Intelligent networked automobile automatic obstacle avoidance method and system

Abstract

The invention relates to the technical field of automobile automatic obstacle avoidance, in particular to an intelligent networked automobile automatic obstacle avoidance method and system.The method comprises the steps that the lateral speed and the tire pressure of an intelligent networked automobile at all moments during driving and the standard lateral speed of the intelligent networked automobile at all moments during driving at the standard tire pressure are collected; forming a lateral speed sequence and a tire pressure sequence of each moment by using the lateral speed and the tire pressure of each moment and all moments before the moment, and extracting each detection lateral subsequence in the lateral speed sequence of each moment; determining the lane changing confidence coefficient of each detection lateral sub-sequence; classifying the mean values of the tire pressure sequences at all the moments and the historical moments thereof to obtain a tire pressure cluster where the mean values of the tire pressure sequences at all the moments are located; determining the tire pressure interference degree of the tire pressure cluster; obtaining a lateral adjustment coefficient of the tire pressure cluster at each moment; and obtaining the lateral speed of the intelligent networked automobile at each moment in the lane changing and obstacle avoiding process. Therefore, the safety of automatic obstacle avoidance of the automobile is improved.

Description

Technical Field

[0001] The present application relates to the technical field of automatic obstacle avoidance for automobiles, and in particular to an automatic obstacle avoidance method and system for intelligent networked automobiles. Background Art

[0002] With the development of intelligent and connected automobile technology, traditional driving modes are gradually shifting towards automation and intelligent decision-making. Various advanced sensors are applied to intelligent connected vehicles. By combining with technologies such as big data and cloud computing, the vehicles have powerful perception, decision-making and execution capabilities, and can monitor the surrounding environment in real time and respond accordingly. Among them, automatic obstacle avoidance is a core function of intelligent connected vehicles. Through the vehicle's own intelligent system, it can quickly and accurately identify and judge obstacles on the road, and take effective avoidance measures in time, thereby greatly reducing the risk of traffic accidents caused by human errors or environmental complexity, and improving driving safety and road traffic efficiency.

[0003] Automatic obstacle avoidance in smart connected cars is to perform emergency obstacle avoidance to the left and right sides when detecting that the vehicle in front suddenly stops or the speed is abnormal, so as to ensure the safety of the vehicle. When performing obstacle avoidance, the existing technology usually plans an obstacle avoidance route based on the distance between the vehicle and the obstacle in front, and then sets the lateral speed according to the planned obstacle avoidance route to achieve obstacle avoidance for the obstacle in front. However, the existing technology does not deeply consider that under different tire pressures, the lateral speed set by the connected car according to the obstacle avoidance route will deviate from the lateral speed of the car when it is actually driving, which may cause the vehicle to lose control due to excessive lateral speed or fail to avoid obstacles safely due to too slow lateral speed. Summary of the invention

[0004] In order to solve the above technical problems, the purpose of this application is to provide an automatic obstacle avoidance method and system for intelligent connected vehicles. The technical solutions adopted are as follows: In a first aspect, an embodiment of the present application provides a method for automatic obstacle avoidance of an intelligent connected vehicle, the method comprising the following steps: Collect the lateral speed and tire pressure of the intelligent connected vehicle at each time when it is driving, and obtain the standard lateral speed of the intelligent connected vehicle at each time when it is driving with standard tire pressure during the pre-delivery test process; The lateral speed and tire pressure at each moment and all previous moments are respectively combined into a lateral speed sequence and a tire pressure sequence at each moment. Based on the distribution of the lateral speed during the lane change process of the vehicle, each detected lateral subsequence in the lateral speed sequence at each moment is extracted; the lane change confidence of each detected lateral subsequence is determined by the numerical difference and change trend difference of the elements on both sides of the peak in each detected lateral subsequence; Classifying the mean values ​​of the tire pressure sequences at each moment and its historical moments to obtain the tire pressure clusters to which the mean values ​​of the tire pressure sequences at each moment belong; extracting each lateral subsequence from the detected lateral subsequences at the moments corresponding to all the mean values ​​in the tire pressure clusters based on the distribution of the lane change confidence; Analyze the distribution difference between the lateral speed in each lateral subsequence and the standard lateral speed, and the difference between the mean value and the standard tire pressure in the tire pressure cluster, and determine the tire pressure interference degree of the tire pressure cluster; take the number of elements between the first element and the maximum value element in the lateral subsequence as the turning time of the lateral subsequence; Based on the difference between the turning time of all lateral subsequences extracted at each moment and the turning time corresponding to the standard lateral speed, combined with the tire pressure interference degree, the lateral adjustment coefficient of the tire pressure cluster at each moment is determined; using the initial lateral speed of the intelligent connected vehicle when changing lanes to avoid obstacles, and the lateral adjustment coefficient of the tire pressure cluster to which the mean of the tire pressure sequence belongs, the lateral speed of the intelligent connected vehicle at each moment during the lane changing and obstacle avoidance process is determined.

[0005] In one embodiment, extracting each detected lateral subsequence in the lateral velocity sequence at each moment includes: The lateral velocity sequence at each moment is divided into subsequences using a sequence segmentation algorithm, the element mean of each subsequence is calculated, and the segmentation threshold of all the element means is obtained using a threshold segmentation algorithm. The subsequences whose element means are greater than the segmentation threshold are taken as the detected lateral subsequences.

