance values ​​of all prediction time points are arranged in chronological order to form a prediction relative distance curve.

[0148] S404: Under the constraint of the predicted relative distance curve, when the relative speed information indicates that the target is approaching the motorcycle, the moment when the predicted relative distance curve first reaches zero distance is determined as the target collision time; when the relative speed information does not indicate that the target is approaching the motorcycle, the target collision time is set to an invalid value.

[0149] Specifically, when determining the target collision time under the constraint of the predicted relative distance curve, the system first determines whether the target is approaching the motorcycle based on the relative speed information. The relative speed information refers to the speed difference between the motorcycle and the target in the direction of travel. If the speed difference is negative, it means that the target is moving in the direction of approaching the motorcycle. When the relative speed information is negative, the system judges each time point based on the predicted relative distance value corresponding to each predicted time point in the predicted relative distance curve until the predicted relative distance is detected to be less than or equal to zero for the first time. The predicted time corresponding to that time point is then identified as the target collision time. When the relative speed information is zero or positive, it means that the target is not approaching the motorcycle, and the target collision time is set to an invalid value.

[0150] S405: Determine the lateral offset based on the difference in projection between the target position information and the motorcycle's current position information in the lateral direction.

[0151] Specifically, when determining the lateral offset, the target position information and the motorcycle's current position information are first projected into the two-dimensional map coordinate system. The coordinate values ​​of the two positions in the lateral direction (i.e., the axis perpendicular to the motorcycle's direction of travel) are extracted. The target position information refers to the position coordinates of the target in the current coordinate system obtained by multi-source ranging fusion, and the motorcycle's current position information refers to the current position coordinates obtained by the motorcycle positioning device. Then, the lateral offset is obtained by calculating the numerical difference between the coordinate values ​​of the target position information in the lateral direction and the coordinate values ​​of the motorcycle's current position information in the lateral direction. This lateral offset is used to subsequently determine whether there is a risk of lateral collision with the target.

[0152] S406: When the lateral offset does not exceed the preset lateral threshold, the relative distance information at the current moment is determined as the target distance; when the lateral offset exceeds the preset lateral threshold, the target distance is set to an invalid value.

[0153] Specifically, when determining the relative distance information at the current moment as the target distance, the lateral offset determined in the previous steps is first used to determine whether it does not exceed a preset lateral threshold. The lateral offset refers to the difference between the projection of the target position information and the current position information of the motorcycle in the lateral direction. The preset lateral threshold is the distance limit used to determine whether the target is in a potential collision path. When the absolute value of the lateral offset is less than or equal to the preset lateral threshold, it means that the target and the motorcycle are on a path where a collision is possible in the lateral direction. In this case, the relative distance information obtained at the current moment is directly used as the target distance. When the absolute value of the lateral offset is greater than the preset lateral threshold, it means that the target and the motorcycle deviate significantly in the lateral direction and do not constitute a potential collision risk. Therefore, the target distance is set to an invalid value to exclude the influence of the target in subsequent collision judgment.

[0154] In one embodiment, step S405, which determines the lateral offset based on the projection difference between the target position information and the motorcycle's current position information in the lateral direction, includes:

[0155] S4051: Project the target location information and the current position information of the motorcycle onto the horizontal coordinate axis perpendicular to the motorcycle's driving direction to obtain the target's horizontal coordinate value and the motorcycle's horizontal coordinate value.

[0156] Specifically, when projecting the target location information and the motorcycle's current location information onto the lateral coordinate axis perpendicular to the motorcycle's driving direction, the two-dimensional spatial coordinates of the target location information and the motorcycle's current location information are first obtained based on a unified geographic coordinate system. The target location information refers to the spatial position of the target identified at the current moment in the geographic coordinate system, and the motorcycle's current location information refers to the spatial position of the motorcycle itself in the geographic coordinate system at the current moment. Then, the unit direction vector of the motorcycle is determined according to its driving direction, and a lateral direction vector perpendicular to this direction vector is constructed. Next, using the projection formula from a point to a line, the target location information and the motorcycle's current location information are projected onto the lateral direction vector, respectively, to obtain the target's lateral coordinate value and the motorcycle's lateral coordinate value.

