Traffic signal data automatic classification control method based on target identification

Abstract

The invention discloses a traffic signal data automatic classification control method based on target identification, relates to the technical field of signal processing, is used for solving the problems of sudden accidents and the like caused by the fact that target vehicle characteristics are not received in time at a fixed frequency, and comprises the steps of setting an exploration time mechanism; when an emergency occurs, a plurality of laser radar detection frequencies are set for screening within the exploration time, the screened laser radar detection frequencies are used for detecting the motion state of a target vehicle, and the vehicle speed and the motion state of the target vehicle are combined to analyze and determine the abnormal degree of the event by using a reinforcement learning method. The system transmits pulses and receives reflected signals to determine the type of a small vehicle target, adjusts the detection frequency of the laser radar by using an analytic hierarchy process in combination with the abnormal degree of an event and the type of the small vehicle target, dynamically updates the touch time by using the detection frequency of the laser radar, calculates a safe distance according to the touch time, and sets an alarm threshold value. When the distance between the vehicles reaches an alarm threshold value, an alarm is given.

Description

Technical Field

[0001] The present invention relates to the technical field of signal processing, and more specifically, to an automatic classification control method for traffic signal data based on target recognition. Background Art

[0002] Signal processing technology plays a vital role in LiDAR and other sensor systems. It is mainly used to improve data quality, enhance signal characteristics and realize real-time analysis. The application of signal processing technology in traffic signal target recognition can predict accidents and play an important auxiliary role in helping drivers avoid them in advance.

[0003] The prior art has the following deficiencies:

[0004] In the past, lidar received the target vehicle characteristics through a preset pulse transmission frequency and made subsequent judgments and processing. When the target vehicle changed its motion state and trajectory, the preset fixed frequency would not receive the target vehicle characteristics in time and could be easily affected by emergencies, leading to traffic problems such as traffic accidents. Summary of the invention

[0005] In order to overcome the above-mentioned defects of the prior art, an embodiment of the present invention provides an automatic classification and control method for traffic signal data based on target recognition, which dynamically adjusts the laser radar detection frequency and the alarm threshold by analyzing the abnormal behavior of the target vehicle and the risk level caused by the collision between the target vehicle and the own vehicle to solve the problems raised in the above-mentioned background technology.

[0006] To achieve the above object, the present invention provides the following technical solutions:

[0007] A method for automatically classifying and controlling traffic signal data based on target recognition comprises the following steps:

[0008] Step S1, setting an exploration time, setting multiple laser radar detection frequencies within the exploration time to detect the signal-to-noise ratio respectively, and selecting the maximum signal-to-noise ratio as the initial laser radar detection frequency;

[0009] Step S2, using the initial laser radar detection frequency to detect the motion state of the own vehicle and the motion state of the target vehicle to calculate the event abnormality index;

[0010] Step S3, receiving the reflected signal characteristics of the detection target to determine the type of small vehicle target, and adjusting the laser radar detection frequency based on the event abnormality index and the type of small vehicle target;

[0011] Step S4, calculating the collision time by the laser radar detection frequency and the distance between the target vehicle and the own vehicle, and setting the alarm threshold according to the collision time.

[0012] In a preferred embodiment, in step S1, the exploration time is used to search for the lidar detection frequency under the minimum signal-to-noise ratio. The signal-to-noise ratio is a measure of the ratio between the strength of the signal and the strength of the background noise. In the lidar system, the signal-to-noise ratio is used to evaluate the ability to detect targets. A higher signal-to-noise ratio can improve the recognition rate and accuracy of the target.

[0013] In a preferred embodiment, in step S2, the motion state of the own vehicle includes the current position, driving speed and driving direction of the own vehicle, and the motion state of the target vehicle includes the current position, driving speed and driving direction of the target vehicle.

[0014] In a preferred embodiment, in step S2, the relative position between the current position of the own vehicle and the current position of the target vehicle is calculated, and the event abnormality index is calculated using a reinforcement learning method in combination with the driving speed and driving direction of the own vehicle and the driving speed and driving direction of the target vehicle.

[0015] In a preferred embodiment, in step S2, the specific steps of calculating the event abnormality index using the reinforcement learning method are as follows:

[0016] Environment settings: set vehicle status, set vehicle actions, and set behavior rewards based on vehicle status and vehicle actions;

[0017] Reinforcement learning: Initialize the event anomaly index, set up a reward mechanism, update the event anomaly index according to the reward mechanism, formulate training rules, set the time step, and update the time anomaly coefficient at each time step.