[0006] In one embodiment, the determination of the lane change confidence includes: For each detected lateral subsequence, the detected lateral subsequence is divided into a first subsequence and a second subsequence using the maximum peak value in the detected lateral subsequence, and the difference between the mean values ​​of the elements of the first subsequence and the second subsequence is recorded as the first difference; Respectively obtain the first-order difference sequences of the first subsequence and the second subsequence, calculate the ratio of the number of positive values ​​and the number of negative values ​​in each first-order difference sequence to the number of all values ​​in the first-order difference sequence, calculate the difference between the ratio of the number of positive values ​​and the number of negative values ​​in each first-order difference sequence, and record it as the second difference; use the average of the second difference between the first subsequence and the second subsequence as the monotonic coefficient of both sides of the corresponding detection lateral subsequence; The lane change confidence is positively correlated with the monotonic coefficient on both sides, and negatively correlated with the number of peaks in the detected lateral subsequence and the first difference.

[0007] In one embodiment, the extracting each lateral subsequence comprises: A threshold segmentation algorithm is used to obtain a segmentation threshold of lane change confidences of all detected lateral subsequences at all corresponding moments in the tire pressure cluster, and the detected lateral subsequences whose lane change confidences are greater than the segmentation threshold are taken as each lateral subsequence.

[0008] In one embodiment, the determination of the tire pressure interference degree includes: The standard lateral speeds at all moments are combined into a standard lateral speed sequence, and based on the standard lateral speed sequence, each standard lateral subsequence is obtained by using the same method as that of obtaining each lateral subsequence from the lateral speed sequence; For the tire pressure cluster, calculate the metric distance between any lateral subsequence and any standard lateral subsequence, and use the average of all the metric distances as the lateral deviation of the tire pressure cluster; record the difference between the mean of the data within the cluster and the standard tire pressure as the third difference, and record the difference between the mean of the elements of all lateral subsequences of the tire pressure cluster and the mean of the elements of all standard lateral subsequences as the fourth difference; The tire pressure interference degree is positively correlated with the lateral deviation degree and the fourth difference, and is negatively correlated with the third difference.

[0009] In one embodiment, determining the lateral adjustment coefficient includes: The turning time of each standard lateral subsequence is determined by the same method as that of obtaining the turning time of each lateral subsequence. For any tire pressure cluster, the difference between the average turning time of all corresponding lateral subsequences and the average turning time of all standard lateral subsequences is calculated as the lane change responsiveness of the any tire pressure cluster. The lateral adjustment coefficient is determined based on the lane change responsiveness and tire pressure interference of any tire pressure cluster and the lateral speed difference of the connected vehicle at adjacent moments.

[0010] In one embodiment, the expression of the lateral adjustment coefficient is: ; In the formula, is the lateral adjustment coefficient of the vth tire pressure cluster at the xth moment, is the tire pressure interference of the vth tire pressure cluster, is the lane change responsiveness of the vth tire pressure cluster, sign() is the sign function, is the lateral speed of the connected car at the xth moment, is the lateral speed of the connected vehicle at the x-1th moment.

[0011] In one embodiment, determining the lateral speed of the intelligent connected vehicle at each moment during the lane change and obstacle avoidance process includes: The tire pressure sequence at the time when the intelligent connected vehicle needs to change lanes for obstacle avoidance is used to obtain a tire pressure cluster to which the average value of the tire pressure sequence at the time of the lane change for obstacle avoidance belongs, and the normalized value of the lateral adjustment coefficient of the tire pressure cluster at each moment in the process of the intelligent connected vehicle changing lanes for obstacle avoidance is calculated. The lateral speed of the intelligent connected vehicle at each moment in the process of the lane change for obstacle avoidance is determined by combining the normalized value with the initial lateral speed.

[0012] In one embodiment, the sum of the normalized value and the value 1 is calculated, and the lateral speed at each moment in the process of the intelligent connected vehicle changing lanes and avoiding obstacles is the product of the initial lateral speed and the sum.

[0013] In a second aspect, an embodiment of the present application further provides an automatic obstacle avoidance system for an intelligent connected vehicle, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of any one of the above methods when executing the computer program.

