Environmental perception system based on intelligent driving

Abstract

The present invention discloses an environment perception system based on intelligent driving, which relates to the field of intelligent driving technology. Environmental data is acquired in real time through sensors such as radar and laser radar, and SLAM technology is used to compare with a basic environmental model to identify false signal areas. The system further evaluates the degree of signal interference through signal strength and path complexity analysis, and divides it into high, medium and low interference levels, and adopts different processing strategies. For low interference levels, sensor data is directly used to update environmental models and dynamic object detection; for high interference levels, the system reduces the driving speed and corrects the path; for general interference levels, the system optimizes the path adjustment range according to the interference prediction results within a fixed time period, so as to ensure the safety, accuracy and stability of the automatic driving system in complex environments.

Description

Technical Field

[0001] The present invention relates to the field of intelligent driving technology, and in particular to an environment perception system based on intelligent driving. Background Art

[0002] Environmental perception based on intelligent driving refers to the use of various sensors (such as cameras, radars, lidars, etc.) to collect real-time environmental data around the vehicle, and process and analyze these data through advanced artificial intelligence algorithms, which can identify dynamic and static objects on the road, such as pedestrians, other vehicles, road signs, traffic lights, and potential obstacles. Through this information processing, the vehicle can form a global understanding of the surrounding environment and provide an accurate environmental model for the autonomous driving system. However, in some complex environments (such as high-rise reflective areas in cities, tunnels, or under bridges), radar and lidar signals may be reflected or interfered, causing the system to receive distorted data, which may cause the algorithm to mistakenly believe that the object does not exist or misjudge the location and movement trajectory of the object, thereby causing wrong decisions. At the same time, due to the sensor's misjudgment of dynamic or static objects in the environment, the vehicle may not be able to correctly plan the driving path. For example, in tunnels or urban areas with dense high-rise buildings, radar reflections cause distortion of the environmental model, and the vehicle may not correctly identify road boundaries or navigation signs, thereby deviating from the normal route. This not only increases the risk of traffic accidents, but may also seriously affect the safety of passengers. Summary of the invention

[0003] The purpose of the present invention is to provide an environment perception system based on intelligent driving to solve the deficiencies in the background technology.

[0004] In order to achieve the above-mentioned object, the present invention provides the following technical solutions: an environment perception system based on intelligent driving, comprising a signal acquisition module, a comparison and analysis module, a signal strength analysis module, a signal path complexity evaluation module, an interference level classification module, an interference processing module and a path optimization module;

[0005] Signal acquisition module: The autonomous driving system uses several radar and lidar sensors installed on the vehicle to obtain reflected signals from different directions during the vehicle's driving process;

[0006] Comparative analysis module: The basic environment model around the vehicle is pre-built through SLAM technology, and the reflection signal data collected in real time is compared and analyzed with the static environment information in the basic environment model, and the false information area is determined based on the analysis results;

[0007] Signal strength analysis module: For false information areas, the signal strength data of radar and lidar are collected according to the changes in signal reflection distance, and the changes in the proportion of abnormally high-intensity signals in the echo signals are analyzed to evaluate the stability of the echo signal strength;

[0008] Signal path complexity assessment module: By comparing the arrival time of the signal, the signal delay state of several radar and lidar sensors detecting the same object is analyzed to assess the complexity of the signal path;

[0009] Interference level classification module: According to the stability of echo signal strength and the complexity of signal path, the interference degree of signal anomaly on environmental perception data is evaluated, and according to the evaluation results, it is divided into high interference level, general interference level and low interference level;

[0010] Interference processing module: For low interference levels, radar and lidar data are directly used to update the basic environment model and dynamic object detection in real time; for high interference levels, driving is reduced based on low interference level data and static environment model, and the driving path is gradually corrected;

[0011] Path optimization module: For general interference levels, further predictions are made on the degree of interference of signal anomalies on environmental perception data within a fixed time period, and the path adjustment range is optimized based on the prediction results.

[0012] Preferably, in the comparison and analysis module, the reflection signal data collected in real time is compared and analyzed with the static environment information in the basic environment model, specifically:

[0013] Extract the location information of static objects from the basic environment model built by SLAM, and compare it with the radar and lidar reflection signals collected in real time;

[0014] The comparison is based on the position, size, and shape characteristics of objects in three-dimensional space. If the real-time data matches the position and size of the object in the environment model, the object is considered to be a known static object and does not require further processing;

[0015] For objects in real-time data that fail to match the environmental model, the system will further analyze the characteristics of their reflection signals to determine whether they are dynamic objects. If the object appears to be static but does not match the position in the basic environmental model, it will be marked as a false information area.

