Vehicle sideslip detection and early warning method and system

Abstract

The invention discloses a vehicle sideslip detection and early warning method and system, and belongs to the technical field of traffic safety detection.The method comprises the following steps that the road condition is detected, a vehicle sideslip risk threshold value is pre-judged in advance, the wheel speed of each wheel is obtained, and the current vehicle average wheel speed is calculated; obtaining the whole vehicle longitudinal acceleration of the vehicle; the data acquisition module sends the acquired data to the information transmission processing module; the information transmission processing module preprocesses the received data, inputs the corrected vehicle speed and the wheel speed of each wheel into a BIRCH algorithm for clustering, finds the vehicle speed and the wheel speed centroid of each wheel after a period of time, and calculates the vehicle slip rate; when the vehicle slip rate exceeds the vehicle sideslip threshold range, an early warning signal is issued, and a vehicle owner is reminded of the sideslip risk level and the direction to be corrected of the vehicle through voice and a vehicle-mounted display screen. According to the invention, real-time monitoring and analysis of the driving state of the vehicle are realized, timely early warning of sideslip of the vehicle is realized, and the driving safety is improved.

Description

Technical Field

[0001] The present invention relates to a method and system for detecting and warning vehicle sideslip, belonging to the technical field of traffic safety detection. Background Art

[0002] In road traffic, vehicle sideslip is a common safety hazard. Due to factors such as braking, rotational inertia, and small adhesion caused by road surface conditions (such as ice, mud, wetness, etc.), the wheels of a vehicle may move laterally, triggering sideslip. This sideslip phenomenon often leads to serious traffic accidents such as collisions, rollovers, and falling into ditches, posing a serious threat to driving safety.

[0003] In the prior art, the detection and warning of vehicle sideslip are mainly carried out by measuring the friction between the vehicle and the ground, calculating the vehicle speed, etc. However, these methods have problems such as cumbersome detection steps, large errors and delays, and the inability to automatically determine the sideslip direction when sideslip occurs, resulting in unsatisfactory warning effects.

[0004] In order to detect vehicle sideslip in a timely manner and give warnings, the present invention proposes a method and system for detecting and warning vehicle sideslip. Summary of the Invention

[0005] To solve the above problems, the present invention proposes a method and system for detecting and warning vehicle sideslip, which can realize timely warning of vehicle sideslip by real-time monitoring and analyzing the driving state of the vehicle, and improve driving safety.

[0006] The technical solution adopted by the present invention to solve its technical problems is as follows: In a first aspect, a method for detecting and warning vehicle sideslip provided by an embodiment of the present invention includes the following steps: Step 1, install a road surface condition sensor at the front bumper of the vehicle, and use the road surface condition sensor to obtain the road surface condition, and pre-judge the vehicle sideslip risk threshold in advance; Step 2, install Hall wheel speed sensors on the four wheels of the vehicle respectively, and use the Hall wheel speed sensors to obtain the wheel speed of each wheel, and calculate the current average wheel speed of the vehicle; Step 3, install a vertical accelerometer at the center position of the bottom of the vehicle, and use the vertical accelerometer to obtain the longitudinal acceleration of the whole vehicle; Step 4, send the obtained wheel speed of each wheel and the longitudinal acceleration of the whole vehicle to the information transmission and processing module through the data acquisition module; Step 5, the information transmission and processing module preprocesses the received data and corrects the current vehicle speed, inputs the corrected vehicle speed and the wheel speed of each wheel into the BIRCH algorithm, constructs a dynamic tree, clusters to find the centroid of the vehicle speed and the wheel speed of each wheel after a period of time, calculates the vehicle slip ratio, and feeds back the data that meets the vehicle side slip threshold to the back-end platform; Step 6, when the vehicle slip ratio exceeds the vehicle side slip threshold range, an early warning signal is issued through the early warning module, a warning icon and the vehicle correction direction are displayed on the in-vehicle display screen, and the driver is prompted by voice about the side slip risk level and the direction that the vehicle should be corrected.

[0007] As a possible implementation of this embodiment, the current average vehicle wheel speed is equal to the average of the wheel speeds detected by the four Hall wheel speed sensors.

[0008] As a possible implementation of this embodiment, the information transmission and processing module uses the Kalman filtering algorithm to perform filtering preprocessing on the received data.

[0009] As a possible implementation of this embodiment, the specific steps for the information transmission and processing module to correct the current vehicle speed are as follows: substituting the current average vehicle wheel speed and the vehicle longitudinal acceleration into the pre-stored speed theoretical formula to correct the current vehicle speed, where V avg represents the average wheel speed of the vehicle during this period, and a x represents the vehicle longitudinal acceleration.