[0014] This application has at least the following beneficial effects: The present application acquires the standard lateral speed of the intelligent connected vehicle at each moment when it is running with the standard tire pressure during the pre-delivery test by collecting the lateral speed and tire pressure at each moment; through the coordinated monitoring of the lateral speed and tire pressure, the comprehensive perception of the dynamic state of the vehicle is improved, and the reliability of obstacle avoidance decision-making is enhanced; the lateral speed and tire pressure at each moment and all previous moments are respectively composed into a lateral speed sequence and a tire pressure sequence at each moment; based on the element distribution in the lateral speed sequence, each detected lateral subsequence in the lateral speed sequence at each moment is extracted; by extracting the elements on both sides of the peak value in each detected lateral subsequence The numerical difference and change trend difference of the elements are used to determine the lane change confidence of each detected lateral subsequence; the dynamic evaluation of the lane change confidence enhances the accuracy of vehicle intention recognition and helps to distinguish normal driving from emergency obstacle avoidance requirements under complex road conditions; the mean of the tire pressure sequence at each moment and its historical moments is classified to obtain the tire pressure cluster where the mean of the tire pressure sequence at each moment is located; based on the distribution of the lane change confidence, each lateral subsequence is extracted from the detected lateral subsequences at the moment corresponding to all the means in the tire pressure cluster to obtain the lateral subsequences; the distribution difference between the lateral speed in each lateral subsequence and the standard lateral speed, as well as the mean in the tire pressure cluster, are analyzed. The difference between the value and the standard tire pressure is used to determine the tire pressure interference of the tire pressure cluster; the tire pressure interference realizes the quantification of the impact of tire pressure changes on lateral speed, so that the adjustment range of lateral speed can be dynamically evaluated according to different tire pressure conditions, improving the accuracy and reliability of lateral speed adjustment; the number of elements between the first element and the maximum value element in the lateral subsequence is used as the turning time of the lateral subsequence; based on the difference between the turning time of all lateral subsequences extracted at each moment and the turning time corresponding to the standard lateral speed, combined with the tire pressure interference, the lateral adjustment coefficient of the tire pressure cluster at each moment is determined; the lateral adjustment coefficient can clearly indicate the difference between the turning time of all lateral subsequences extracted at each moment and the turning time corresponding to the standard lateral speed. The adjustment direction and amplitude of the lateral speed under the same tire pressure conditions can avoid the risk of vehicle loss of control or obstacle avoidance failure caused by abnormal tire pressure. The initial lateral speed of the intelligent connected vehicle when changing lanes to avoid obstacles and the lateral adjustment coefficient of the tire pressure cluster to which the tire pressure sequence mean belongs are used to determine the lateral speed of the intelligent connected vehicle at each moment in the process of changing lanes to avoid obstacles. The lateral speed of the vehicle can be effectively adjusted dynamically according to the tire pressure of the vehicle when avoiding obstacles, achieving coordinated compensation of environmental interference and vehicle status, overcoming the problem of understeering or oversteering caused by tire pressure fluctuations, improving the obstacle avoidance effect of the vehicle, and avoiding additional traffic risks. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0016] Figure 1 A flowchart of a method for automatic obstacle avoidance for an intelligent networked vehicle provided in accordance with an embodiment of the present application; Figure 2 Flow chart for adjusting lateral speed for car lane change and obstacle avoidance. DETAILED DESCRIPTION

[0017] In order to further explain the technical means and effects adopted by this application to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, features and effects of an automatic obstacle avoidance method and system for an intelligent connected vehicle proposed in this application, in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0018] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.

[0019] The following is a detailed description of a specific solution of an automatic obstacle avoidance method and system for an intelligent connected vehicle provided by the present application in conjunction with the accompanying drawings.

[0020] See also Figure 1 , which shows a flowchart of a method for automatic obstacle avoidance of an intelligent networked vehicle provided by an embodiment of the present application, the method comprising the following steps: S1, collecting the lateral speed and tire pressure of the intelligent connected vehicle at each time when it is running, and obtaining the standard lateral speed of the intelligent connected vehicle at each time when it is running with the standard tire pressure during the pre-delivery test process.

[0021] This embodiment takes any normally traveling intelligent networked vehicle as an example for analysis, and collects the lateral speed of the intelligent networked vehicle in real time through the on-board IMU sensor of the intelligent networked vehicle, specifically: the acceleration and angular velocity of the intelligent networked vehicle in the three-dimensional direction of the three-dimensional coordinate system are obtained in real time, and then, according to the acceleration and angular velocity data, the lateral speed value of the intelligent networked vehicle is obtained through the vehicle dynamics model. The lateral speed in this embodiment is the specific value of the lateral speed, regardless of the direction; the tire pressure of the intelligent networked vehicle is collected in real time through the central control system of the intelligent networked vehicle, specifically: since the front wheel is the steering wheel of the vehicle, its tire pressure directly affects the steering response speed of the vehicle, therefore, this embodiment obtains the two front tire pressures of the intelligent networked vehicle in real time through the central control system of the intelligent networked vehicle, and calculates the average of the two front tire pressures as the tire pressure of the intelligent networked vehicle analyzed in this embodiment. At the same time, the standard tire pressure of the intelligent networked vehicle at the time of leaving the factory and the standard lateral speed of the intelligent networked vehicle at each moment when the vehicle is driven with the standard tire pressure during the factory test are obtained through the central control system. The process of obtaining the vehicle's standard tire pressure and standard lateral speed is the same as the process of obtaining the real-time lateral speed and tire pressure of an intelligent connected vehicle.

[0022] Among them, obtaining the lateral speed of the vehicle through the IMU sensor and the vehicle dynamics model is an existing well-known technology, and the specific process is not repeated; the lateral speed and tire pressure of the intelligent networked vehicle are collected synchronously, and the collection frequency of the lateral speed, tire pressure, and standard lateral speed of the intelligent networked vehicle is 10Hz, which can be set by the implementer according to the actual situation, and this embodiment does not limit it here. It should be noted that the tire pressure of the intelligent networked vehicle analyzed later in this embodiment is the average of the two front tire pressures.

[0023] In order to eliminate the dimensional influence between the data, the collected lateral speed, tire pressure, and standard lateral speed are normalized respectively. This embodiment adopts the maximum-minimum value normalization processing method. The implementer can choose other existing feasible normalization methods at will, and this embodiment does not limit it here.

[0024] It should be noted that, when the vehicle speed is slow, the connected car can usually avoid obstacles by emergency braking without changing lanes. When emergency lane change is required, the impact of the vehicle speed change on the lateral speed in a very short time is not particularly significant, and the impact is usually reflected indirectly through steering input and path curvature. Therefore, this embodiment does not take the vehicle speed into consideration when avoiding obstacles.