[0157] Furthermore, the calculation method involves linearly projecting the target position information onto the direction perpendicular to the motorcycle's current travel direction within a two-dimensional plane. Let the motorcycle's current position information be represented by two-dimensional coordinates. The direction vector of the motorcycle's travel is Then its unit normal vector (i.e., the unit vector perpendicular to the direction of travel) is Let the target location information be... Then the projected coordinates of the target in the horizontal direction are: .

[0158] S4052: Calculate the difference between the horizontal coordinates of the information target and the horizontal coordinates of the information motorcycle to obtain the initial horizontal offset.

[0159] Specifically, based on the acquired target location information and the motorcycle's current location information, the numerical difference between the target's lateral coordinate value and the motorcycle's lateral coordinate value is determined. Let the target's lateral coordinate value be... The motorcycle's lateral coordinate value is The initial lateral offset is This initial lateral offset is used to characterize the degree of positional deviation of the target relative to the motorcycle in the lateral direction.

[0160] S4053: The initial lateral offset of the information is corrected according to the attitude correction angle of the motorcycle's driving direction to obtain the lateral offset.

[0161] Specifically, the initial lateral offset is corrected by cosine projection according to the attitude correction angle of the motorcycle's driving direction. Let the initial lateral offset be... The attitude correction angle for the motorcycle's direction of travel is The corrected lateral offset is The attitude correction angle refers to the angle between the actual driving direction of the motorcycle and the longitudinal direction in the set reference coordinate system. This correction operation is used to eliminate the interference of the motorcycle's dynamic attitude changes such as turning and yaw on the accuracy of lateral offset calculation, so that the lateral offset can more accurately reflect the relative deviation between the target and the motorcycle in the direction perpendicular to the driving direction.

[0162] In one embodiment, step S50, namely generating a helmet prompt instruction based on the warning trigger information, includes:

[0163] S501: Match the warning trigger information with the preset prompt type mapping rules to determine the prompt category information.

[0164] Specifically, the trigger status and prompt priority parameter in the warning trigger information are input into the prompt type mapping function. The prompt type mapping function is a set of rules that associate different types of warning trigger information with prompt categories according to preset rules. The trigger status is a status indicator indicating whether there is a collision risk to the target. The prompt priority parameter is a risk level value calculated based on factors such as the target's approach speed, the slope of the predicted relative distance curve, and the target type. In the prompt type mapping function, it is first determined whether there is a situation that requires a prompt based on the trigger status. If so, the corresponding prompt category code is further searched in the mapping rules according to the prompt priority parameter. The prompt category information includes one or more combinations of visual prompts, sound prompts, or vibration prompts.

[0165] S502: Retrieve the corresponding prompt content from the prompt content library according to the prompt category information, and convert the prompt content sequence into prompt signal parameters, including prompt method, prompt intensity and prompt duration.

[0166] Specifically, the prompt category information is input into the index retrieval function of the prompt content library. The prompt content library refers to a set of preset prompt text, prompt audio files, or vibration mode codes according to different prompt category information. Through the index mapping rules in the prompt content library, the prompt content sequence that matches the prompt category information is located. The prompt content sequence is parsed according to the format conversion rules. The parsing operation includes encoding the presentation mode of the prompt content into a prompt mode, extracting the execution intensity parameter of the prompt content into a prompt intensity, and mapping the duration field of the prompt content into a prompt duration. The prompt mode includes one or more combinations of voice broadcast, graphic flashing, or controller vibration. The prompt intensity represents the parameter value of visual brightness level, volume amplitude, or vibration amplitude. The prompt duration represents the length of time required to maintain a single prompt or continuous prompt.

[0167] S503: The prompt method, prompt intensity, and prompt duration are uniformly encapsulated to obtain the helmet prompt instruction.

[0168] Specifically, the prompt intensity and prompt duration are written as field parameters into the instruction encapsulation format template. The instruction encapsulation format template is a data structure that conforms to the helmet controller communication protocol. The field corresponding to the prompt mode is used to identify the form of expression used to execute the prompt, the prompt intensity field is used to set the power or amplitude parameter of the prompt output device, and the prompt duration field is used to set the time period of the prompt output. During the encapsulation process, the prompt mode, prompt intensity, and prompt duration are formatted and encoded according to the preset byte arrangement order to generate a data frame containing complete control parameters. The data frame is the helmet prompt instruction.