[0016] Preferably, in the signal strength analysis module, a time series analysis is performed on the signal strength in the T time period in the false information area, and an abnormal signal ratio fluctuation index is generated according to the change in the proportion of abnormally high-intensity signals in the echo signal. The method for obtaining the abnormal signal ratio fluctuation index is:

[0017] In the T time period, the abnormal signal ratio at time point t is calculated, that is, the ratio of abnormal high-intensity signals to all signals is marked as , and establish the corresponding time series set, calculate the similarity between each time point in the time series and its lagged value , the expression is: ;W is the sliding window size, is the lag time interval, is the abnormal signal ratio at time point i, For time point The abnormal signal ratio, is the mean value of the abnormal signal ratio at time point t in the sliding window W, and the calculation formula is: ; Sliding window from time Start from the beginning and gradually move forward until the entire time period T is covered. Calculate the fluctuation of the abnormal signal ratio in the entire time period T, that is, calculate the abnormal signal ratio fluctuation index, which is expressed as: ; Where FD is the abnormal signal proportion fluctuation index.

[0018] Preferably, in the signal path complexity evaluation module, the signal delay states of the same object detected by several radar and lidar sensors are analyzed, and the delay deviation anomaly index is generated after comparing the delay differences of the signals of several sensors. The method for obtaining the delay deviation anomaly index is:

[0019] The signal delay data reaching the same object is obtained from n sensors, and the signal delay difference data of different sensors are sorted into a matrix form. The matrix X is expressed as: ; It represents the signal delay difference detected by the nth sensor at the mth time point. After the signal delay difference in the matrix is ​​standardized, the linear relationship between the delay data of different sensors is calculated. The expression is: ; In the formula, T is the matrix transpose, Z is the standardized delay data matrix, C is an n×n matrix, which expresses the covariance between the signal delay difference data. The covariance matrix C is subjected to eigenvalue decomposition to obtain the eigenvalue and the corresponding eigenvector ; is the eigenvalue of the nth principal component, is the eigenvector corresponding to the nth principal component; ; Project the original standardized data Z onto the selected principal component to obtain the principal component score S, which is expressed as: ; In the formula, is a matrix composed of the eigenvectors of the first k principal components. According to the principal component scores, the delay deviation anomaly index is calculated. The expression is: ; In the formula, is the score of the jth principal component, is the delay deviation anomaly index.

[0020] Preferably, in the interference level classification module, the abnormal signal proportion fluctuation index and the delay deviation abnormal index are converted into a first eigenvector, and the first eigenvector is used as the input of the machine learning model. The machine learning model uses each group of first eigenvectors to predict the interference degree value label of the environmental perception data as the prediction target, and takes minimizing the sum of prediction errors of the interference degree value labels of all environmental perception data as the training target. The machine learning model is trained until the sum of prediction errors converges, and the model training is stopped. The interference degree value of the environmental perception data is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0021] Preferably, the interference degree value of the acquired environmental perception data is compared with a gradient standard threshold, the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and the interference degree value of the environmental perception data is compared with the first standard threshold and the second standard threshold respectively;

[0022] If the interference level of the environmental perception data is greater than the second standard threshold, it means that the interference level of the signal anomaly on the environmental perception data is high, and a high anomaly signal is generated, which is classified as a high interference level;

[0023] If the interference level of the environmental perception data is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it means that the interference level of the signal anomaly on the environmental perception data is general, and a general anomaly signal is generated, which is classified as a general interference level;

[0024] If the interference level value of the environmental perception data is less than the first standard threshold, it means that the interference level of the signal anomaly on the environmental perception data is low. At this time, a low anomaly signal is generated and classified as a low interference level.

[0025] Preferably, in the path optimization module, for the general interference level, the interference degree of the signal anomaly on the environmental perception data within a fixed time period is further predicted;

[0026] For the general interference level, that is, the interference degree value of the environmental perception data generated within a fixed time period is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the interference degree values ​​of the environmental perception data generated in the subsequent fixed time period that are greater than or equal to the first standard threshold and less than or equal to the second standard threshold are collected, and the corresponding data set is established, the mean and standard deviation of the data set are calculated, and after analysis and prediction, the path adjustment range is optimized according to the prediction results.

[0027] Preferably, if the mean value of the interference degree of the environmental perception data in the data set is greater than or equal to the reference threshold of the mean value of the interference degree of the environmental perception data, and the standard deviation of the interference degree of the environmental perception data is less than the reference threshold of the standard deviation of the interference degree of the environmental perception data, it means that the interference degree of the environmental perception data is high but the fluctuation is small, and the interference is stable, and the path is adjusted to a moderate degree to avoid the continuous influence of the high interference area;

[0028] If the mean value of the interference degree of the environmental perception data is greater than or equal to the reference threshold of the mean value of the interference degree of the environmental perception data, and the standard deviation of the interference degree of the environmental perception data is greater than or equal to the reference threshold of the standard deviation of the interference degree of the environmental perception data, it means that the interference degree of the environmental perception data is high and fluctuates greatly, and the interference is strong and unstable. The path is adjusted significantly to avoid the area to prevent unpredictable signal interference from causing adverse effects on path planning;

[0029] If the mean value of the interference degree of the environmental perception data is less than the reference threshold value of the mean value of the interference degree of the environmental perception data, and the standard deviation of the interference degree of the environmental perception data is greater than or equal to the reference threshold value of the standard deviation of the interference degree of the environmental perception data, it means that the interference degree of the environmental perception data is low but fluctuates greatly, and the interference is sometimes strong and sometimes weak. A small range of path adjustment is performed to reduce the risk of interference;

[0030] If the mean value of the interference degree of the environmental perception data is less than the reference threshold of the mean value of the interference degree of the environmental perception data, and the standard deviation of the interference degree of the environmental perception data is less than the reference threshold of the standard deviation of the interference degree of the environmental perception data, the interference degree of the environmental perception data is low and the fluctuation is small, the interference is stable and weak, and there is no need to adjust the path. The vehicle continues to travel along the original path.