[0025] S2, the lateral speed and tire pressure at each moment and all previous moments are respectively composed into a lateral speed sequence and a tire pressure sequence at each moment, and based on the element distribution in the lateral speed sequence, each detected lateral subsequence in the lateral speed sequence at each moment is extracted; and the lane change confidence of each detected lateral subsequence is determined by the numerical difference and change trend difference of the elements on both sides of the peak in each detected lateral subsequence.

[0026] Since the driving direction of a connected car will inevitably change when it changes lanes to avoid obstacles, lane changing and turning operations will occur. Therefore, in order to accurately analyze the impact of tire pressure on the lateral speed of a connected car, it is first necessary to filter out the tire pressure data of the connected car when it changes lanes and turns from the historical driving data of the connected car.

[0027] When the connected car is driving in a straight line without directional deviation, its lateral speed will be 0; when the connected car is changing lanes or turning normally, lateral acceleration will be generated, so that the lateral speed will increase from zero. When the lateral speed reaches the maximum value, it means that the vehicle is changing lanes and turning with the maximum lateral displacement; when the car changes lanes and turns, and continues to drive in a straight line, its lateral speed begins to gradually decrease and returns to zero. Therefore, when the vehicle changes lanes and turns, the lateral speed of the vehicle will gradually increase from zero to a peak value, and then gradually decrease to zero. However, considering that if there are external interferences such as side winds and road inclinations when the vehicle is driving, the lateral speed will also change, so this embodiment divides the lateral speed data.

[0028] In this embodiment, the lateral speed at each moment and all previous moments are combined into a lateral speed sequence at each moment in chronological order, the tire pressure at each moment and all previous moments are combined into a tire pressure sequence at each moment in chronological order, and the standard lateral speeds at all moments during the factory test of the intelligent connected vehicle are combined into a standard lateral speed sequence in chronological order.

[0029] For the lateral speed sequence at each moment, this embodiment divides the lateral speed sequence into multiple subsequences through a sequence segmentation algorithm. Since the lateral speed of the connected vehicle is zero when it is traveling in a straight line and is not disturbed by the outside world, the subsequences with changed lateral speeds can be selected through the mean difference of each subsequence. The sequence segmentation algorithm used in this embodiment is the BG (Bernaola Galvan) sequence segmentation algorithm, which is an existing well-known technology. The implementer can choose other existing feasible sequence segmentation algorithms at will, and this embodiment does not limit it here.

[0030] For each subsequence after the lateral speed sequence at each moment is segmented, the mean of all elements in each subsequence is calculated and recorded as the first mean, and the first mean of all subsequences of the lateral speed sequence is used as the input of the Otsu threshold method to obtain the segmentation threshold, and the subsequences whose first mean is greater than the segmentation threshold are used as each detected lateral subsequence. The Otsu threshold method can extract subsequences with large fluctuations in the lateral speed sequence, which can reduce the amount of data processed and improve efficiency on the one hand, and avoid the false detection phenomenon that a subsequence with all 0s may be mistakenly detected as a subsequence when the car changes lanes on the other hand.

[0031] This embodiment takes the u-th detected lateral subsequence after the lateral velocity sequence at any moment is divided as an example for analysis. First, the peak value in the u-th detected lateral subsequence is obtained by the peak detection algorithm, and the u-th detected lateral subsequence is divided into a first subsequence and a second subsequence according to the maximum value of all peak values. That is, all elements on the left side of the maximum peak value in the u-th detected lateral subsequence form the first subsequence, and all elements on the right side of the maximum peak value form the second subsequence. For example, the u-th detected lateral subsequence is [1, 2, 3, 5, 4, 2, 3], then the maximum peak value therein is 5, and all elements on the left side of the maximum peak value are 1, 2, 3, then the first subsequence is [1, 2, 3]. Similarly, all elements on the right side of the maximum peak value are 4, 2, 3, then the second subsequence is [4, 2, 3]. It should be understood that if there is an element U in the detected lateral subsequence that is greater than the elements on its left and right sides, then the element U is a peak value. The mean of all elements in the first subsequence is calculated and recorded as the second mean. The mean of all elements in the second subsequence is calculated and recorded as the third mean. The difference between the second mean and the third mean is recorded as the first difference. The smaller the first difference is, the more similar the data distribution on both sides of the peak value in the u-th detected lateral subsequence is, and the more consistent it is with the lateral speed characteristics during normal lane change or turning. Among them, this embodiment adopts the AMPD peak detection algorithm, which is an existing well-known technology. The implementer can choose other existing feasible peak detection algorithms at will, and this embodiment does not limit it here.

[0032] It should be noted that the difference indicates the degree of difference between two variables, and can be calculated specifically by using the absolute value of the difference, the square of the difference, the ratio, etc. In this embodiment, the absolute value of the difference between the second mean and the third mean is recorded as the first difference.

[0033] Obtain the first-order difference sequence P1 of the first subsequence and the first-order difference sequence P2 of the second subsequence, and calculate the ratio z1 of the number of positive values ​​in the first-order difference sequence P1 to the number of all values ​​in the first-order difference sequence P1, and the ratio z2 of the number of negative values ​​in the first-order difference sequence P1 to the number of all values ​​in the first-order difference sequence P1, record the absolute value of the difference between z1 and z2 as the second difference, and accordingly, for the first-order difference sequence P2, adopt the same method as that for calculating the second difference of the first-order difference sequence P1 to obtain the second difference corresponding to the first-order difference sequence P2, and use the average of the second differences between the first-order difference sequence P1 and the first-order difference sequence P2 as the monotonic coefficient on both sides of the u-th detection lateral subsequence. The second difference of each first-order difference sequence can reflect whether the signs of the elements in each first-order difference sequence are all positive or all negative; the monotonic coefficients on both sides can reflect whether the data on both sides of the maximum value of the u-th detected lateral subsequence conform to the changing characteristics of the lateral speed data when the vehicle changes lanes or turns.

