Method, device, processor and computer-readable storage medium for implementing automatic driving function degradation processing in severe weather and environment

Abstract

The present invention relates to a method for realizing degradation processing of an automatic driving function in severe weather and environment, the method comprising: identifying the state of the surrounding environment through a perception system module provided on the driving vehicle, obtaining the risks of the currently affected sensors and the risks of the weather conditions; inputting the judgment result into a degradation state management module to evaluate the severity of the external environment, and obtaining the risk level of the current vehicle environment; implementing a corresponding degradation strategy according to the degradation level result output by the degradation state management module, and performing corresponding automatic driving decision execution. The present invention also relates to a corresponding device, processor and storage medium thereof. The method, device, processor and storage medium thereof for realizing degradation processing of an automatic driving function in severe weather and environment of the present invention can identify different types of weather and environment that are not conducive to driving, and can freely match and combine different degradation strategies according to different environments.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving, and specifically refers to a method, device, processor and computer-readable storage medium thereof for implementing autonomous driving function degradation processing in severe weather and environment. Background Art

[0002] At present, in the vehicle driving scene, facing bad weather (such as rain, fog and snow) and dark environment, the driver's perception and range of the surrounding environment will drop sharply, increasing the risk of driving in such environment. If the automatic driving system is used, the sensor interference or occlusion in extreme environment will also affect the intelligent driving system's perception performance of the surrounding environment.

[0003] Prior art CN202210272904.4 discloses a method of using a camera sensor and wiper gear position for fusion judgment, and adjusting the speed or following strategy according to the amount of rain and snow. The problem with this solution is that the camera only judges whether there is rain or snow, and the amount of rain and snow depends on the wiper gear position. If the algorithm of the rain sensor is poor, it is very easy to make a wrong judgment. At the same time, this solution cannot cope with foggy weather.

[0004] Prior art CN202211552096.3 discloses a method of using laser radar point cloud noise data to judge rain, fog and snow weather. When the radar noise is too large, the confidence of the result output by the sensor is reduced, which is easy to affect driving and reduce speed. The problem with this solution is that it is only based on the data of a single laser radar and is easily interfered by the external environment, resulting in misidentification. Summary of the invention

[0005] The purpose of the present invention is to overcome the shortcomings of the above-mentioned prior art and to provide a method, device, processor and computer-readable storage medium for realizing automatic driving function degradation processing in severe weather and environment.

[0006] In order to achieve the above-mentioned purpose, the method, device, processor and computer-readable storage medium for implementing automatic driving function degradation processing in severe weather and environment of the present invention are as follows:

[0007] The method for implementing automatic driving function degradation processing in severe weather and environment is mainly characterized in that the method comprises the following steps:

[0008] (1) Identify the state of the surrounding environment through the perception system module installed on the driving vehicle, and obtain the risks of the currently affected sensors and the risks of the weather conditions;

[0009] (2) Inputting the judgment result into the degradation status management module to evaluate the severity of the external environment and obtain the risk level of the current vehicle environment;

[0010] (3) The system implements a corresponding downgrade strategy for the vehicle according to the downgrade level result output by the downgrade status management module, and performs corresponding autonomous driving decision execution.

[0011] Preferably, the sensor risk assessment process is performed in the following manner:

[0012] (a1) Configure the basic weight sensor_weights(SW) of each visual sensor in the perception function, where SW is the weight of each sensor;

[0013] (a2) Set the sensor degradation level risk parameter sensor_status_risk (SSR), the sensor weather impact factor weather_impact (WI), and the risk severity of weather factors weather_risk (WR) according to the actual situation. SSR represents the degree of limitation of each sensor through machine learning scene recognition, WI represents the impact of each sensor in different weather environments, and WR represents the risk severity of each weather impact on the system.

[0014] (a3) Calculate the comprehensive risk of each sensor according to the following formula:

[0015]

[0016] Among them, i represents the defined weather type, j represents the defined sensor type, N and M represent the number of defined weather types and sensor types respectively. sensor Normalize and take the value [0,1].