[0018] In a preferred embodiment, in step S3, when receiving the detection target reflection signal feature to determine the small vehicle target type, the small vehicle target type is determined by comparing the reflection signals and shapes of different types of small vehicles in the database, and different characteristic coefficients are calculated according to the small vehicle target type. If the target vehicle is a non-motor vehicle, the characteristic coefficient is set to 0. If the target vehicle is a motor vehicle, the characteristic coefficient calculation steps are as follows:

[0019] Select a detection time to record the target vehicle size and target vehicle speed of all small vehicles and merge them into a vehicle size data set and a vehicle speed data set respectively, de-unitize the data in the vehicle size data set and the vehicle speed data set and sort them numerically, set the classification interval after sorting in ascending order, set the split ratio to divide the data set into multiple small data sets, obtain the numerical interval of the divided small data sets and calculate the average value, select the median in the average value, and calculate the ratio of the average value to the median of each small data set as the vehicle size coefficient or the vehicle speed coefficient;

[0020] After detecting the target vehicle, the numerical range of the small data set corresponding to the target vehicle size and the target vehicle speed is determined, and the average value is calculated using the vehicle size coefficient and vehicle speed coefficient of the corresponding small data set as the characteristic coefficient of the target vehicle.

[0021] In a preferred embodiment, in step S3, the characteristic coefficient of the target vehicle and the time anomaly coefficient obtained at each time step are used to calculate the logistic regression coefficient as the adjustment ratio of the lidar detection frequency, and the ratio of the previous lidar detection frequency to the adjustment ratio is used as the adjusted lidar detection frequency at each time step.

[0022] In a preferred embodiment, in step S4, the system calculates the collision time and dynamically sets the safety distance according to the laser radar detection range. The specific steps are as follows:

[0023] Each time the lidar detection frequency is updated, the collision speed is calculated by performing a same-direction difference and opposite sum operation based on the own vehicle speed and the target vehicle speed, and then compared with the previously recorded collision speed. The larger speed is selected for calculation, and the ratio of the distance between the target vehicle and the own vehicle to the screened collision speed is recorded as the collision time, and compared with the previously recorded collision time. The smaller one is selected as the collision assessment time, and the safety range ratio is set to calculate the safety distance. The product of the collision assessment time multiplied by the safety range ratio and the collision speed is used as the safety range.

[0024] In a preferred embodiment, in step S4, the motion state of the target vehicle and the motion state of the own vehicle are detected at each time step through the adjusted laser radar detection frequency to dynamically update the safety range. When the distance between the own vehicle and the target vehicle reaches the safety range, the alarm sounds.

[0025] Technical effects and advantages of the automatic classification control method of traffic signal data based on target recognition of the present invention:

[0026] The present invention sets an exploration time mechanism. When an emergency occurs, multiple laser radar detection frequencies are set within the exploration time for screening, and the target vehicle's motion state is detected by using the screened laser radar detection frequencies, thereby improving the accuracy of vehicle data detection, improving the data basis for subsequent abnormal analysis of the target vehicle, and comprehensively determining the abnormal degree of the event by combining the vehicle's own speed and the target vehicle's motion state. The system transmits pulses and receives reflected signals to determine the type of small vehicle target, and determines the risk of collision by analyzing the characteristics of different vehicles through the target type, thereby improving the adjustment accuracy of the laser radar detection frequency and the alarm threshold. The laser radar detection frequency is adjusted in combination with the abnormal degree of the event and the small vehicle target type, and finally, the touch time is dynamically updated according to the laser radar detection frequency, and the alarm threshold is set by calculating the safe distance according to the touch time. When the vehicle distance reaches the alarm threshold, an alarm is triggered to ensure the driver's safety. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 The present invention is a schematic diagram of an automatic classification control method for traffic signal data based on target recognition. DETAILED DESCRIPTION

[0028] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0029] The present invention sets an exploration time mechanism. When an emergency occurs, multiple laser radar detection frequencies are set within the exploration time for screening. The target vehicle's motion state is detected using the screened laser radar detection frequencies. The abnormality of the event is determined by combining the vehicle's own speed and the target vehicle's motion state. The system transmits pulses and receives reflected signals to determine the type of small vehicle target. The laser radar detection frequency is adjusted based on the abnormality of the event and the type of small vehicle target. Finally, the touch time is dynamically updated according to the laser radar detection frequency. The safe distance is calculated according to the touch time to set the alarm threshold. When the distance between the vehicles reaches the alarm threshold, an alarm is triggered to ensure the driver's safety.