[0010] As a possible implementation of this embodiment, the early warning module includes a voice broadcast warning device and an in-vehicle display screen. When the vehicle slip ratio exceeds the vehicle side slip threshold range, the early warning module is activated, a warning icon and the direction that the vehicle steering wheel should be corrected are displayed on the in-vehicle display screen, and at the same time, the driver is prompted by voice about the side slip risk level and the direction that the vehicle should be corrected.

[0011] As a possible implementation of this embodiment, the steps of inputting the corrected vehicle speed and the wheel speed of each wheel into the BIRCH algorithm, constructing a dynamic tree, clustering to find the centroid of the vehicle speed and the wheel speed of each wheel after a period of time, and calculating the vehicle slip ratio include: Storing the corrected vehicle speed and the wheel speed of each wheel in the database as multi-dimensional data points, and initializing the CF tree as an empty tree; Inserting data points and globally clustering to dynamically establish a CF tree; Merging the CF vectors of non-overlapping clusters to quickly calculate the cluster centroid; The centroid of each cluster represents the typical value of the corrected vehicle speed and the wheel speed, and is substituted into the vehicle slip ratio formula Calculate the vehicle slip ratio, where S represents the slip ratio, Vref represents the corrected vehicle speed, and Vr represents the wheel speed.

[0012] As a possible implementation of this embodiment, the dynamically building the CF tree includes: Set parameters, where the parameters include the maximum number of child nodes B of the branch node, the maximum number of MinClusters L of the leaf node, and the cluster maximum diameter threshold T; Recursively select the nearest child node. Starting from the root node, select the nearest child node based on the average distance between classes until reaching the leaf node; Leaf node processing and splitting. Check whether the nearest MinCluster in the leaf node can absorb the data point. If it can absorb, update the CF value; otherwise, try to add a new MinCluster. If the leaf node is full, split the pair of MinClusters with the farthest distance and reassign other MinClusters to the new leaf node; Backtracking update and tree structure adjustment. Update the parent node. If the leaf node splits, backtrack and update the CF value of the parent node upward, and may recursively split non-leaf nodes until the root node.

[0013] As a possible implementation of this embodiment, the vehicle sideslip threshold includes mild sideslip S a , moderate sideslip S b and severe sideslip S c , and their threshold ranges are [10%, 15%], (15%, 20%], and (20%, +∞] respectively.

[0014] As a possible implementation of this embodiment, the dynamic optimization of the vehicle sideslip threshold is as follows: on rainy or snowy days, the vehicle sideslip threshold is lowered by 5%; when the vehicle is in a curve scene, combined with the steering angle data, if the steering angle > 30°, the vehicle sideslip threshold is additionally lowered by 3%; the instantaneous peak < 0.5 seconds only records the log and does not trigger an alarm; when the vehicle slip ratio continuously exceeds the vehicle sideslip threshold for 2 seconds, the warning level is immediately upgraded.

[0015] In a second aspect, a vehicle sideslip detection and warning system provided by an embodiment of the present invention includes a data acquisition module, an information transmission and processing module, and a warning module; the data acquisition module includes a road surface condition sensor, a Hall wheel speed sensor, and a vertical accelerometer. The road surface condition sensor is installed at the front bumper of the vehicle, the Hall wheel speed sensors are respectively installed on the four wheels of the vehicle, and the vertical accelerometer is installed at the center position of the vehicle bottom; the data acquisition module uses the road surface condition sensor to detect the road surface condition, pre-judge the vehicle sideslip risk threshold in advance, uses the Hall wheel speed sensor to obtain the wheel speed of each wheel, and uses the vertical accelerometer to obtain the longitudinal acceleration of the entire vehicle; the information transmission and processing module is used to receive in real time the data detected by the data acquisition module, and perform algorithm analysis on the collected data in real time, calculate the current vehicle slip ratio, and feedback the data that meets the vehicle sideslip threshold to the backend platform; the warning module is used to issue a warning signal, display a warning icon and the vehicle correction direction on the in-vehicle display screen, and at the same time give a voice prompt to the vehicle owner about the sideslip risk level and the direction that the vehicle should correct.

[0016] As a possible implementation manner of this embodiment, the information transmission and processing module substitutes the current average vehicle wheel speed and the longitudinal acceleration of the entire vehicle into a pre-stored speed theoretical formula to correct the current vehicle speed, and then inputs the corrected vehicle speed and the wheel speed of each wheel into the BIRCH algorithm to construct a dynamic tree, clusters to find the centroid of the vehicle speed and the wheel speed of each wheel after a period of time, and finally substitutes it into the vehicle slip ratio formula for calculation and comparison with the set vehicle sideslip threshold range.