[0017] Preferably, the weather condition risk assessment process is performed in the following manner:

[0018] (b1) Define the basic risk weights of weather factors weather_weights (WW), where WW is the weight of each weather factor;

[0019] (b2) Define the risk severity of each weather factor weather_risk(WR) according to the actual situation;

[0020] (b3) Calculate the comprehensive risk of current weather conditions according to the following formula:

[0021]

[0022] Where i represents the defined weather type and N is the number of defined weather types.

[0023] Preferably, the perception system module performs judgment and processing inside the module by combining the degree of influence of the camera, lidar and millimeter-wave radar and the effect of fusion imaging, so as to obtain the degree of environmental influence after multi-sensor fusion.

[0024] Preferably, the degradation status management module outputs the corresponding perceived function impact level light according to the acquired environmental impact level, including: Normal, Degraded, Partially Failed or Failed; and calculates the system comprehensive risk level R in the following manner: system :

[0025] R system =α·R weather +β·R sensor

[0026] Among them, α is the weight of weather condition risk, R weather is the weather status risk level, β is the weight of sensor status risk, R sensor It is the risk level of the sensor status.

[0027] Preferably, the step (3) is specifically as follows:

[0028] The system performs corresponding function and parameter adjustments through the global function controller and the global parameter adjuster according to the degradation level results output by the degradation status management module, so as to implement different autonomous driving degradation strategies.

[0029] Preferably, the method further comprises:

[0030] Manually correct the currently acquired sensor results, including: environment type, environment severity and whether to enforce manual operation. Enforcement means that manual setting can be selected as reference or highest priority processing.

[0031] The main features of the device for implementing automatic driving function degradation processing in severe weather and environment are as follows:

[0032] a processor configured to execute computer executable instructions;

[0033] A memory stores one or more computer executable instructions, which, when executed by the processor, implement the steps of the method for implementing the degradation of the autonomous driving function in the above-mentioned severe weather and environment.

[0034] The processor for implementing degradation processing of autonomous driving functions in severe weather and environment has the main feature that the processor is configured to execute computer executable instructions, and when the computer executable instructions are executed by the processor, the steps of the method for implementing degradation processing of autonomous driving functions in severe weather and environment described above are implemented.

[0035] The main feature of the computer-readable storage medium is that a computer program is stored thereon, and the computer program can be executed by a processor to implement the steps of the method for implementing the degradation of the automatic driving function in the above-mentioned severe weather and environment.

[0036] The method, device, processor and computer-readable storage medium thereof for implementing degradation processing of the autonomous driving function in severe weather and environment of the present invention are adopted. In view of the limitations of the perception of surrounding environmental conditions in severe weather and the singleness of the degradation strategy, various types of weather and surrounding environmental conditions that are not conducive to driving can be identified, and the impact of the vehicle's perception performance can be judged based on multi-sensor data to reduce the situation of degradation upon entering the weather conditions, thereby formulating different degradation strategies according to the characteristics of different weather and environmental conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a flow chart of the method for implementing automatic driving function degradation processing in the severe weather and environment of the present invention. DETAILED DESCRIPTION

[0038] In order to more clearly describe the technical content of the present invention, further description is given below in conjunction with specific embodiments.

[0039] Before describing in detail embodiments according to the present invention, it should be noted that, hereinafter, the terms "comprises", "includes" or any other variations are intended to cover non-exclusive inclusion, whereby a process, method, article or apparatus comprising a series of elements includes not only these elements, but also other elements not explicitly listed or inherent to such process, method, article or apparatus.

[0040] See also Figure 1 As shown, the method for implementing the degradation of the automatic driving function in the severe weather and environment, wherein the method comprises the following steps:

[0041] (1) Identify the state of the surrounding environment through the perception system module installed on the driving vehicle, and obtain the risks of the currently affected sensors and the risks of the weather conditions;

[0042] (2) Inputting the judgment result into the degradation status management module to evaluate the severity of the external environment and obtain the risk level of the current vehicle environment;

[0043] (3) The system implements a corresponding downgrade strategy for the vehicle according to the downgrade level result output by the downgrade status management module, and performs corresponding autonomous driving decision execution.