[0034] Based on the above analysis, this embodiment calculates the lane change confidence of the u-th detected lateral subsequence, and the specific calculation method is: ; In the formula, is the lane change confidence of the u-th detected lateral subsequence, is the monotonic coefficient on both sides of the u-th detection lateral subsequence, is the number of peaks in the u-th detected lateral subsequence, is the first difference of the u-th detected lateral subsequence, is a preset value greater than 0. In order to avoid the denominator being 0, in this embodiment , implementers can set it according to actual conditions.

[0035] It should be understood that the monotonic coefficients on both sides The larger it is, the more the u-th detected lateral sub-sequence data conforms to the characteristics of the vehicle changing lanes normally and turning, that is, the lateral speed gradually increases and then gradually decreases. If it is 1, it means that the u-th detected lateral subsequence has only one peak, which is consistent with the characteristics of the vehicle changing lanes and turning. When the value is larger or equal to 0, it reflects that the lateral speed in the u-th detected lateral subsequence is more likely to be affected by external interference. It can reflect whether the change of lateral speed is smooth and symmetrical. It can comprehensively reflect whether the lateral speed data change in the u-th detected lateral subsequence conforms to the data characteristics of normal lane change or turning. The greater the lane change confidence, the more the lateral speed distribution in the u-th detected lateral subsequence conforms to the numerical distribution characteristics of normal lane change of the vehicle, thereby effectively screening out subsequences that conform to the lane change and turning characteristics from complex driving data.

[0036] S3, classifying the means of the tire pressure sequences at each moment and its historical moments to obtain the tire pressure cluster to which the means of the tire pressure sequences at each moment belong; extracting each lateral subsequence from the detected lateral subsequences at all corresponding moments in the tire pressure cluster based on the distribution of lane change confidence of the detected lateral subsequence.

[0037] In order to more accurately analyze the impact of tire pressure on lateral speed, this embodiment performs cluster analysis on tire pressure data to ensure that the analysis results are referenceable under the same tire pressure conditions. This embodiment obtains the tire pressure sequence of each moment and the N moments closest to it. In this embodiment, N=30, which can be set by the implementer according to the actual situation, and this embodiment does not limit it here. Calculate the mean of all tire pressures in each tire pressure sequence, recorded as the fourth mean, and for any moment, use the fourth mean of the tire pressure sequence of any moment and the N moments closest to it as the input of the DPC (Density Peak Clustering) density peak clustering algorithm, and output each cluster cluster, recorded as a tire pressure cluster, each tire pressure cluster represents a tire pressure condition. This embodiment uses a cross-validation method to obtain the cutoff distance of the DPC clustering algorithm. The DPC density peak clustering algorithm and the cross-validation method are existing well-known technologies. The implementer can choose other existing feasible clustering algorithms, and this embodiment does not limit it here.

[0038] This embodiment takes the vth tire pressure cluster at any moment as an example for analysis. Since the tire pressure sequence of this embodiment is the tire pressure sequence at each moment, each tire pressure sequence corresponds to one moment, so the lateral speed sequence at the corresponding moment of the tire pressure sequence corresponding to each of the fourth mean values ​​in the vth tire pressure cluster is obtained, and the lane change confidence of all detected lateral subsequences corresponding to the lateral speed sequences at all corresponding moments in the vth tire pressure cluster is used as the input of the Otsu threshold method, and the output is the segmentation threshold, which is recorded as the first segmentation threshold, and the detected lateral subsequences with lane change confidence greater than the first segmentation threshold are used as each lateral subsequence.

[0039] Similarly, for the standard lateral speed sequence, the lane change confidence of each detected lateral subsequence is calculated after segmentation using the lateral speed sequence at any moment, and the same calculation method as each lateral subsequence is further obtained. The standard lateral speed sequence is segmented, and the lane change confidence of each segmented sequence is calculated. The lane change confidence is used to perform threshold segmentation to obtain each standard lateral subsequence.

[0040] By extracting lateral subsequences and standard lateral subsequences through the Otsu threshold method, the lateral speed data of the connected vehicle when changing lanes or turning can be extracted, so as to further analyze the difference in lateral speed caused by tire pressure differences.

[0041] S4, analyzing the distribution difference between the lateral speed in each lateral subsequence and the standard lateral speed, as well as the difference between the fourth mean value in the tire pressure cluster and the standard tire pressure, to determine the tire pressure interference degree of the tire pressure cluster; taking the number of elements from the first element to the maximum value element in the lateral subsequence as the turning time of the lateral subsequence.

[0042] For all lateral subsequences corresponding to the vth tire pressure cluster, calculate the metric distance between any lateral subsequence and any standard lateral subsequence, and calculate the average of the metric distances of all lateral subsequences of the vth tire pressure cluster as the lateral deviation of the vth tire pressure cluster. The larger the lateral deviation, the greater the deviation between the lateral speed change characteristics caused by the tire pressure change and the standard lateral speed change characteristics, which reflects that the tire pressure change has a greater impact on the vehicle's lane change, and the greater the need for adjusting the lateral speed. Among them, the metric distance described in this implementation is calculated using the DTW (Dynamic Time Warping) distance, and the implementer can choose other existing feasible metric distance calculation methods.