[0017] As a possible implementation manner of this embodiment, the warning device includes a voice broadcast warning device and an in-vehicle display screen. When the vehicle slip ratio exceeds the threshold range, the warning module is activated, and a warning icon and the direction that the vehicle steering wheel should correct are displayed on the in-vehicle display screen, and at the same time, the vehicle owner is given a voice prompt about the sideslip risk level and the direction that the vehicle should correct.

[0018] The technical solution of the embodiment of the present invention may have the following beneficial effects: The present invention realizes the real-time monitoring and analysis of the vehicle driving state, thereby realizing the timely warning of vehicle sideslip; processes and analyzes data through the BIRCH algorithm and the Kalman filter algorithm, accurately screens and excludes bad data, and improves the accuracy and reliability of the obtained data; at the same time, dynamically adjusts the structure of the clustering tree, so that the clustering result can better fit the distribution of data points, further improving the warning accuracy. By real-time monitoring and analyzing the vehicle driving state, the timely warning of vehicle sideslip is realized, effectively avoiding traffic accidents caused by sideslip and improving driving safety. Description of the Drawings

[0019] Figure 1It is a flowchart of a method for detecting and warning of vehicle sideslip shown according to an exemplary embodiment; Figure 2 It is a schematic structural diagram of a system for detecting and warning of vehicle sideslip shown according to an exemplary embodiment; Figure 3 It is a diagram of the installation position of a data acquisition module shown according to an exemplary embodiment; Figure 4 It is a schematic diagram of clustering results obtained by clustering using the BIRCH algorithm shown according to an exemplary embodiment; Figure 5 It is a box plot for comparing the distribution of slip ratios of each cluster shown according to an exemplary embodiment; Figure 6 It is a heat map for comparing the characteristic statistics of each cluster; Figure 7 It is a time series diagram of the slip ratio. Detailed implementation manners

[0020] The present invention will be further described below in conjunction with the drawings and embodiments: To clearly illustrate the technical features of this solution, the present invention will be elaborated in detail below through specific implementation manners and in conjunction with its drawings. The following disclosure provides many different embodiments or examples for implementing different structures of the present invention. To simplify the disclosure of the present invention, the components and settings of specific examples are described below. In addition, the present invention may repeat reference numerals and / or letters in different examples. This repetition is for the purpose of simplification and clarity, and does not itself indicate the relationship between the various embodiments and / or settings discussed. It should be noted that the components illustrated in the drawings are not necessarily drawn to scale. The present invention omits the description of well-known components and processing technologies and processes to avoid unnecessarily limiting the present invention.

[0021] As Figure 1 shown, a method for detecting and warning of vehicle sideslip provided by an embodiment of the present invention includes the following steps: Step 1, install a road surface condition sensor at the front bumper of the vehicle, and use the road surface condition sensor to obtain the road surface condition to pre-judge the vehicle sideslip risk threshold; Step 2, install Hall wheel speed sensors on the four wheels of the vehicle respectively, and use the Hall wheel speed sensors to obtain the wheel speed of each wheel and calculate the current average wheel speed of the vehicle; Step 3, install a vertical accelerometer at the center position of the vehicle bottom, and use the vertical accelerometer to obtain the longitudinal acceleration of the whole vehicle; Step 4, send the obtained wheel speed of each wheel and the longitudinal acceleration of the whole vehicle to the information transmission and processing module through the data acquisition module; Step 5: The information transmission and processing module preprocesses the received data and corrects the current vehicle speed. It inputs the corrected vehicle speed and the wheel speed of each wheel into the BIRCH algorithm to construct a dynamic tree, clusters to find the centroid of the vehicle speed and the wheel speed of each wheel after a period of time, calculates the vehicle slip ratio, and feeds back the data that meets the vehicle sideslip threshold to the back-end platform. Step 6: When the vehicle slip ratio exceeds the vehicle sideslip threshold range, the warning module issues a warning signal, displays a warning icon and the vehicle correction direction on the in-vehicle display screen, and uses voice prompts to inform the vehicle owner of the sideslip risk level and the direction that the vehicle should correct.

[0022] As a possible implementation of this embodiment, the current average vehicle wheel speed is equal to the average of the wheel speeds detected by the four Hall wheel speed sensors.

[0023] As a possible implementation of this embodiment, the information transmission and processing module uses the Kalman filter algorithm to perform filtering preprocessing on the received data.