[0044] As a preferred embodiment of the present invention, the sensor risk assessment process is performed in the following manner:

[0045] (a1) Configure the basic weight sensor_weights(SW) of each visual sensor in the perception function, where SW is the weight of each sensor;

[0046] (a2) Set the sensor degradation level risk parameter sensor_status_risk (SSR), the sensor weather impact factor weather_impact (WI), and the risk severity of weather factors weather_risk (WR) according to the actual situation. SSR represents the degree of limitation of each sensor through machine learning scene recognition, WI represents the impact of each sensor in different weather environments, and WR represents the risk severity of each weather impact on the system.

[0047] (a3) Calculate the comprehensive risk of each sensor according to the following formula:

[0048]

[0049] Among them, i represents the defined weather type, j represents the defined sensor type, N and M represent the number of defined weather types and sensor types respectively. semsor Normalize and take the value [0,1].

[0050] As a preferred embodiment of the present invention, the weather condition risk assessment process is performed in the following manner:

[0051] (b1) Define the basic risk weights of weather factors weather_weights (WW), where WW is the weight of each weather factor;

[0052] (b2) Define the risk severity of each weather factor weather_risk(WR) according to the actual situation;

[0053] (b3) Calculate the comprehensive risk of current weather conditions according to the following formula:

[0054]

[0055] Where i represents the defined weather type and N is the number of defined weather types.

[0056] As a preferred embodiment of the present invention, the perception system module performs judgment and processing inside the module by combining the degree of influence and the effect of fusion imaging of the camera, lidar and millimeter-wave radar, so as to obtain the degree of environmental impact after multi-sensor fusion.

[0057] As a preferred embodiment of the present invention, the degradation status management module outputs the corresponding perceived function impact level light according to the acquired environmental impact level, including: Normal, Degraded, Partially Failed or Failed; and calculates the system comprehensive risk level R in the following manner: system :

[0058] R system =α·R weather +β·R sensor

[0059] Among them, α is the weight of weather condition risk, R weather is the weather status risk level, β is the weight of sensor status risk, R sensor is the risk level of the sensor status. In practical applications, based on the calculated R system , define different R system The corresponding system degradation degree is shown in the following table as an example. The actual parameter values ​​are adjusted according to the calibration effect:

[0060]

[0061] As a preferred embodiment of the present invention, the step (3) is specifically as follows:

[0062] The system performs corresponding function and parameter adjustments through the global function controller and the global parameter adjuster according to the degradation level results output by the degradation status management module, so as to implement different autonomous driving degradation strategies.

[0063] As a preferred embodiment of the present invention, the method further comprises:

[0064] Manually correct the currently acquired sensor results, including: environment type, environment severity and whether to enforce manual operation. Enforcement means that manual setting can be selected as reference or highest priority processing.

[0065] The device for implementing automatic driving function degradation processing in severe weather and environment, wherein the device comprises:

[0066] a processor configured to execute computer executable instructions;

[0067] A memory stores one or more computer executable instructions, which, when executed by the processor, implement the steps of the method for implementing the degradation of the autonomous driving function in the above-mentioned severe weather and environment.

[0068] The processor for implementing degradation processing of the autonomous driving function in severe weather and environment, wherein the processor is configured to execute computer-executable instructions, and when the computer-executable instructions are executed by the processor, the steps of the method for implementing degradation processing of the autonomous driving function in severe weather and environment described above are implemented.

[0069] The computer-readable storage medium stores a computer program thereon, and the computer program can be executed by a processor to implement the steps of the method for implementing the degradation of the autonomous driving function in the above-mentioned severe weather and environment.