[0030] Embodiment, a method for automatically classifying and controlling traffic signal data based on target recognition, such as Figure 1 As shown, the following steps are included:

[0031] Step S1, setting an exploration time, setting multiple laser radar detection frequencies within the exploration time to detect the signal-to-noise ratio respectively, and selecting the maximum signal-to-noise ratio as the initial laser radar detection frequency;

[0032] Step S2, using the initial laser radar detection frequency to detect the motion state of the own vehicle and the motion state of the target vehicle to calculate the event abnormality index;

[0033] Step S3, receiving the reflected signal characteristics of the detection target to determine the type of small vehicle target, and adjusting the laser radar detection frequency based on the event abnormality index and the type of small vehicle target;

[0034] Step S4, calculating the collision time by the laser radar detection frequency and the distance between the target vehicle and the own vehicle, and setting the alarm threshold according to the collision time.

[0035] The specific implementation is as follows:

[0036] In step S1, the exploration time is used to search for the lidar detection frequency under the minimum signal-to-noise ratio. The lidar detection frequency is the frequency at which the lidar transmits and receives laser pulses. The greater the frequency of transmitting and receiving laser pulses, the higher the target detection details and resolution, but also increases the energy consumption.

[0037] The signal-to-noise ratio is a measure of the ratio between the strength of the signal and the strength of the background noise. In radar and lidar systems, the signal-to-noise ratio is used to evaluate the ability to detect targets. A higher signal-to-noise ratio can improve the recognition rate and accuracy of the target.

[0038] The exploration time is set using the percentile method. When the laser radar detects the target vehicle, the speed and direction of the target vehicle are obtained. The system collects the speed and direction of its own vehicle and calculates the driving time when its own vehicle and the target vehicle travel to the shortest distance. The exploration time is set according to the N% percentile ratio of the driving time. For example, if the driving time is 5 seconds and 20% is taken as the exploration time, the exploration time is 1 second.

[0039] The radar echo data for a period of time is recorded before the laser radar detects the target vehicle, and the average noise value is calculated as the standard noise. During the exploration time, the system randomly sets multiple laser radar detection frequencies to emit laser pulses to the target vehicle and receive the laser pulse intensity reflected by the target vehicle. The ratio of the laser pulse intensity of different laser radar detection frequencies to the standard noise is calculated to obtain the signal-to-noise ratio, and the calculated signal-to-noise ratio is compared. The laser radar detection frequency with the maximum signal-to-noise ratio is selected as the initial laser radar detection frequency.

[0040] It should be noted that when randomly setting multiple lidar detection frequencies, the switching time can be set according to the actual situation, that is, a lidar detection frequency is randomly generated each time the switching time passes during the exploration time. For example, if the exploration time is 1 second and the switching time is 0.1 second, a lidar detection frequency is generated every 0.1 second during the exploration time, and the average value of the laser pulse intensity received within 0.1 second is calculated as the laser pulse intensity corresponding to the lidar detection frequency.

[0041] In step S2, the motion state of the own vehicle includes the current position, driving speed and driving direction of the own vehicle, and the motion state of the target vehicle includes the current position, driving speed and driving direction of the target vehicle. The relative positions of the own vehicle and the target vehicle are calculated based on the current position of the own vehicle and the current position of the target vehicle.

[0042] The current position of the own vehicle and the current position of the target vehicle can be expressed by coordinates. The system uses the Beidou system to determine the coordinates of the own vehicle and uses the laser radar to detect the coordinates of the target vehicle. When calculating the relative position between them, the square root of the coordinate difference obtained by subtracting the horizontal and vertical coordinates of the two coordinates is obtained. The event abnormality index is calculated using the calculated relative position between each other, the speed and direction of the own vehicle, and the speed and direction of the target vehicle using the reinforcement learning method. The specific steps are as follows:

[0043] Environment setting: Set the state, record the relative position of the own vehicle and the target vehicle, as well as their respective driving speeds and directions; set the actions, including acceleration, deceleration, left turn, right turn, and maintaining the status quo; set the reward mechanism, give negative rewards when touching, give positive rewards based on time when there is no collision, and then give behavioral rewards for the set actions respectively.

[0044] Reinforcement Learning:

[0045] Step 1: Initialization. Set the initial event anomaly index according to the relative position of the own vehicle and the target vehicle. The initial event anomaly index can be set according to the distance between the own vehicle and the target vehicle. For example, when the distance between the own vehicle and the target vehicle is 60 meters, set the initial event anomaly index to 60.