[0024] As a possible implementation of this embodiment, the specific steps for the information transmission and processing module to correct the current vehicle speed are as follows: Substitute the current average vehicle wheel speed and the vehicle longitudinal acceleration into the pre-stored speed theoretical formula to correct the current vehicle speed, where V avg represents the average wheel speed of the vehicle during this period, and a x represents the vehicle longitudinal acceleration.

[0025] As a possible implementation of this embodiment, the warning module includes a voice broadcast warning device and an in-vehicle display screen. When the vehicle slip ratio exceeds the vehicle sideslip threshold range, the warning module is activated. A warning icon and the direction that the vehicle steering wheel should correct are displayed on the in-vehicle display screen, and at the same time, the vehicle owner is informed of the sideslip risk level and the direction that the vehicle should correct through voice prompts.

[0026] As a possible implementation of this embodiment, the process of inputting the corrected vehicle speed and the wheel speed of each wheel into the BIRCH algorithm to construct a dynamic tree, clustering to find the centroid of the vehicle speed and the wheel speed of each wheel after a period of time, and calculating the vehicle slip ratio includes: Store the corrected vehicle speed and the wheel speed of each wheel in the database as multi-dimensional data points, and initialize the CF tree as an empty tree; Insert the data points and perform global clustering to dynamically establish the CF tree; Merge the CF vectors of non-overlapping clusters and quickly calculate the cluster centroid; The centroid of each cluster represents the typical value of the corrected vehicle speed and the wheel speed, and substitute it into the vehicle slip ratio formula Calculate the vehicle slip ratio, where S represents the slip ratio, Vref represents the corrected vehicle speed, and Vr represents the wheel speed.

[0027] As a possible implementation of this embodiment, the dynamically building the CF tree includes: Set parameters, where the parameters include the maximum number of child nodes B of the branch node, the maximum number of MinClusters L of the leaf node, and the cluster maximum diameter threshold T; Recursively select the nearest child node. Starting from the root node, select the nearest child node based on the average distance between classes until reaching the leaf node; Leaf node processing and splitting. Check whether the nearest MinCluster in the leaf node can absorb the data point. If it can absorb, update the CF value. Otherwise, try to add a new MinCluster; if the leaf node is full, split the pair of MinClusters with the farthest distance, and reassign other MinClusters to the new leaf node; Backtracking update and tree structure adjustment. Update the parent node. If the leaf node is split, backtrack and update the CF value of the parent node upward, and may recursively split non-leaf nodes until the root node.

[0028] As a possible implementation of this embodiment, the vehicle sideslip threshold includes mild sideslip S a 、moderate sideslip S b and severe sideslip S c , and their threshold ranges are [10%, 15%], (15%, 20%], and (20%, +∞] respectively.

[0029] As a possible implementation of this embodiment, the dynamic optimization of the vehicle sideslip threshold is as follows: on rainy or snowy days, the vehicle sideslip threshold is lowered by 5%; when the vehicle is in a curve scenario, combined with the steering angle data, if the steering angle > 30°, the vehicle sideslip threshold is additionally lowered by 3%; instantaneous peak < 0.5 seconds only records the log and does not trigger an alarm; when the vehicle slip ratio continuously exceeds the vehicle sideslip threshold for 2 seconds, the warning level is immediately upgraded.

[0030] Compared with the prior art, the present invention has the following advantages: 1. Process and analyze data through the BIRCH algorithm and the Kalman filter algorithm, accurately screen and exclude bad data, make the obtained data accurate and precise, and can intuitively present the interval of data intensive distribution, which is intuitive and perceivable. Effectively detect the sideslipping vehicle in time, achieve timely warning, greatly reduce the traffic tragedies caused by vehicle sideslip, and effectively guarantee the traffic safety of people's travel.

[0031] 2. By using the BIRCH algorithm to construct a dynamically changing tree from the existing data, the accuracy of clustering data segmentation is greatly improved, the error is further reduced, and the data obtained from the analysis is more accurate and reliable. At the same time, various factors such as temperature and humidity that affect the road surface adhesion coefficient are fully considered, which helps to accurately and timely detect skidding vehicles and improves driving safety.

[0032] 3. The data processing and algorithms of the present invention have a high level of intelligence, can automatically complete tasks such as data collection, processing, analysis, and prediction, reduce manual intervention, and improve work efficiency. At the same time, the system of the present invention also supports remote monitoring and operation, and can realize real-time monitoring and remote control of the vehicle state through network connection, further improving the intelligence level of the system.

[0033] 4. The algorithm of the present invention modifies the CF value of non-leaf nodes from bottom to top and splits nodes when necessary to adapt to newly inserted data points, dynamically adjusting the structure of the clustering tree, so that the clustering results can better fit the distribution of data points.