[0043] The absolute value of the difference between the mean of all the fourth means in the vth tire pressure cluster and the standard tire pressure is recorded as the third difference. The absolute value of the difference between the mean of all lateral subsequence elements corresponding to the vth tire pressure cluster and the mean of all standard lateral subsequence elements is calculated and recorded as the fourth difference. The tire pressure interference degree of the vth tire pressure cluster is calculated as follows: ; In the formula, is the tire pressure interference degree of the vth tire pressure cluster; is the lateral deviation of the vth tire pressure cluster; is the fourth difference of the vth tire pressure cluster; is the third difference of the vth tire pressure cluster; is a preset value greater than 0. In order to avoid the denominator being 0, in this embodiment The implementer can set it according to the actual situation, and this embodiment does not limit it here.

[0044] It should be understood that the lateral deviation It can measure the degree of deviation between the lateral speeds when the vehicle changes lanes or turns under different tire pressures, thereby reflecting the significance of the impact of tire pressure on lateral speed. The influence of a unit change in tire pressure on lateral speed can be quantified. The greater the influence, the greater the tire pressure interference. It can reflect whether the influence of the vth tire pressure condition on the lateral speed is significant. The greater the tire pressure interference, the more significant the influence of the vth tire pressure condition on the lateral speed.

[0045] During the process of changing lanes and turning, the grip and steering direction generated by the tire pressure will change with the turning stage. During the initial turning, the steering angle is large; when the steering angle reaches the maximum value, the steering must be reversed to return the vehicle to the correct driving angle and maintain normal driving. Therefore, the entire turning process includes two directions: turning direction and returning to the correct direction.

[0046] Since tire pressure will cause changes in lateral speed, the lateral speed needs to be adjusted according to the difference between the tire pressure and the standard tire pressure in different steering stages, so as to more accurately adapt to lane changing needs and improve the vehicle's handling and stability.

[0047] Furthermore, in each lateral subsequence and each standard lateral subsequence, the number of elements from the first element to the maximum value element in the lateral subsequence is recorded as the turning time of each lateral subsequence, and accordingly, the turning time of each standard lateral subsequence is obtained.

[0048] S5, based on the difference between the turning time of all lateral subsequences extracted at each moment and the turning time corresponding to the standard lateral speed, combined with the tire pressure interference degree, determine the lateral adjustment coefficient of the tire pressure cluster at each moment; use the initial lateral speed of the intelligent connected vehicle when changing lanes to avoid obstacles, and the lateral adjustment coefficient of the tire pressure cluster to which the mean of the tire pressure sequence belongs, to determine the lateral speed of the intelligent connected vehicle at each moment during the lane changing and obstacle avoidance process.

[0049] Taking the vth tire pressure cluster as an example, the difference between the mean turning time a1 of all corresponding lateral subsequences and the mean turning time a2 of all standard lateral subsequences is calculated as the lane change responsiveness of the vth tire pressure cluster. The lane change responsiveness can reflect the impact of tire pressure difference on the lateral speed of the vehicle when turning. If the lane change responsiveness is greater than 0, it means that the actual time spent on lane change of the vehicle has been extended under the vth tire pressure condition. At this time, the lateral speed needs to be increased to ensure timely obstacle avoidance; if the lane change responsiveness is less than 0, it means that the time spent on lane change of the vehicle has been reduced. At this time, the lateral speed needs to be reduced to avoid excessive deviation of the vehicle; if the lane change responsiveness is 0, it means that there is no need to adjust the lateral speed.

[0050] Based on the above analysis, this embodiment calculates the lateral adjustment coefficient of each tire pressure cluster at each time, and the specific calculation method is: ; In the formula, is the lateral adjustment coefficient of the vth tire pressure cluster at the xth moment, is the tire pressure interference of the vth tire pressure cluster, is the lane change responsiveness of the vth tire pressure cluster, sign() is the sign function, when the internal parameter of the sign function is less than 0, it outputs -1, when the internal parameter of the sign function is greater than 0, it outputs 1, when the internal parameter of the sign function is equal to 0, it outputs 0, is the lateral speed of the connected car at the xth moment, is the lateral speed of the connected car at the x-1th moment. Since the lateral speed of the car will gradually increase when turning, if , it reflects that the car is turning at the current moment. This means that the lane change has been completed and the direction is returning to the correct direction.

[0051] It should be understood that if the vehicle is currently turning, the tire pressure will interfere with the steering efficiency of the vehicle, so the lateral speed needs to be adjusted to ensure the safe obstacle avoidance of the connected car. If the lane change operation has been completed to achieve obstacle avoidance and the vehicle is returning to the correct direction, the lateral speed of the vehicle is gradually decreasing and tending to stabilize. Adjusting the lateral speed at this time may cause the vehicle to swing or lose control, increasing additional traffic risks. Therefore, there is no need to make major adjustments to maintain the overall balance and stability of the vehicle. The lateral adjustment coefficient can comprehensively reflect the demand direction and adjustment intensity of the vehicle's lateral speed adjustment due to tire pressure changes in different steering stages, ensuring that the lateral speed adjustment strategy can both respond to abnormal tire pressure and adapt to dynamic driving scenarios. The larger the lateral adjustment coefficient, the greater the need to adjust the lateral speed in the vth tire pressure situation.