[0031] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0032] 1. The present invention solves the problem of misjudgment caused by sensor signal interference in complex environments through the collaborative work of multiple modules. Through modules such as signal acquisition, SLAM comparative analysis, signal strength and path complexity evaluation, the system can identify and classify different levels of interference. For low interference levels, sensor data is directly used for environmental model update and dynamic object detection to ensure the stability and accuracy of vehicle driving; for high interference levels, the system reduces vehicle speed and corrects the path to prevent wrong decisions caused by signal interference, greatly improving the safety of the autonomous driving system in complex environments.

[0033] 2. The present invention predicts and optimizes the path for general interference levels, collects and analyzes the interference level values ​​of environmental perception data, and dynamically adjusts the vehicle's driving path in combination with its mean and standard deviation. For situations where the interference level is high but the fluctuation is small, the system makes moderate path adjustments, and for high interference situations with large fluctuations, it makes large-scale path optimization to avoid unstable interference areas. The system can flexibly adjust the vehicle's driving strategy according to changes in environmental interference, ensuring the safety, stability and driving efficiency of autonomous driving. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0035] Figure 1 The figure is a flow chart of the method of the present invention.

[0036] Figure 2 It is a system module diagram of the present invention. DETAILED DESCRIPTION

[0037] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are 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.

[0038] For examples, see Figure 1 and Figure 2 As shown, the environment perception system based on intelligent driving described in this embodiment includes a signal acquisition module, a comparison and analysis module, a signal strength analysis module, a signal path complexity evaluation module, an interference level classification module, an interference processing module and a path optimization module;

[0039] Signal acquisition module: The autonomous driving system uses several radar and lidar sensors installed on the vehicle to obtain reflected signals from different directions during the vehicle's driving process;

[0040] Comparative analysis module: The basic environment model around the vehicle is pre-built through SLAM technology, and the reflection signal data collected in real time is compared and analyzed with the static environment information in the basic environment model, and the false information area is determined based on the analysis results;

[0041] Signal strength analysis module: For false information areas, the signal strength data of radar and lidar are collected according to the changes in signal reflection distance, and the changes in the proportion of abnormally high-intensity signals in the echo signals are analyzed to evaluate the stability of the echo signal strength;

[0042] Signal path complexity assessment module: By comparing the arrival time of the signal, the signal delay state of several radar and lidar sensors detecting the same object is analyzed to assess the complexity of the signal path;

[0043] Interference level classification module: According to the stability of echo signal strength and the complexity of signal path, the interference degree of signal anomaly on environmental perception data is evaluated, and according to the evaluation results, it is divided into high interference level, general interference level and low interference level;

[0044] Interference processing module: For low interference levels, radar and lidar data are directly used to update the basic environment model and dynamic object detection in real time; for high interference levels, driving is reduced based on low interference level data and static environment model, and the driving path is gradually corrected;

[0045] Path optimization module: For general interference levels, further predictions are made on the degree of interference of signal anomalies on environmental perception data within a fixed time period, and the path adjustment range is optimized based on the prediction results.

[0046] In the signal acquisition module, multiple radar and lidar sensors are installed in the front, rear, left, right and top of the autonomous vehicle. The layout of these sensors ensures that the vehicle's 360-degree field of view is covered, blind spots are minimized, and reflected signals from different directions are captured. The installation angle of each sensor is carefully designed to reduce reflection blind spots while covering potential reflection areas in complex urban environments (such as high-rise buildings, tunnels, etc.). After installation, all sensors are calibrated to ensure that they work in the same coordinate system. Through precise time synchronization, all radars and lidars are able to collect data at the same time point, ensuring that the subsequent data fusion process is consistent and accurate.

[0047] Each radar and lidar sensor periodically sends out signal waves (radar sends out electromagnetic waves, and lidar sends out laser beams). These signals will encounter obstacles, static objects, or dynamic objects (such as vehicles and pedestrians) in the environment and generate reflections. The sensor receives the reflected signal and records relevant data, such as the distance, relative speed, azimuth, reflection intensity, etc. of the object. Each sensor generates a point cloud or distance data set based on the data it collects. This data contains information such as the three-dimensional position, outline, and size of the surrounding objects. When sensors in different directions detect the same object, a direction mark will be added to each point cloud or data block for subsequent data fusion and analysis.