[0046] Step 2: Set up a reward mechanism, which is an update mechanism for the event anomaly index. The event anomaly index is limited to 100. When it is touched, a negative reward is given to deduct 100 from the event anomaly index. Set the judgment frequency and make judgments based on the judgment frequency. Each time the judgment is made, if the vehicle does not collide, a positive reward is given to increase the event anomaly index by 1. A positive reward is given when the vehicle is gradually moving away from the target to increase the event anomaly index by 10. A negative reward is given when the vehicle is gradually approaching the target to deduct 10 from the event anomaly index. Set up behavioral rewards. When the vehicle is moving in the same direction as the target vehicle, the event anomaly index is increased by 5, and if it is moving in the opposite direction to the target vehicle, the event anomaly index is deducted by 5. For example, when the acceleration or deceleration direction or the steering direction of the vehicle is consistent with the target vehicle's driving speed direction, it is in the same direction. If there is an angle between the acceleration or deceleration direction or the steering direction of the vehicle and the target vehicle's driving speed direction, the positive angle direction of the angle less than a right angle among the east, south, west and north is taken as the driving direction of the vehicle and compared with the target vehicle's driving speed direction; otherwise, it is in the opposite direction. In addition, when the own vehicle is moving at a right angle to the target vehicle, the event abnormality index does not change.

[0047] Step 3: Formulate training rules to set time steps with judgment frequency, observe the status of the own vehicle and the target vehicle at each time step, randomly perform an action to update the position and status of the own vehicle, receive rewards and record the updated event anomaly index.

[0048] It should be noted that the Beidou system is a global satellite navigation system that can be used to determine the vehicle coordinates. The reinforcement learning method belongs to the category of machine learning. It optimizes the decision-making strategy through interactive learning with the environment. The reinforcement learning method can be used to set and update the event anomaly index, which is convenient for subsequent adjustment of the lidar detection frequency and alarm threshold. The time step is a basic time unit for system updates.

[0049] In step S3, the target reflection signal characteristics are detected to determine the type of small vehicle target. The type of small vehicle target is determined by comparing the reflection intensity comparison table and shape comparison library of reflection signals of different types of small vehicles in the open source database. The lidar detection frequency is dynamically adjusted based on the event anomaly index and the type of small vehicle target.

[0050] By obtaining the event anomaly index at each time step and combining it with the type of small vehicle target, the laser radar detection frequency is adjusted using the hierarchical analysis method. The specific steps are as follows:

[0051] When a small vehicle appears within the detection range of the laser radar, the target vehicle is classified as a motor vehicle or a non-motor vehicle by emitting pulses and receiving the detection target reflection signal characteristics. If the target vehicle is a motor vehicle, it is further divided into the target vehicle size in the detection target reflection signal characteristics and the target vehicle driving speed setting range for division.

[0052] When setting the interval, a detection experiment can be carried out. A detection time is selected for the detection experiment. The target vehicle size and target vehicle speed of all small cars in the detection time are recorded and merged into a vehicle size data set and a vehicle speed data set respectively. The data in the vehicle size data set and the vehicle speed data set are de-unitized and numerically sorted. After sorting in order from small to large, the classification interval is set. The data set is segmented using the percentile method, and M% is set as the segmentation ratio. The data set is divided into multiple small data sets according to M%. Taking the vehicle size data set as an example, the minimum and maximum values ​​in the divided small data sets are used as comparison value intervals and the average value is calculated. The median in the average value is selected, and the ratio of the average value to the median of each small data set is calculated as the vehicle size coefficient. The vehicle speed coefficient can be obtained similarly for the vehicle speed data set.

[0053] When the laser radar detects a target vehicle, the detected target vehicle size and target vehicle speed are compared with the comparison value range, the vehicle size coefficient and vehicle speed coefficient of the range to which the target vehicle size and target vehicle speed belong are calculated, and the average value is calculated as the characteristic coefficient of the target vehicle. In addition, if the detection target reflection signal characteristics determine that the target vehicle is a non-motor vehicle, the characteristic coefficient of the target vehicle defaults to 0.