[0034] As Figure 2 shown, a vehicle skidding detection and warning system provided by an embodiment of the present invention includes a data acquisition module, an information transmission and processing module, and a warning module; the data acquisition module includes a road surface condition sensor, a Hall wheel speed sensor, and a vertical accelerometer. As Figure 3 shown, the Hall wheel speed sensors 1 are respectively installed on the four wheels of the vehicle, the vertical accelerometer 2 is installed at the center position of the vehicle bottom, and the road surface condition sensor 3 is installed at the front bumper of the vehicle; the data acquisition module uses the road surface condition sensor to detect the road surface condition, pre-judge the vehicle skidding risk threshold in advance, uses the Hall wheel speed sensor to obtain the wheel speed of each wheel, and uses the vertical accelerometer to obtain the longitudinal acceleration of the whole vehicle; the information transmission and processing module is used to receive the data detected by the data acquisition module in real time, and perform algorithm analysis on the collected data in real time, calculate the current vehicle slip ratio, and feedback the data that meets the vehicle skidding threshold to the back-end platform; the warning module is used to issue a warning signal, display a warning icon and the vehicle correction direction on the in-vehicle display screen, and at the same time give a voice prompt to the vehicle owner about the skidding risk level and the direction that the vehicle should correct, and give a skidding warning to the vehicle owner.

[0035] The road surface condition sensor is installed at the front bumper of the vehicle to detect the road surface condition and pre-judge the vehicle skidding risk threshold in advance; the Hall wheel speed sensors are respectively installed on the four wheels of the vehicle, and the current average wheel speed of the vehicle is equal to the average value of the wheel speeds detected by the four Hall wheel speed sensors; the vertical accelerometer is installed at the center position of the vehicle bottom to detect the current longitudinal acceleration of the whole vehicle; the detected information is sent to the core processor through the information transmission and processing module to obtain the slip ratio of the current vehicle.

[0036] As a possible implementation of this embodiment, the information transmission and processing module substitutes the current average wheel speed of the vehicle and the longitudinal acceleration of the whole vehicle into a pre-stored speed theoretical formula to correct the current vehicle speed, where V avg represents the average wheel speed of the vehicle during this period, and a x represents the longitudinal acceleration of the whole vehicle; then, the corrected vehicle speed and the wheel speed of each wheel are input into the BIRCH algorithm to construct a dynamic tree, and the centroid of the vehicle speed and the wheel speed of each wheel after a period of time is found by clustering. Finally, it is substituted into the vehicle slip ratio formula to calculate the vehicle slip ratio and compare it with the set vehicle sideslip threshold range.

[0037] The vehicle slip ratio is obtained by the following method: When the tire generates traction or braking force, relative movement will occur between the tire and the ground. The slip ratio is the proportion of the sliding component in the wheel movement. To fully consider the influence of various factors such as temperature and humidity on the road surface adhesion coefficient, as Figure 3 shown, the road surface condition sensor 3 can be used to obtain road surface condition data in real time, and combined with the road surface condition data, it is calculated and analyzed whether the current slip ratio of the vehicle is within the threshold range, which helps to accurately and timely detect the sideslipping vehicle and improve driving safety.

[0038] As a possible implementation of this embodiment, the warning device includes a voice broadcast warning device and an in-vehicle display screen. When the vehicle slip ratio exceeds the threshold range, the warning module is activated, and a warning icon and the direction that the vehicle steering wheel should be corrected are displayed on the in-vehicle display screen. At the same time, a voice prompts the vehicle owner of the sideslip risk level and the direction that the vehicle should be corrected, and a sound and image warning of sideslip is given to the vehicle owner.

[0039] The core processor mainly uses the BIRCH algorithm to process data. The following is a specific description: Input data to preprocess and initialize the data: The corrected vehicle speed (Vref) and the wheel speed (Vr) are used as multi-dimensional data points. Initialize the CF tree: Set parameters "B = 2" (maximum number of child nodes of the branch node), "L = 2" (maximum number of MinCluster of the leaf node), "T = 1.5" (threshold of the maximum diameter of the cluster), and initialize an empty tree. In the BIRCH algorithm, the constructed CF tree contains various CF values (N, LS, SS), where N represents the number of data points belonging to this cluster; LS refers to the linear sum of the feature vectors of all data points in the cluster. Let the data points in the cluster be (x1, x2,..., xN), then Ls can be expressed as ; SS refers to the sum of the squares of the feature vectors of all data points in the cluster, which characterizes the dispersion degree of the data points in the cluster. Its calculation formula is 。The combination of CF values (N, LS, SS) can effectively represent and manage the information of clusters in a tree structure.