[0048] In the comparative analysis module, the system first collects initial data in the environment through multiple sensors such as radar, lidar, and cameras. After denoising and preprocessing, the initial data is used as the input of the SLAM algorithm to build a three-dimensional static environment model around the vehicle. The system uses SLAM algorithms (such as lidar SLAM or visual SLAM) to build a three-dimensional map in an unknown environment and determine the position of the vehicle in the map in the process. The core of SLAM technology is to continuously update the environment model through sensor data and simultaneously update the position of the vehicle in this environment. The model includes static environment information such as roads, buildings, and bridges as the framework of the basic environment model.

[0049] After the SLAM process is completed, the system stores the information of static objects on the vehicle's driving path in the basic environment model. The model contains accurate object position, size, boundary information, etc., representing the static environment structure around the vehicle. While the vehicle is driving, multiple radar and lidar sensors installed continuously transmit signals and receive reflected data to form point cloud information or distance data containing surrounding objects. The data of each sensor includes key information such as the position, reflection intensity, distance of the object, and marks the timestamp and sensor direction of the collected data. The data from different sensors are synchronized by timestamp, and all data points are aligned to the same coordinate system to ensure that the data can be directly compared with the basic environment model constructed by SLAM.

[0050] The system extracts the location information of static objects from the basic environment model constructed by SLAM and compares it with the radar and lidar reflection signals collected in real time;

[0051] The comparison is based on the position, size, and shape features of objects in three-dimensional space. If the real-time data matches the position and size of the object in the environment model, the object is considered to be a known static object and does not require further processing;

[0052] For objects in real-time data that fail to match the environmental model, the system will further analyze the characteristics of their reflection signals to determine whether they are dynamic objects (such as pedestrians or vehicles). If the object appears to be static but does not match the position in the basic environmental model, it will be marked as a false information area.

[0053] In the signal strength analysis module, for the false information area, the system extracts the signal strength from the echo signals of the radar and lidar and stores it together with the reflection distance. The intensity value of each reflected signal is usually related to the material and shape of the reflecting surface. Too high or too low intensity may indicate that the signal has been reflected or interfered. When collecting data, the system records the echo intensity at each moment and its relationship with distance to ensure that the intensity data matches the distance data. The system performs a time series analysis of the signal strength over a period of time in the false information area to observe the dynamic changes in signal strength. Stable signal strength should remain within a certain range over time, while false signals may show significant fluctuations or abnormally high intensity.

[0054] The signal strength in the T time period in the false information area is analyzed in time series, and the abnormal signal ratio fluctuation index is generated according to the change in the proportion of abnormal high-intensity signals in the echo signal to evaluate the stability of the echo signal strength. The abnormal signal ratio fluctuation index is obtained as follows:

[0055] In the T time period, the abnormal signal ratio at time point t is calculated, that is, the ratio of abnormal high-intensity signals to all signals is marked as , and establish the corresponding time series set, calculate the similarity between each time point in the time series and its lagged value , the expression is: ;W is the sliding window size, is the lag time interval, usually one or more time steps, is the abnormal signal ratio at time point i, For time point The abnormal signal ratio, is the mean value of the abnormal signal ratio at time point t in the sliding window W, and the calculation formula is: The sliding window starts from time t=1 and moves forward step by step until it covers the entire time period T. The fluctuation of the abnormal signal ratio in the entire time period T is calculated, that is, the abnormal signal ratio fluctuation index is calculated. The expression is: ; Where FD is the abnormal signal proportion fluctuation index.

[0056] When the abnormal signal ratio fluctuation index is large, it means that the proportion of abnormal high-intensity signals in the echo signal fluctuates greatly. This means that the echo signal strength changes significantly at different time points, which may be affected by external interference or multiple reflections, resulting in increased signal instability. At this time, the signal quality of the environmental perception system in this area is poor, the probability of misjudging false signals is high, and the stability is low.

[0057] On the contrary, when the abnormal signal ratio fluctuation index is small, it means that the abnormal high-intensity signal ratio is relatively stable, the change amplitude of the echo signal strength is small, and it shows high stability. This usually indicates that there is less signal interference, and the sensor can more accurately perceive the reflection of objects in the environment. At this time, the signal strength is more consistent, and the system's perception accuracy and reliability of the environment are also improved accordingly.

[0058] Signal Path Complexity Assessment Module: This module analyzes the signal delay states of several radar and lidar sensors detecting the same object by comparing the arrival times of the signals, and assesses the complexity of the signal path.

[0059] Several radar and lidar sensors are installed on the autonomous vehicle, and each sensor detects the reflected signal of the same object from different directions. Each sensor records the time difference from the transmission signal to the reception of the reflected signal, that is, the signal flight time. By comparing the arrival time of the signal of the same object from different sensors, the delay difference of the signal traveling on different paths is calculated.

[0060] Ideally, the path of a signal from the source to the object and back to the receiver is a straight line. However, in complex environments (such as tall buildings, bridges, tunnels, etc.), the signal may be reflected multiple times, and the delay time will increase compared to the direct path. By comparing the delay differences of the signals from several sensors, the system can detect whether there is an abnormal delay. If the signal delay time of some sensors is significantly higher than that of other sensors, it may be because the signal has passed through a complex reflection path.