[0054] After obtaining the characteristic coefficient of the target vehicle, the event anomaly index obtained by detecting the target vehicle is processed, and the event anomaly index of the target vehicle is ratioed with the set maximum limit of the event anomaly index, and the calculation result is used as the anomaly coefficient. The higher the anomaly coefficient, the higher the event anomaly index of the target vehicle under the condition that the maximum limit of the event anomaly index does not change, and the more likely it is that the own vehicle and the target vehicle will collide. It is necessary to increase the frequency of lidar detection and reduce the detection interval.

[0055] The abnormal coefficient and characteristic coefficient of the comprehensive target vehicle are used to calculate the logistic regression coefficient using the logistic regression method to adjust the laser radar detection frequency. The calculation formula is: P = 1 / 1 + e -z , where P is the logistic regression coefficient, z is the sum of the characteristic coefficient and the abnormal coefficient of the target vehicle, and e is the natural base. The calculated logistic regression coefficient is used as the adjustment ratio of the laser radar detection frequency, then: T new =T old / P,T newis the adjusted lidar detection frequency, T old is the lidar detection frequency before adjustment.

[0056] It should be noted that the larger the size of the target vehicle or the faster the speed, the greater the harm caused by the collision, and the laser radar detection frequency needs to be increased, that is, the detection interval time needs to be reduced. The characteristic coefficient is the average of the vehicle size coefficient and vehicle speed coefficient of the target vehicle. When the characteristic coefficient of the target vehicle is larger, the vehicle size coefficient or vehicle speed coefficient of the target vehicle is larger, that is, the larger the vehicle size and vehicle speed of the target vehicle are, the more it is necessary to reduce the detection interval time. The segmentation ratio can be set according to actual requirements. For example, if the segmentation ratio is set to 20%, the data set will be divided into five small data sets, and five comparison value intervals will appear, etc., which will not be repeated here.

[0057] In step S4, the system calculates the collision time and sets a safety distance according to the laser radar detection range. When the target vehicle enters the set safety distance, the alarm is triggered. The safety distance can be dynamically set by calculating the collision time based on the laser radar detection frequency and the distance between the target vehicle and the vehicle itself. Each time the laser radar detection frequency is updated, the speed of the own vehicle and the target vehicle are subtracted in the same direction and summed in opposite directions. For example, when the own vehicle and the target vehicle are traveling in the same direction, the speed of the target vehicle is subtracted from the speed of the own vehicle. When the own vehicle and the target vehicle are traveling in opposite directions, the speed of the target vehicle and the speed of the own vehicle are summed as the collision speed.

[0058] It should be explained that since there are many driving directions between vehicles, the driving speeds of the own vehicle and the target vehicle are calculated by taking the driving speed components of the driving directions on the parallel line of the distance between the two vehicles. For example, when the driving directions of the two vehicles and the angle between the parallel line of the distance between the two vehicles are both 30 degrees, the vehicle driving speed is the speed component after taking the cosine of 30 degrees of the original driving speed on the average line of the distance between the two vehicles.

[0059] The collision speed at each lidar detection frequency is recorded and compared with the previously recorded collision speed. The larger speed is selected for calculation. The ratio of the distance between the target vehicle and the own vehicle to the screened collision speed is recorded as the collision time. When the lidar detection frequency is updated at each time step, the collision time will also be dynamically updated. The minimum value of the collision time recorded in the past is selected as the collision assessment time, and the safety range ratio is set to calculate the safety distance. The product of the collision assessment time multiplied by the safety range ratio and the collision speed is used as the safety range. When the distance between the own vehicle and the target vehicle reaches the safety range, the alarm will sound.

[0060] It should be noted that the safety range ratio is used to control the size of the set safety range, which is set by the user of this system through the user interface and will not be analyzed here.

[0061] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product.

[0062] Those of ordinary skill in the art will appreciate that the modules and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application of the technical solution and the invention constraints. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0063] In addition, each functional module in each embodiment of the present application may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module.

[0064] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0065] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for automatic classification and control of traffic signal data based on target recognition, characterized in that: The following steps are included: Step S1, setting an exploration time, setting multiple laser radar detection frequencies within the exploration time to detect the signal-to-noise ratio respectively, and selecting the maximum signal-to-noise ratio as the initial laser radar detection frequency; Step S2, using the initial laser radar detection frequency to detect the motion state of the own vehicle and the motion state of the target vehicle to calculate the event abnormality index; Step S3, receiving the reflected signal characteristics of the detection target to determine the type of small vehicle target, and adjusting the laser radar detection frequency based on the event abnormality index and the type of small vehicle target; Step S4, calculating the collision time by the laser radar detection frequency and the distance between the target vehicle and the own vehicle, and setting the alarm threshold according to the collision time.