[0169] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A multi-source ranging fusion collision warning method for motorcycle helmets, characterized in that, The aforementioned multi-source ranging fusion collision warning method for motorcycle helmets includes: Acquire synchronous detection information, which includes near-range ranging information, far-range ranging information, and attitude information; The motorcycle's driving condition is determined based on the attitude information in the synchronous detection information. When the driving condition is cornering, the scanning frequency and range of the long-distance ranging information in the outer direction are adjusted and the echo in the inner direction is suppressed. When the driving condition is straight, the continuous scanning of the long-distance ranging information is maintained, and the scanning weight of the short-distance ranging information is directionally corrected to obtain multi-source detection information with attitude compensation. The environmental visibility status is determined based on multi-source detection information. When detecting low visibility conditions, the weights of near-range and long-range ranging information in the multi-source detection information are updated in the data fusion process based on the echo intensity distribution, point cloud sparsity, and stability parameters in the multi-source detection information to obtain the fused target location information and fused target velocity information. Calculate the target collision time and target distance based on target location and speed information. Compare the target collision time with a preset collision time threshold and the target distance with a preset distance threshold. When either condition is met, generate a warning trigger message. Based on the warning trigger information, generate helmet prompt instructions.

2. The multi-source ranging fusion collision warning method for motorcycle helmets according to claim 1, characterized in that, The motorcycle's driving condition is determined based on the attitude information in the synchronous detection information. When the driving condition is cornering, the scanning frequency and range of the long-distance ranging information in the outer direction are adjusted, and the echo in the inner direction is suppressed, including: Obtain the tilt angle and yaw rate from the attitude information; When the tilt angle is greater than the preset tilt angle threshold and the yaw rate is greater than the preset yaw rate threshold, the motorcycle is determined to be in a cornering condition. In cornering conditions, the scanning frequency of the outer direction in the long-distance ranging information is increased to the upper limit of the preset scanning frequency, and the scanning range of the outer direction is expanded to the preset expansion range. At the same time, the scanning frequency of the inner direction is reduced to the lower limit of the preset scanning frequency, and the scanning range of the inner direction is shrunk to the preset shrinkage range, so as to reduce the interference of the inner direction echo.

3. The multi-source ranging fusion collision warning method for motorcycle helmets according to claim 1, characterized in that, When the driving condition is straight, continuous scanning of long-range ranging information is maintained, and the scanning weight of short-range ranging information is directionally corrected to obtain attitude-compensated multi-source detection information, including: When the tilt angle value is less than or equal to the preset tilt angle threshold and the yaw rate is less than or equal to the preset yaw rate threshold, the motorcycle is determined to be in a straight-line working condition. Under straight-line conditions, the omnidirectional continuous scanning mode of long-range ranging information is kept unchanged. Based on the vehicle orientation parameters in the attitude information and the target occurrence frequency distribution of each scanning sector within the preset time window, the target density prediction results of each scanning sector in the next monitoring cycle are predicted. The directional scanning weight of the near-range ranging information in each scanning sector is dynamically adjusted according to the target density prediction results to obtain attitude-compensated multi-source detection information.

4. The multi-source ranging fusion collision warning method for motorcycle helmets according to claim 3, characterized in that, Based on the vehicle orientation parameters in the attitude information and the target occurrence frequency distribution of each scanning sector within a preset time window, the target density prediction results for each scanning sector in the next monitoring cycle are predicted, including: Obtain real-time speed information and road curvature information for motorcycles; The window length of the preset time window is determined based on real-time vehicle speed information and road curvature information, and the sampling time interval is determined based on the window length; Based on the sampling time interval, time normalization is performed on the number of target detections in each scanning sector, and the weighted historical target occurrence frequency is calculated in combination with the echo signal quality factor; Based on the correlation between the weighted historical target occurrence frequency and vehicle orientation parameters, the target occurrence trend value of each scanning sector is calculated; Based on the target occurrence trend value, the probability of target occurrence in each scanning sector in the next monitoring period is estimated, and the potential target density prediction result of each scanning sector is obtained.

5. The multi-source ranging fusion collision warning method for motorcycle helmets according to claim 3, characterized in that, Based on the target density prediction results, the directional scanning weights of near-range ranging information in each scanning sector are dynamically adjusted to obtain attitude-compensated multi-source detection information, including: Obtain the target density value of each scanned sector from the target density prediction results; The target density value is compared with a preset target density threshold to determine the priority of each scan sector; The directional scanning weights of near-range ranging information in each scanning sector are adjusted by increasing or decreasing according to priority to obtain the adjusted directional scanning weights. The near-range ranging information is weighted according to the adjusted directional scanning weights to obtain attitude-compensated multi-source detection information.