[0061] The signal delay status of the same object detected by several radar and lidar sensors is analyzed, and the delay deviation anomaly index is generated after comparing the delay differences of the signals of several sensors to evaluate the complexity of the signal path. The delay deviation anomaly index is obtained as follows:

[0062] The signal delay data reaching the same object is obtained from n sensors, and the signal delay difference data of different sensors are sorted into a matrix form. The matrix X is expressed as: ; It represents the signal delay difference detected by the nth sensor at the mth time point. After the signal delay difference in the matrix is ​​standardized, the linear relationship between the delay data of different sensors is calculated. The expression is: ; In the formula, T is the matrix transpose, Z is the standardized delay data matrix, C is an n×n matrix, which expresses the covariance between the signal delay difference data. The covariance matrix C is subjected to eigenvalue decomposition to obtain the eigenvalue and the corresponding eigenvector ; is the eigenvalue of the nth principal component, indicating the contribution of the principal component to the data variance. is the eigenvector corresponding to the nth principal component, indicating the direction of the principal component; ; Project the original standardized data Z onto the selected principal component to obtain the principal component score S, which is expressed as: ; In the formula, is a matrix composed of the eigenvectors of the first k principal components. According to the principal component scores, the delay deviation anomaly index is calculated. The expression is: ; In the formula, is the score of the jth principal component, is the delay deviation anomaly index.

[0063] The larger the delay deviation anomaly index, the greater the difference in signal delays detected by each sensor, indicating that there are more anomalies in the signal path. This usually reflects that the signal has passed through a complex propagation path, such as multiple reflections, obstructions, or detours, resulting in significant differences in the time it takes for each sensor to receive the signal. In this case, the complexity of the signal path is high, indicating that there are multiple interference sources in the environment, which may affect the accuracy and stability of the autonomous driving system.

[0064] The smaller the delay deviation anomaly index, the smaller the signal delay difference between the sensors, indicating that the signal path is relatively direct, the propagation process is simple, and it is not affected by multiple reflections or interference. In this case, the complexity of the signal path is low, and the signal time received by each sensor is relatively consistent, indicating that there is less environmental interference, the system has a higher accuracy in environmental perception, and the signal path is more stable.

[0065] Interference level classification module: According to the stability of the echo signal strength and the complexity of the signal path, the interference degree of the signal anomaly on the environmental perception data is evaluated, and according to the evaluation results, it is divided into high interference level, general interference level and low interference level.

[0066] The abnormal signal proportion fluctuation index and the delay deviation abnormal index are converted into the first eigenvector, and the first eigenvector is used as the input of the machine learning model. The machine learning model predicts the interference degree value label of the environmental perception data for each group of first eigenvectors as the prediction target, and minimizes the sum of prediction errors of the interference degree value labels of all environmental perception data as the training target. The machine learning model is trained until the sum of prediction errors converges and the model training is stopped. The interference degree value of the environmental perception data is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

[0067] The method for obtaining the interference degree value of the environmental perception data is as follows: from the first feature vector training data of the trained machine learning model, the corresponding function expression is obtained: ; In the formula, is the output function of the model, FD is the abnormal signal proportion fluctuation index, is the delay deviation anomaly index, is the interference degree value of the environmental perception data.

[0068] Compare the interference degree value of the acquired environmental perception data with a gradient standard threshold, where the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and compare the interference degree value of the environmental perception data with the first standard threshold and the second standard threshold respectively;

[0069] If the interference level of the environmental perception data is greater than the second standard threshold, it means that the interference level of the signal anomaly on the environmental perception data is high, and a high anomaly signal is generated, which is classified as a high interference level;

[0070] If the interference level of the environmental perception data is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it means that the interference level of the signal anomaly on the environmental perception data is general, and a general anomaly signal is generated, which is classified as a general interference level;

[0071] If the interference level value of the environmental perception data is less than the first standard threshold, it means that the interference level of the signal anomaly on the environmental perception data is low. At this time, a low anomaly signal is generated and classified as a low interference level.

[0072] Interference processing module: For low interference levels, the data from radar and lidar are directly used to update the basic environmental model and the detection of dynamic objects in real time; for high interference levels, the driving is reduced based on the data of low interference levels and the static environmental model, and the driving path is gradually corrected.

[0073] When the system detects that the interference level of radar and lidar signals is low, or the interference level value of environmental perception data is less than the first standard threshold, that is, the low interference level, it means that the signal interference in the environment is not significant, and the data collected by the sensor has a high reliability. The system directly uses the radar and lidar data collected in real time to update the basic environment model around the vehicle. At this time, the data accuracy is high and can accurately reflect the static objects in the environment (such as roads, buildings, traffic signs, etc.). Dynamic objects around the vehicle, such as pedestrians and vehicles, are captured in real time through radar and lidar sensors. The system tracks the movement trajectory of these dynamic objects through multi-frame data fusion and determines their speed and direction. Because it is in a low-interference environment, the system has high detection and tracking accuracy for dynamic objects, and decisions can be adjusted in real time. In a low-interference environment, the system can rely on sensor data to plan the vehicle's driving path to ensure safe driving. The vehicle's driving decision is mainly based on the detection results of dynamic objects and the update of the environmental model, including obstacle avoidance, lane change and other operations. When the vehicle is driving on an open or interference-free road, the radar and lidar detect the position and speed of surrounding vehicles and pedestrians, and the system directly uses these data to update the environmental model in real time and generate a driving path. Based on the movement trajectory of objects captured by the sensor, the system predicts the next path. If it detects that the vehicle ahead is slowing down, the autonomous driving system will plan to slow down or change lanes in advance to ensure driving safety.