[0040] Store the corrected vehicle speed and the wheel speed data of each wheel in the database as data points. Initialize the CF tree as an empty tree, and then insert the data points. For global clustering, the CF tree will be dynamically established as new data points are added. The insertion process includes the following steps: Step1: Recursively select the nearest child node. Starting from the root node, select the nearest child node based on the average distance between classes (such as Euclidean distance) until reaching the leaf node. Ensure that the new data point is assigned to the currently most similar cluster to maintain the local clustering quality.

[0041] Step2: Leaf node processing and splitting. Check whether the nearest MinCluster in the leaf node can absorb the data point (cluster diameter ≤ T). If it can be absorbed, update the CF values (N, LS, SS); otherwise, try to add a new MinCluster. If the leaf node is full (the number of MinClusters = L), then split the pair of MinClusters with the farthest distance and reassign the other MinClusters to the new leaf nodes. By dynamically splitting, balance the clustering accuracy and computational efficiency to ensure that the cluster diameter does not exceed the threshold T.

[0042] Step3: Backtracking update and tree structure adjustment. Update the parent node. If the leaf node is split, then backtrack upward to update the CF value of the parent node and may recursively split non-leaf nodes until the root node to maintain the global statistical consistency of the CF tree and ensure that the hierarchical structure of the tree adapts to the changes in the data distribution.

[0043] Global clustering and result output. CF vector merging. Merge the CF vectors of non-overlapping clusters (such as "CF1 + CF2 = (N1 + N2, LS1 + LS2, SS1 + SS2)"), which supports fast calculation of the cluster centroids (Vref_center = LS1 / N, Vr_center = LS2 / N).

[0044] Output the clustering result: The centroid of each cluster represents the typical values of Vref and Vr. Substitute them into the vehicle slip ratio formula for calculation. The clustering result is as Figure 4 shown.

[0045] Explanation of the clustering result analysis chart: 1. Chart composition: X-axis: vehicle speed (km / h); Y-axis: wheel speed (km / h); 2. Color points: Different colors represent different clustering labels; 3. Cluster 0 (purple): Uniform driving (vehicle speed is stable at 25 - 35 km / h), high matching degree between vehicle speed and wheel speed, low slip ratio (<5%), belonging to a safe state; 4. Cluster 1 (blue): Gradually accelerating (vehicle speed gradually increases from 20 km / h to 40 km / h), vehicle speed slightly higher than wheel speed, medium slip ratio (5–15%), requires monitoring but no immediate risk; 5. Cluster 2 (yellow): Vehicle braking (vehicle speed fluctuates violently, such as suddenly dropping from 30 km / h to 20 km / h), high slip ratio (>20%), may trigger a side-slip warning.

[0046] As Figure 5 shown, boxes of different colors represent the data distribution of 25%–75%. The horizontal line inside the box is the median, the red dashed line is the data range, and the white dot is the mean. The box plot can directly compare the concentration and dispersion of the slip ratios of different clusters. If the overall box of a certain cluster is higher than the threshold, it is determined as a high-risk cluster, and the points outside the box may be instantaneous skidding or noise data.

[0047] Figure 6 The heat map is for the comparison of the statistical quantities of each cluster feature. As Figure 6 shown, the heat map determines the typical vehicle speed / wheel speed of each cluster through the mean value, and evaluates the data fluctuation through the standard deviation. Figure 6 In it, the darker the color, the larger the value. Among them, a low standard deviation indicates concentrated data and more reliable risk judgment; a high standard deviation requires further inspection of outliers.

[0048] Figure 7 The time series plot of the slip ratio. According to Figure 7 it, it can be observed whether the slip ratio continuously exceeds the vehicle side-slip threshold. Different background color blocks show the duration of different driving states, and the red dashed line is the set controllable safety threshold for side-slip vehicles. If it is exceeded, a warning is triggered.

[0049] After all data points are analyzed by the algorithm for clustering, the slip ratio S of the vehicle is obtained in real time 1 、S 2 、S 3 ……S x， Compare S x with the pre-set threshold. If S x is within the range of S a 、S b 、S c in one of them, the corresponding warning module is activated. A warning icon and the vehicle correction direction are displayed on the in-vehicle display screen. At the same time, the driver is prompted by voice about the side-slip risk level and the direction that the vehicle should correct, to give a side-slip warning to the driver. The side-slip risk levels are shown in Table 1.