[0070] The technical solution will be further described in detail below:

[0071] 1. Environmental Identification:

[0072] The recognition of surrounding environment status mainly relies on three data sources: cameras, lidar, and manual input.

[0073] The impact of rain, fog, snow or dark light environment on visual sensors is mainly reflected in the imaging features. Scene classification is used to identify rainy and foggy weather or scenes that affect driving. The scene types that can be identified and output to the state management module in the form of enumeration include but are not limited to: rainy days, foggy days, snowy days, etc. LiDAR measures distance by emitting lasers and receiving light beams reflected by objects, thereby generating dense three-dimensional point clouds. In rainy and foggy weather, the laser beam is affected by absorption, reflection and refraction to produce noise interference, resulting in a large number of noise points in the three-dimensional point cloud, interfering with the perception of the drivable space. The result of noise interference generated by the laser radar inside the system reflects the extent of rain, snow and fog in the environment, and therefore also outputs the degree of influence of the sensor.

[0074] The affected sensors mainly output the impact of external weather on the sensor system from two aspects: weather risk and sensor risk:

[0075] (1) Weather risks

[0076] Weather risks include the type of weather environment and the severity of the weather environment.

[0077] Weather environment type:

[0078] The actual environment reflects what is happening in the real world, and there is no need to take downgrade measures for abnormal weather with low severity. Therefore, it is only necessary to identify the surrounding environment as efficiently and widely as possible to reduce missed detections, and use machine learning to use high recall indicators to label and train it to improve the detection rate of scene recognition.

[0079] Environmental severity:

[0080] The severity indicates the severity of the weather environment for detection, which has a high reference for the degree of system degradation. Therefore, a high precision index is used for annotation to improve the detection accuracy. The purpose of this step is to obtain quantitative data of the severity of the environment for subsequent calculations.

[0081] First, define the basic risk weights of weather factors weather_weights (WW), WW rain , WW snow , WW fog ,…, WW is the weight of each weather (for example, WW rain =0.3, WW snow =0.2) Different weather type weights can be added according to the system design, and the sum of all weather weights equals 1. The design of different weights represents the sensitivity of the system to different types of weather. The larger the weight, the more sensitive it is to that type of weather, and the more likely it is to cause degradation.

[0082] Define the risk severity of each weather factor weather_risk(WR). The threshold is defined using the quantitative data of environmental severity obtained through machine learning mentioned above, such as WR rain Rainfall, WR fog Visibility, etc. Here, the severity is graded, multiple grades are distinguished, and for different coefficients, the visibility in fog is WR. fog For example, if visibility is greater than 1000m, WR fog =0, visibility is less than 300m, then WR fog =0.3, visibility is less than 100m, then WR fog =0.6, and so on, the value range is [0,1].

[0083] Calculating the comprehensive risk of weather i represents the defined weather type. To prevent the occurrence of a single weather type and the severity of the weather type reaching the highest level, but ignoring its danger due to the low proportion of the basic risk weight, it can be set that when one of the weather types WR x If the risk value exceeds the set value (for example, when WR>0.8), the risk value is increased.

[0084] (2) Sensor risks

[0085] Different sensors are affected differently by weather, so the degree of functional limitation of each sensor needs to be calculated according to different weather types.

[0086] First, configure the basic weights sensor_weights (SW) of the sensor in the perception function. SW camera , S.W. lidar etc., the sum of all sensor weights is 1.

[0087] Set the sensor degradation level risk parameter sensor_status_risk (SSR) and the sensor weather impact factor weather_impact (WI). SSR means that each sensor uses machine learning to identify the scene and judge its own current limitation level, define multiple limitation levels and assign coefficients of [0,1]. The larger the coefficient, the more serious the level. WI indicates the impact level of each sensor in different weather environments. For example, WI rain,camera Indicates the degree of impact of the camera on rainy days, using a coefficient of [0,1], the larger the coefficient, the more serious the impact. The following table is a partial illustration of the parameter settings, and the specific parameters need to be adjusted according to the sensor and algorithm performance:

[0088]

[0089]

[0090] Calculating the integrated risk of a sensor system i represents the defined weather type, j represents the defined sensor type, and finally R sensor Normalize and maintain the value [0,1].