[0074] When the system detects a high level of interference in radar and lidar signals, or the interference level of environmental perception data is greater than the second standard threshold, that is, a high interference level, it means that there are significant signal reflection, occlusion and other problems in the environment, and the sensor data is distorted or unreliable.

[0075] In high-interference areas, radar and lidar data may be severely interfered with (e.g., multiple reflections, environmental noise), and therefore cannot be directly used to update the environmental model. The system uses only reliable data from low-interference areas through multi-sensor data fusion. The system corrects the interfered area information based on data with low interference levels and historical static environmental models. The static environmental model contains known stable features such as roads and buildings, which can help compensate for areas where sensor data is inaccurate.

[0076] When the interference level is high, the system determines that the environment is not conducive to safe high-speed driving, and the vehicle will reduce its driving speed to avoid misjudgment or emergency situations. Slowing down can give the system more time to process data and adjust the path. Since the accuracy of dynamic object detection is reduced under high interference levels, the system will rely less on real-time detection data. By comparing with low-interference data and static environment models, the system makes path corrections. At this time, path planning relies more on existing environmental perception models. When gradually adjusting the path, the system will drive conservatively to avoid the vehicle entering areas that may be severely disturbed to ensure driving safety.

[0077] When a vehicle enters a complex urban area with tall buildings or a tunnel, the LiDAR may misjudge the vehicle due to multiple reflections or signal obstruction. The system filters out high-interference sensor data and relies only on previously collected reliable data and static environmental models to determine the location of buildings, road boundaries, etc. Multiple reflections from the LiDAR between tall buildings or in tunnels result in inaccurate environmental perception. At this point, the vehicle will slow down and try to rely on static models and GPS signals to avoid collisions and misoperations. Assuming that the radar and LiDAR are interfered with when the vehicle enters a tunnel and misjudge that there is an obstacle ahead, the system will gradually correct the vehicle's path based on historical data from low-interference periods to avoid sudden braking or misdirection.

[0078] Path optimization module: For general interference levels, further predictions are made on the degree of interference of signal anomalies on environmental perception data within a fixed time period, and the path adjustment range is optimized based on the prediction results.

[0079] For the general interference level, that is, the interference degree value of the environmental perception data generated within a fixed time period is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the interference degree values ​​of the environmental perception data generated in the subsequent fixed time period that are greater than or equal to the first standard threshold and less than or equal to the second standard threshold are collected, and the corresponding data set is established, the mean and standard deviation of the data set are calculated, and after analysis and prediction, the path adjustment range is optimized according to the prediction results.

[0080] If the mean value of the interference degree of the environmental perception data in the data set is greater than or equal to the reference threshold of the mean value of the interference degree of the environmental perception data, and the standard deviation of the interference degree of the environmental perception data is less than the reference threshold of the standard deviation of the interference degree of the environmental perception data, it means that the interference degree of the environmental perception data is high but the fluctuation is small, and the interference is stable. Since the interference is stable, it is recommended to make moderate adjustments to the path to avoid the continuous influence of the high interference area, but no drastic adjustment is required;

[0081] If the mean value of the interference degree of the environmental perception data is greater than or equal to the reference threshold of the mean value of the interference degree of the environmental perception data, and the standard deviation of the interference degree of the environmental perception data is greater than or equal to the reference threshold of the standard deviation of the interference degree of the environmental perception data, it means that the interference degree of the environmental perception data is high and fluctuates greatly, and the interference is strong and unstable. Due to the strong and unstable interference, it is recommended to significantly adjust the path and try to avoid the area to prevent unpredictable signal interference from adversely affecting the path planning;

[0082] If the mean value of the interference degree of the environmental perception data is less than the reference threshold value of the mean value of the interference degree of the environmental perception data, and the standard deviation of the interference degree of the environmental perception data is greater than or equal to the reference threshold value of the standard deviation of the interference degree of the environmental perception data, it means that the interference degree of the environmental perception data is low but fluctuates greatly, and the interference is sometimes strong and sometimes weak. Since the interference is generally small but occasionally fluctuates greatly, it is recommended to make a small range of path adjustment to reduce the risk of interference, but there is no need for excessive adjustment;

[0083] If the mean interference level of the environmental perception data is less than the reference threshold of the mean interference level of the environmental perception data, and the standard deviation of the interference level of the environmental perception data is less than the reference threshold of the standard deviation of the interference level of the environmental perception data, the interference level of the environmental perception data is low and has small fluctuations, and the interference is stable and weak. Since the interference is small and stable, it is recommended that the route does not need to be adjusted, and the vehicle can continue to travel along the original route without special avoidance.