[0050] Table 1 Vehicle side-slip thresholds corresponding to side-slip risk levels

[0051] Dynamic threshold optimization: 1. On rainy or snowy days, the threshold is lowered by 5% (for example, the mild sideslip is adjusted to 8% - 15%). When the vehicle is in a cornering scenario: Combining the steering angle data, if the steering angle > 30°, the threshold is additionally lowered by 3%.

[0052] 2. Vehicle sideslip duration judgment: Instantaneous peak (< 0.5 seconds): Only record the log, do not trigger an alarm to avoid false alarms; Continuously exceeding the threshold (> 2 seconds): Immediately upgrade the warning level.

[0053] The technical solution of the present invention mainly has the following main parts: 1. Installation and configuration of the data acquisition module: Install Hall wheel speed sensors on the four wheels of the vehicle respectively to obtain the wheel speed of each wheel; install a vertical accelerometer at the center position of the vehicle bottom to obtain the longitudinal acceleration of the whole vehicle.

[0054] 2. Setting of the data processing module: The information transmission and processing module receives the data detected by the data acquisition module and performs algorithm analysis. First, correct the current vehicle speed according to the pre-stored speed theoretical formula. Then, use the BIRCH algorithm to perform clustering analysis on the corrected vehicle speed and the wheel speed of each wheel, find the centroid and substitute it into the vehicle slip ratio formula to calculate the slip ratio.

[0055] 3. Triggering and display of the warning module: When the vehicle slip ratio exceeds the set threshold range, the warning module is activated. A warning icon and the direction that the vehicle steering wheel should be corrected are displayed on the in-vehicle display screen, and at the same time, the voice prompts the owner of the sideslip risk level and the direction that the vehicle should be corrected.

[0056] 4. Dynamic threshold optimization: Adjust the threshold range according to weather and road conditions (such as rainy days, snowy days, corners, etc.) to improve the accuracy and timeliness of the warning. For example, on rainy or snowy days, the threshold is lowered by 5%; when the vehicle is in a cornering scenario and the steering angle is greater than 30°, the threshold is additionally lowered by 3%. In addition, for the case where the instantaneous peak duration is less than 0.5 seconds, only record the log and do not trigger an alarm; for the case where the continuously exceeding the threshold time is greater than 2 seconds, immediately upgrade the warning level.

[0057] Finally, it should be noted that: The above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit them. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art should understand that: It is still possible to modify the specific implementation manners of the present invention or make equivalent replacements, and any modification or equivalent replacement without departing from the spirit and scope of the present invention shall be covered by the protection scope of the claims of the present invention.

Claims

1. A vehicle skidding detection and early warning method, characterized in that: The following steps are involved: Step 1: Install a road condition sensor on the front bumper of the vehicle, and use the road condition sensor to obtain road conditions and predict the vehicle sideslip risk threshold in advance; Step 2, installing Hall-type wheel speed sensors on the four wheels of the vehicle respectively, and using the Hall-type wheel speed sensors to obtain the wheel speed of each wheel, and calculating the current average wheel speed of the vehicle; Step 3, installing a vertical accelerometer at the center of the bottom of the vehicle, and using the vertical accelerometer to obtain the longitudinal acceleration of the vehicle; Step 4, sending the acquired wheel speed of each wheel and the longitudinal acceleration of the whole vehicle to the information transmission processing module through the data acquisition module; Step 5: The information transmission processing module pre-processes the received data and corrects the current vehicle speed, inputs the corrected vehicle speed and the wheel speed of each wheel into the BIRCH algorithm, constructs a dynamic tree, clusters to find the vehicle speed and the wheel speed centroid of each wheel after a period of time, calculates the vehicle slip rate, and feeds back the data that meets the vehicle side slip threshold to the back-end platform; Step 6: When the vehicle slip rate exceeds the vehicle sideslip threshold range, a warning signal is issued through the warning module, a warning icon and the vehicle correction direction are displayed on the vehicle display screen, and the owner is prompted by voice the sideslip risk level and the direction in which the vehicle should be corrected.

2. The vehicle sideslip detection and early warning method according to claim 1, characterized in that: The current average wheel speed of the vehicle is equal to an average value of the wheel speeds detected by the four Hall-type wheel speed sensors.

3. The vehicle sideslip detection and early warning method according to claim 1 is characterized in that: The specific steps of the information transmission processing module to correct the current vehicle speed are: substituting the current vehicle average wheel speed and the vehicle longitudinal acceleration into the pre-stored speed theoretical formula To correct the current vehicle speed, V avg represents the average wheel speed of the vehicle during this period, a x Indicates the longitudinal acceleration of the vehicle.