[0052] The autonomous obstacle avoidance system of a connected car can monitor the surrounding environment in real time through sensors, radars, cameras and other devices, identify potential obstacles, and provide environmental information for obstacle avoidance. When the vehicle detects that it needs to change lanes to avoid obstacles, it will plan an obstacle avoidance path based on the information of surrounding obstacles, and then automatically avoid obstacles by controlling the lateral speed and steering angle of the connected car. Among them, planning the obstacle avoidance path of the connected car is an existing well-known technology, and the specific process will not be repeated.

[0053] When the vehicle needs to change lanes to avoid obstacles, first obtain the tire pressure sequence at the starting time of the vehicle's lane change and obstacle avoidance, cluster the mean of the tire pressure sequences at the starting time and the N nearest moments before the starting time using the DPC density peak clustering algorithm to obtain each tire pressure cluster, and obtain the tire pressure cluster where the mean of the tire pressure sequence at the starting time is located. For example, the tire pressure cluster where the mean of the tire pressure sequence at the starting time is located is the vth tire pressure cluster. In order to improve the obstacle avoidance effect for the connected vehicle and avoid the phenomenon of vehicle loss of control or inability to completely avoid obstacles, this embodiment optimizes the lateral speed of the vehicle at each moment in the process of lane change and obstacle avoidance according to the lateral adjustment coefficient, and determines the lateral speed of the intelligent connected vehicle at each moment in the process of lane change and obstacle avoidance. Taking the xth moment as an example, the specific expression is: ; In the formula, is the lateral speed at the xth moment after adjustment under the vth tire pressure condition, that is, the lateral speed at the xth moment during the lane change and obstacle avoidance process of the intelligent connected vehicle, is the initial lateral speed of the vehicle at the start of lane change to avoid obstacles, is the lateral adjustment coefficient of the vth tire pressure cluster at the xth moment, is a normalization function, which normalizes the lateral adjustment coefficient to between (-1,1). It should be noted that (-1,1) is an open interval. The flow chart of the lateral speed adjustment for lane change and obstacle avoidance is as follows: Figure 2 shown.

[0054] From the above expression, we can see that is the normalized value of the lateral adjustment coefficient of the vth tire pressure cluster at the xth moment, is the sum of the normalized value and the value 1, and the adjusted lateral speed at the xth moment under the vth tire pressure condition is the product of the initial lateral speed at the start moment of the vehicle changing lanes to avoid obstacles and the sum. The same calculation method as the lateral speed at the xth moment in the process of the intelligent connected vehicle changing lanes to avoid obstacles is used to obtain the lateral speed at each moment in the process of the intelligent connected vehicle changing lanes to avoid obstacles, that is, the adjusted lateral speed at each moment in the process of the intelligent connected vehicle changing lanes to avoid obstacles.

[0055] Finally, the lateral speed adjusted at each moment during the vehicle's lane change and obstacle avoidance process is used to change lanes and avoid obstacles according to the planned obstacle avoidance path, ensuring that the vehicle can travel according to the planned obstacle avoidance path, improving the vehicle's obstacle avoidance effect, avoiding additional traffic risks, and ensuring the stability and safety of the obstacle avoidance process.

[0056] Based on the same inventive concept as the above method, an embodiment of the present application also provides an automatic obstacle avoidance system for an intelligent connected vehicle, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor implements the steps of any one of the above-mentioned automatic obstacle avoidance methods for an intelligent connected vehicle when executing the computer program.

[0057] It should be noted that the above sequence of the embodiments of the present application is for description only and does not represent the advantages and disadvantages of the embodiments. The above is a description of a specific embodiment of this specification. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0058] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

[0059] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the principles of the present application should be included in the protection scope of the present application.

Claims

1. A method for automatic obstacle avoidance of an intelligent networked vehicle, characterized in that: The method comprises the following steps: Collect the lateral speed and tire pressure of the intelligent connected vehicle at each time when it is driving, and obtain the standard lateral speed of the intelligent connected vehicle at each time when it is driving with standard tire pressure during the pre-delivery test process; The lateral speed and tire pressure at each moment and all previous moments are respectively combined into a lateral speed sequence and a tire pressure sequence at each moment. Based on the distribution of the lateral speed during the lane change process of the vehicle, each detected lateral subsequence in the lateral speed sequence at each moment is extracted; the lane change confidence of each detected lateral subsequence is determined by the numerical difference and change trend difference of the elements on both sides of the peak in each detected lateral subsequence; Classifying the mean values ​​of the tire pressure sequences at each moment and its historical moments to obtain the tire pressure clusters to which the mean values ​​of the tire pressure sequences at each moment belong; extracting each lateral subsequence from the detected lateral subsequences at the moments corresponding to all the mean values ​​in the tire pressure clusters based on the distribution of the lane change confidence; Analyze the distribution difference between the lateral speed in each lateral subsequence and the standard lateral speed, and the difference between the mean value and the standard tire pressure in the tire pressure cluster, and determine the tire pressure interference degree of the tire pressure cluster; take the number of elements between the first element and the maximum value element in the lateral subsequence as the turning time of the lateral subsequence; Based on the difference between the turning time of all lateral subsequences extracted at each moment and the turning time corresponding to the standard lateral speed, combined with the tire pressure interference degree, the lateral adjustment coefficient of the tire pressure cluster at each moment is determined; using the initial lateral speed of the intelligent connected vehicle when changing lanes to avoid obstacles, and the lateral adjustment coefficient of the tire pressure cluster to which the mean of the tire pressure sequence belongs, the lateral speed of the intelligent connected vehicle at each moment during the lane changing and obstacle avoidance process is determined.