[0084] In this embodiment, the autonomous driving system collects reflected signals from different directions through the radar and lidar sensors on the vehicle, and compares and analyzes the real-time collected signals with the static environmental information based on the basic environmental model constructed by SLAM technology to identify the false information area. For the false information area, the system collects the changes in the signal reflection distance and signal strength, analyzes the proportion of abnormally high-intensity signals, evaluates the signal stability, and evaluates the complexity of the signal path by analyzing the difference in the arrival time of the signal. According to the stability of the signal and the complexity of the path, the system divides the interference degree of the signal anomaly on the environmental perception data into high interference, general interference and low interference levels. At low interference levels, sensor data can be directly used to update the environmental model in real time and detect dynamic objects; at high interference levels, the system reduces the vehicle speed and gradually corrects the path based on low interference data and static environmental models. For general interference levels, the system further predicts the interference degree within a fixed time period, and optimizes the path adjustment range based on the prediction results to ensure safety and efficiency.

[0085] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0086] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0087] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks or optical disks.

[0088] The above description is only a specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the 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.

Claims

1. An environment perception system based on intelligent driving, characterized by: It includes signal acquisition module, comparison and analysis module, signal strength analysis module, signal path complexity evaluation module, interference level classification module, interference processing module and path optimization module; Signal acquisition module: The autonomous driving system uses several radar and lidar sensors installed on the vehicle to obtain reflected signals from different directions during the vehicle's driving process; Comparative analysis module: The basic environment model around the vehicle is pre-built through SLAM technology, and the reflection signal data collected in real time is compared and analyzed with the static environment information in the basic environment model, and the false information area is determined based on the analysis results; Signal strength analysis module: For false information areas, the signal strength data of radar and lidar are collected according to the changes in signal reflection distance, and the changes in the proportion of abnormally high-intensity signals in the echo signals are analyzed to evaluate the stability of the echo signal strength; Signal path complexity assessment module: By comparing the arrival time of the signal, the signal delay state of several radar and lidar sensors detecting the same object is analyzed to assess the complexity of the signal path; Interference level classification module: According to the stability of echo signal strength and the complexity of signal path, the interference degree of signal anomaly on environmental perception data is evaluated, and according to the evaluation results, it is divided into high interference level, general interference level and low interference level; Interference processing module: For low interference levels, radar and lidar data are directly used to update the basic environment model and dynamic object detection in real time; for high interference levels, driving is reduced based on low interference level data and static environment model, and the driving path is gradually corrected; Path optimization module: For general interference levels, further predictions are made on the degree of interference of signal anomalies on environmental perception data within a fixed time period, and the path adjustment range is optimized based on the prediction results.

2. The environment perception system based on intelligent driving according to claim 1, characterized in that: In the comparison and analysis module, the reflection signal data collected in real time is compared and analyzed with the static environment information in the basic environment model, specifically: Extract the location information of static objects from the basic environment model built by SLAM, and compare it with the radar and lidar reflection signals collected in real time; The comparison is based on the position, size, and shape characteristics of objects in three-dimensional space. If the real-time data matches the position and size of the object in the environment model, the object is considered to be a known static object and does not require further processing; For objects in real-time data that fail to match the environmental model, the system will further analyze the characteristics of their reflection signals to determine whether they are dynamic objects. If the object appears to be static but does not match the position in the basic environmental model, it will be marked as a false information area.

3. The environment perception system based on intelligent driving according to claim 1, characterized in that: In the signal strength analysis module, the signal strength in the T time period in the false information area is analyzed in time series, and the abnormal signal ratio fluctuation index is generated according to the change in the abnormal high-intensity signal ratio in the echo signal. The method for obtaining the abnormal signal ratio fluctuation index is: In the T time period, the abnormal signal ratio at time point t is calculated, that is, the ratio of abnormal high-intensity signals to all signals is marked as , and establish the corresponding time series set, calculate the similarity between each time point in the time series and its lagged value , the expression is: ;W is the sliding window size, is the lag time interval, is the abnormal signal ratio at time point i, For time point The abnormal signal ratio, is the mean value of the abnormal signal ratio at time point t in the sliding window W, and the calculation formula is: The sliding window starts from time t=1 and moves forward step by step until it covers the entire time period T. The fluctuation of the abnormal signal ratio in the entire time period T is calculated, that is, the abnormal signal ratio fluctuation index is calculated. The expression is: ; Where FD is the abnormal signal proportion fluctuation index.