4. The vehicle sideslip detection and early warning method according to claim 1, characterized in that: The early warning module includes a voice broadcast warning device and a vehicle-mounted display screen. When the vehicle slip rate exceeds the vehicle sideslip threshold range, the early warning module is activated, and a warning icon and the direction in which the vehicle steering wheel should be corrected are displayed on the vehicle-mounted display screen. At the same time, a voice prompt is given to the owner of the sideslip risk level and the direction in which the vehicle should be corrected.

5. The vehicle sideslip detection and early warning method according to claim 1, characterized in that: The corrected vehicle speed and the wheel speed of each wheel are input into the BIRCH algorithm, a dynamic tree is constructed, clustering is performed to find the vehicle speed and the wheel speed centroid of each wheel after a period of time, and the vehicle slip rate is calculated, including: The corrected vehicle speed and the wheel speed of each wheel are stored in the database as multi-dimensional data points, and the CF tree is initialized to an empty tree; Insert data points, perform global clustering, and dynamically build a CF tree; Merge the CF vectors of disjoint clusters and quickly calculate the cluster centroid; The centroid of each cluster represents the typical value of the corrected vehicle speed and wheel speed, which is substituted into the vehicle slip rate formula Calculate the vehicle slip rate, S represents the slip rate, Vref represents the corrected vehicle speed, and Vr represents the wheel speed.

6. The vehicle sideslip detection and early warning method according to claim 5, characterized in that: The dynamically establishing CF tree includes: Set parameters, including the maximum number of child nodes B of a branch node, the maximum number of MinClusters L of a leaf node, and the maximum cluster diameter threshold T; Recursively select the nearest child node, starting from the root node, and select the nearest child node based on the average distance between classes until reaching the leaf node; Processing and splitting leaf nodes: Check whether the nearest MinCluster in the leaf node can absorb the data point. If it can, update the CF value; otherwise, try to add a new MinCluster. If the leaf node is full, split the pair of MinClusters with the farthest distance and reallocate other MinClusters to the new leaf node. Backtracking update and tree structure adjustment, update the parent node, if the leaf node is split, backtrack upward to update the CF value of the parent node, and may recursively split non-leaf nodes until the root node.

7. The vehicle sideslip detection and early warning method according to any one of claims 1 to 6, characterized in that: The vehicle sideslip threshold includes a slight sideslip S a , Moderate side slip S b and severe side slip S c , and their threshold ranges are [10%, 15%], (15%, 20%], and (20%, +∞], respectively.

8. The vehicle sideslip detection and early warning method according to claim 1, characterized in that: The dynamic optimization of the vehicle sideslip threshold is as follows: on rainy or snowy days, the vehicle sideslip threshold is lowered by 5%; when the vehicle is in a curve scene, combined with the steering angle data, if the steering angle is >30°, the vehicle sideslip threshold is further lowered by 3%; if the instantaneous peak value is <0.5 seconds, only the log is recorded and no alarm is triggered; when the vehicle slip rate continues to exceed the vehicle sideslip threshold for 2 seconds, the warning level is immediately upgraded.

9. A vehicle skidding detection and warning system, characterized in that: It includes a data acquisition module, an information transmission processing module, and an early warning module; the data acquisition module includes a road condition sensor, a Hall wheel speed sensor and a vertical accelerometer, the road condition sensor is installed at the front bumper of the vehicle, the Hall wheel speed sensor is installed on the four wheels of the vehicle respectively, and the vertical accelerometer is installed at the center of the bottom of the vehicle; the data acquisition module uses the road condition sensor to detect the road condition, predicts the vehicle side slip risk threshold in advance, uses the Hall wheel speed sensor to obtain the wheel speed of each wheel, and uses the vertical accelerometer to obtain the vehicle longitudinal acceleration of the vehicle; the information transmission processing module is used to receive the data detected by the data acquisition module in real time, and perform algorithm analysis on the collected data in real time, calculate the current vehicle slip rate and feed back the data that meets the vehicle side slip threshold to the back-end platform; the early warning module is used to issue early warning signals, and the warning icon and the vehicle correction direction are displayed on the vehicle display screen, and the voice prompts the owner of the side slip risk level and the direction in which the vehicle should be corrected.

10. The vehicle sideslip detection and warning system according to claim 9, characterized in that: The information transmission processing module substitutes the current vehicle average wheel speed and the longitudinal acceleration of the whole vehicle into the pre-stored speed theory formula to correct the current vehicle speed, and then inputs the corrected vehicle speed and the wheel speed of each wheel into the BIRCH algorithm, constructs a dynamic tree, clusters to find the vehicle speed and the wheel speed centroid of each wheel after a period of time, and finally substitutes them into the vehicle slip rate formula for calculation and compares with the set vehicle sideslip threshold range.

Images