2. The automatic obstacle avoidance method for an intelligent networked vehicle according to claim 1, characterized in that: The step of extracting each detected lateral subsequence from the lateral velocity sequence at each moment includes: The lateral velocity sequence at each moment is divided into subsequences using a sequence segmentation algorithm, the element mean of each subsequence is calculated, and the segmentation threshold of all the element means is obtained using a threshold segmentation algorithm. The subsequences whose element means are greater than the segmentation threshold are taken as the detected lateral subsequences.

3. The automatic obstacle avoidance method for an intelligent networked vehicle according to claim 1, characterized in that: The determination of the lane change confidence includes: For each detected lateral subsequence, the detected lateral subsequence is divided into a first subsequence and a second subsequence using the maximum peak value in the detected lateral subsequence, and the difference between the mean values ​​of the elements of the first subsequence and the second subsequence is recorded as the first difference; Respectively obtain the first-order difference sequences of the first subsequence and the second subsequence, calculate the ratio of the number of positive values ​​and the number of negative values ​​in each first-order difference sequence to the number of all values ​​in the first-order difference sequence, calculate the difference between the ratio of the number of positive values ​​and the number of negative values ​​in each first-order difference sequence, and record it as the second difference; use the average of the second difference between the first subsequence and the second subsequence as the monotonic coefficient of both sides of the corresponding detection lateral subsequence; The lane change confidence is positively correlated with the monotonic coefficient on both sides, and negatively correlated with the number of peaks in the detected lateral subsequence and the first difference.

4. The automatic obstacle avoidance method for an intelligent networked vehicle according to claim 1, characterized in that: The extraction obtains each lateral subsequence, including: A threshold segmentation algorithm is used to obtain a segmentation threshold of lane change confidences of all detected lateral subsequences at all corresponding moments in the tire pressure cluster, and the detected lateral subsequences whose lane change confidences are greater than the segmentation threshold are taken as each lateral subsequence.

5. The automatic obstacle avoidance method for an intelligent networked vehicle according to claim 1, characterized in that: The determination of the tire pressure interference degree includes: The standard lateral speeds at all moments are combined into a standard lateral speed sequence, and based on the standard lateral speed sequence, each standard lateral subsequence is obtained by using the same method as that of obtaining each lateral subsequence from the lateral speed sequence; For the tire pressure cluster, calculate the metric distance between any lateral subsequence and any standard lateral subsequence, and use the average of all the metric distances as the lateral deviation of the tire pressure cluster; record the difference between the mean of the data within the cluster and the standard tire pressure as the third difference, and record the difference between the mean of the elements of all lateral subsequences of the tire pressure cluster and the mean of the elements of all standard lateral subsequences as the fourth difference; The tire pressure interference degree is positively correlated with the lateral deviation degree and the fourth difference, and is negatively correlated with the third difference.

6. The automatic obstacle avoidance method for an intelligent networked vehicle as claimed in claim 5, characterized in that: The determination of the lateral adjustment coefficient includes: The turning time of each standard lateral subsequence is determined by the same method as that of obtaining the turning time of each lateral subsequence. For any tire pressure cluster, the difference between the average turning time of all corresponding lateral subsequences and the average turning time of all standard lateral subsequences is calculated as the lane change responsiveness of the any tire pressure cluster. The lateral adjustment coefficient is determined based on the lane change responsiveness and tire pressure interference of any tire pressure cluster and the lateral speed difference of the connected vehicle at adjacent moments.

7. The automatic obstacle avoidance method for an intelligent networked vehicle according to claim 6, characterized in that: The expression of the lateral adjustment coefficient is: ; In the formula, is the lateral adjustment coefficient of the vth tire pressure cluster at the xth moment, is the tire pressure interference of the vth tire pressure cluster, is the lane change responsiveness of the vth tire pressure cluster, sign() is the sign function, is the lateral speed of the connected car at the xth moment, is the lateral speed of the connected vehicle at the x-1th moment.

8. The automatic obstacle avoidance method for an intelligent networked vehicle according to claim 1, characterized in that: The determining of the lateral speed at each moment during the process of the intelligent connected vehicle changing lanes and avoiding obstacles includes: The tire pressure sequence at the time when the intelligent connected vehicle needs to change lanes for obstacle avoidance is used to obtain a tire pressure cluster to which the average value of the tire pressure sequence at the time of the lane change for obstacle avoidance belongs, and the normalized value of the lateral adjustment coefficient of the tire pressure cluster at each moment in the process of the intelligent connected vehicle changing lanes for obstacle avoidance is calculated. The lateral speed of the intelligent connected vehicle at each moment in the process of the lane change for obstacle avoidance is determined by combining the normalized value with the initial lateral speed.

9. The automatic obstacle avoidance method for an intelligent networked vehicle as claimed in claim 8, characterized in that: The sum of the normalized value and the value 1 is calculated, and the lateral speed at each moment in the process of the intelligent connected vehicle changing lanes and avoiding obstacles is the product of the initial lateral speed and the sum.

10. An automatic obstacle avoidance system for an intelligent network-connected vehicle, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 9 are implemented.

Images