4. The environment perception system based on intelligent driving according to claim 3 is characterized in that: In the signal path complexity evaluation module, the signal delay status of the same object detected by several radar and lidar sensors is analyzed, and the delay deviation anomaly index is generated after comparing the delay differences of the signals of several sensors. The method for obtaining the delay deviation anomaly index is as follows: The signal delay data reaching the same object is obtained from n sensors, and the signal delay difference data of different sensors are sorted into a matrix form. The matrix X is expressed as: ; It represents the signal delay difference detected by the nth sensor at the mth time point. After the signal delay difference in the matrix is ​​standardized, the linear relationship between the delay data of different sensors is calculated. The expression is: ; In the formula, T is the matrix transpose, Z is the standardized delay data matrix, C is an n×n matrix, which expresses the covariance between the signal delay difference data. The covariance matrix C is subjected to eigenvalue decomposition to obtain the eigenvalue and the corresponding eigenvector ; is the eigenvalue of the nth principal component, is the eigenvector corresponding to the nth principal component; ; Project the original standardized data Z onto the selected principal component to obtain the principal component score S, which is expressed as: ; In the formula, is a matrix composed of the eigenvectors of the first k principal components. According to the principal component scores, the delay deviation anomaly index is calculated. The expression is: ; In the formula, is the score of the jth principal component, is the delay deviation anomaly index.

5. The environment perception system based on intelligent driving according to claim 4 is characterized in that: In the interference level classification module, the abnormal signal proportion fluctuation index and the delay deviation abnormal index are converted into the first eigenvector, and the first eigenvector is used as the input of the machine learning model. The machine learning model predicts the interference degree value label of the environmental perception data with each group of first eigenvectors as the prediction target, and minimizes the sum of prediction errors of the interference degree value labels of all environmental perception data as the training target. The machine learning model is trained until the sum of prediction errors converges, and the model training is stopped. The interference degree value of the environmental perception data is determined according to the model output results, wherein the machine learning model is a polynomial regression model.

6. The environment perception system based on intelligent driving according to claim 5, characterized in that: Compare the interference degree value of the acquired environmental perception data with a gradient standard threshold, where the gradient standard threshold includes a first standard threshold and a second standard threshold, and the first standard threshold is less than the second standard threshold, and compare the interference degree value of the environmental perception data with the first standard threshold and the second standard threshold respectively; If the interference level of the environmental perception data is greater than the second standard threshold, it means that the interference level of the signal anomaly on the environmental perception data is high, and a high anomaly signal is generated, which is classified as a high interference level; If the interference level of the environmental perception data is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, it means that the interference level of the signal anomaly on the environmental perception data is general, and a general anomaly signal is generated, which is classified as a general interference level; If the interference level value of the environmental perception data is less than the first standard threshold, it means that the interference level of the signal anomaly on the environmental perception data is low. At this time, a low anomaly signal is generated and classified as a low interference level.

7. The environment perception system based on intelligent driving according to claim 1, characterized in that: In the path optimization module, for the general interference level, further prediction is made on the interference degree of signal anomalies on environmental perception data within a fixed time period; For the general interference level, that is, the interference degree value of the environmental perception data generated within a fixed time period is greater than or equal to the first standard threshold and less than or equal to the second standard threshold, the interference degree values ​​of the environmental perception data generated in the subsequent fixed time period that are greater than or equal to the first standard threshold and less than or equal to the second standard threshold are collected, and the corresponding data set is established, the mean and standard deviation of the data set are calculated, and after analysis and prediction, the path adjustment range is optimized according to the prediction results.

8. The environment perception system based on intelligent driving according to claim 7, characterized in that: If the mean value of the interference degree of the environmental perception data in the data set is greater than or equal to the reference threshold of the mean value of the interference degree of the environmental perception data, and the standard deviation of the interference degree of the environmental perception data is less than the reference threshold of the standard deviation of the interference degree of the environmental perception data, it means that the interference degree of the environmental perception data is high but the fluctuation is small, and the interference is stable. The path is adjusted to a moderate degree to avoid the continuous influence of the high interference area; If the mean value of the interference degree of the environmental perception data is greater than or equal to the reference threshold of the mean value of the interference degree of the environmental perception data, and the standard deviation of the interference degree of the environmental perception data is greater than or equal to the reference threshold of the standard deviation of the interference degree of the environmental perception data, it means that the interference degree of the environmental perception data is high and fluctuates greatly, and the interference is strong and unstable. The path is adjusted significantly to avoid the area to prevent unpredictable signal interference from causing adverse effects on path planning; If the mean value of the interference degree of the environmental perception data is less than the reference threshold value of the mean value of the interference degree of the environmental perception data, and the standard deviation of the interference degree of the environmental perception data is greater than or equal to the reference threshold value of the standard deviation of the interference degree of the environmental perception data, it means that the interference degree of the environmental perception data is low but fluctuates greatly, and the interference is sometimes strong and sometimes weak. A small range of path adjustment is performed to reduce the risk of interference; If the mean value of the interference degree of the environmental perception data is less than the reference threshold of the mean value of the interference degree of the environmental perception data, and the standard deviation of the interference degree of the environmental perception data is less than the reference threshold of the standard deviation of the interference degree of the environmental perception data, the interference degree of the environmental perception data is low and the fluctuation is small, the interference is stable and weak, and there is no need to adjust the path. The vehicle continues to travel along the original path.

Images