2. The method for automatic classification and control of traffic signal data based on target recognition according to claim 1, characterized in that: In step S1, the exploration time is used to search for the lidar detection frequency under the minimum signal-to-noise ratio. The signal-to-noise ratio is a measure of the ratio between the strength of the signal and the strength of the background noise. In the lidar system, the signal-to-noise ratio is used to evaluate the ability to detect targets. A higher signal-to-noise ratio can improve the recognition rate and accuracy of the target.

3. The method for automatic classification and control of traffic signal data based on target recognition according to claim 1, characterized in that: In step S2, the motion state of the own vehicle includes the current position, driving speed and driving direction of the own vehicle, and the motion state of the target vehicle includes the current position, driving speed and driving direction of the target vehicle.

4. The method for automatic classification and control of traffic signal data based on target recognition according to claim 3 is characterized in that: In step S2, the relative position between the current position of the own vehicle and the current position of the target vehicle is calculated, and the event abnormality index is calculated using the reinforcement learning method in combination with the driving speed and driving direction of the own vehicle and the driving speed and driving direction of the target vehicle.

5. The method for automatic classification and control of traffic signal data based on target recognition according to claim 4, characterized in that: In step S2, the specific steps of calculating the event abnormality index using the reinforcement learning method are as follows: Environment settings: set vehicle status, set vehicle actions, and set behavior rewards based on vehicle status and vehicle actions; Reinforcement learning: Initialize the event anomaly index, set up a reward mechanism, update the event anomaly index according to the reward mechanism, formulate training rules, set the time step, and update the time anomaly coefficient at each time step.

6. The method for automatic classification and control of traffic signal data based on target recognition according to claim 1, characterized in that: In step S3, when receiving the detection target reflection signal feature to determine the small vehicle target type, the small vehicle target type is determined by comparing the reflection signals and shapes of different types of small vehicles in the database, and different characteristic coefficients are calculated according to the small vehicle target type. If the target vehicle is a non-motor vehicle, the characteristic coefficient is set to 0; if the target vehicle is a motor vehicle, The steps for calculating the characteristic coefficient are as follows: Select a detection time to record the target vehicle size and target vehicle speed of all small vehicles and merge them into a vehicle size data set and a vehicle speed data set respectively, de-unitize the data in the vehicle size data set and the vehicle speed data set and sort them numerically, set the classification interval after sorting in ascending order, set the split ratio to divide the data set into multiple small data sets, obtain the numerical interval of the divided small data sets and calculate the average value, select the median in the average value, and calculate the ratio of the average value to the median of each small data set as the vehicle size coefficient or the vehicle speed coefficient; After detecting the target vehicle, the numerical range of the small data set corresponding to the target vehicle size and the target vehicle speed is determined, and the average value is calculated using the vehicle size coefficient and vehicle speed coefficient of the corresponding small data set as the characteristic coefficient of the target vehicle.

7. The method for automatic classification and control of traffic signal data based on target recognition according to claim 6, characterized in that: In step S3, the characteristic coefficient of the target vehicle and the time anomaly coefficient obtained at each time step are used to calculate the logistic regression coefficient as the adjustment ratio of the lidar detection frequency, and the ratio of the previous lidar detection frequency to the adjustment ratio is used as the adjusted lidar detection frequency at each time step.

8. The method for automatic classification and control of traffic signal data based on target recognition according to claim 1, characterized in that: In step S4, the system calculates the collision time and dynamically sets the safety distance according to the laser radar detection range. The specific steps are as follows: Each time the lidar detection frequency is updated, the collision speed is calculated by performing a same-direction difference and opposite sum operation based on the own vehicle speed and the target vehicle speed, and then compared with the previously recorded collision speed. The larger speed is selected for calculation, and the ratio of the distance between the target vehicle and the own vehicle to the screened collision speed is recorded as the collision time, and compared with the previously recorded collision time. The smaller one is selected as the collision assessment time, and the safety range ratio is set to calculate the safety distance. The product of the collision assessment time multiplied by the safety range ratio and the collision speed is used as the safety range.

9. The method for automatic classification and control of traffic signal data based on target recognition according to claim 8, characterized in that: In step S4, the motion state of the target vehicle and the motion state of the own vehicle are detected at each time step through the adjusted laser radar detection frequency to dynamically update the safety range. When the distance between the own vehicle and the target vehicle reaches the safety range, the alarm sounds.

Images