6. The multi-source ranging fusion collision warning method for motorcycle helmets according to claim 1, characterized in that, The environmental visibility status is determined based on multi-source detection information. When detecting low visibility conditions, the weights of near-range and long-range ranging information in the multi-source detection information are updated during data fusion based on echo intensity distribution, point cloud sparsity, and stability parameters. This yields the fused target location information and fused target velocity information, including: Initial weight parameters are assigned to near-range and long-range ranging information based on echo intensity distribution; The initial weight parameters are corrected based on the sparsity of the point cloud to obtain the corrected weight parameters; The corrected weight parameters are normalized by the stability parameters to obtain the updated weight parameters. During the data fusion process, updated weight parameters are applied to perform weighted fusion of near-range and long-range ranging information to obtain fused target location information and fused target velocity information.

7. A multi-source ranging fusion collision warning method for motorcycle helmets according to claim 6, characterized in that, During the data fusion process, updated weight parameters are applied to perform weighted fusion of near-range and long-range ranging information to obtain fused target location and velocity information, including: By multiplying the near-range and far-range ranging information by the corresponding updated weight parameters, we obtain the near-range weighted ranging information and the far-range weighted ranging information. The consistency of the near-range weighted ranging information and the far-range weighted ranging information is verified, and data points with ranging deviations exceeding a preset threshold are removed to obtain the near-range weighted ranging information and the far-range weighted ranging information with consistency verification. The near-range weighted ranging information from the consistency check is fused with the far-range weighted ranging information from the consistency check to obtain the fused ranging result; The target's three-dimensional spatial position is calculated by combining the fused ranging results with attitude information, thus obtaining the fused target position information; The displacement change rate of the target position information to be fused is calculated within a preset fusion period to obtain the target velocity information to be fused.

8. The multi-source ranging fusion collision warning method for motorcycle helmets according to claim 1, characterized in that, Calculate the target collision time and target distance based on target location and velocity information, including: Based on the projection of the target location information and the motorcycle's current location information onto the direction of travel, the relative distance in the direction of travel is determined, and the relative distance information is obtained. Based on the projection of the target speed information and the motorcycle's current speed information onto the direction of travel, the relative speed in the direction of travel is determined, and the relative speed information is obtained. Time-series prediction is performed on relative distance and relative velocity information to obtain the predicted relative distance curve; Under the constraint of the predicted relative distance curve, when the relative speed information indicates that the target is approaching the motorcycle, the moment when the predicted relative distance curve first reaches zero distance is determined as the target collision time; when the relative speed information does not indicate that the target is approaching the motorcycle, the target collision time is set to an invalid value. The lateral offset is determined based on the difference in the projection of the target location information and the motorcycle's current location information in the lateral direction. When the lateral offset does not exceed the preset lateral threshold, the relative distance information at the current moment is determined as the target distance; when the lateral offset exceeds the preset lateral threshold, the target distance is set to an invalid value.

9. A multi-source ranging fusion collision warning method for motorcycle helmets according to claim 8, characterized in that, Based on the difference in the lateral projection between the target location information and the motorcycle's current location information, the lateral offset is determined, including: Project the target location information and the current location information of the motorcycle onto the horizontal coordinate axis perpendicular to the motorcycle's direction of travel to obtain the target's horizontal coordinate value and the motorcycle's horizontal coordinate value. The initial lateral offset is obtained by calculating the difference between the lateral coordinates of the information target and the lateral coordinates of the information motorcycle. The initial lateral offset of the information is corrected according to the attitude correction angle of the motorcycle's driving direction to obtain the lateral offset.

10. A multi-source ranging fusion collision warning method for motorcycle helmets according to claim 1, characterized in that, Based on the aforementioned warning trigger information, a helmet prompt instruction is generated, including: The warning trigger information is matched with the preset prompt type mapping rules to determine the prompt category information; Based on the prompt category information, the corresponding prompt content is retrieved from the prompt content library, and the prompt content sequence is converted into prompt signal parameters, which include prompt method, prompt intensity, and prompt duration. The prompt method, prompt intensity, and prompt duration are uniformly encapsulated to obtain helmet prompt instructions.

Images