[0091] Manual input will output three results: environment type, severity of adverse environment, and whether it is mandatory. Due to the performance limitations of sensors and algorithms, the output results may not be completely accurate. At this time, manual operation can be added on the vehicle interface to make certain corrections to the sensor results. At the same time, you can choose whether to force the option, indicating whether the manual setting is a reference or the highest priority. The results of the manual output of the status management module on the vehicle include the environment type (enumeration: no abnormal weather, rainy days, foggy days, snowy days, snow on the ground, etc.), the severity of the environment (enumeration: normal, limited function, severely limited function, function failure) and whether it is mandatory (Boolean value: yes, no). In the human-computer interaction interface, the above display names can be modified to meet the driver's needs for quick understanding.

[0092] 2. Status Information Processing

[0093] The affected states of different sensors are first processed by the perception system module. The perception module first combines the degree of impact of the camera, lidar, and millimeter-wave radar and the effect of fusion imaging, processes it inside the module, determines the degree of environmental impact after multi-sensor fusion, and outputs the degree of impact of the perception function to the status management module: Normal, Degraded, Partially Failed, Failed (the function level is determined according to actual needs). The calculation method of the comprehensive system risk level of task management is R system =α·R weather +β·R sensor , α and β are weights, representing the relative importance of sensor status and weather to the overall system risk, α+β=1. For example, when the sensor status is more important, α=0.7, β=0.3 can be set. system Develop corresponding driving strategies, such as R system ≤0.3 means Normal, 0.3 <R system ≤0.6 indicates Degrade. In addition, the perception module forwards the environment type judgment data output by the camera to the status management module.

[0094] The status management module also receives status information from manual input, and determines the final intelligent driving function degradation level in combination with the received perception status information. The judgment method is: receive the severity of the function impact from the perception and manual output, and take the most serious level of the two. If a mandatory flag signal is received from the manual information source, the manual function impact severity is taken as the final result.

[0095] 3. Functional degradation

[0096] When the state management module completes the judgment of information from multiple sources, it will output the final degradation level result to the downstream algorithm modules (perception, decision-making, planning, etc.). Each algorithm module adopts different degradation strategies according to different levels and environmental weather information. The degradation strategies include but are not limited to:

[0097] Reduce driving speed

[0098] Increase the following distance

[0099] Adjust the acceleration and braking speed

[0100] Reduce lane change intentions

[0101] Increase lane-changing safety distance

[0102] Reduce the confidence ratio of the most affected sensors

[0103] The above downgrade strategy:

[0104] (1) Different thresholds are set according to different degradation levels. As the degradation level increases, the restrictions on vehicle performance are increased. For example, as the degradation level increases, the speed of the autonomous driving vehicle is reduced by a percentage based on the maximum target cruising speed.

[0105] (2) The downgrade strategy function realizes the hybrid structure of memory global unified control and module autonomous adjustment. Through the function controller and parameter adjustment, the strategy is divided into global management and individual algorithm modules to ensure global consistent controllability while providing autonomy and responsiveness within the algorithm module. The specific design is as follows:

[0106] 1. After receiving the information of sensor risk, weather risk and manual input, the task management module calculates the degradation level in the degradation status management module according to the comprehensive risk of the system.

[0107] 2. The global function controller and global parameter adjuster receive different degradation levels and adjust the functions according to the defined degradation strategy. The global controller is responsible for unified management of the function switches of each module (function cutting or degradation), and handles cross-module dependencies, for example: turning off the lane change function when downgrading to Danger. The global parameter adjuster is responsible for managing global shared parameters, such as resource allocation, global thresholds, etc., and checking whether the parameters adjusted autonomously by the algorithm module meet the global constraints.

[0108] 3. The two global managers send the relevant constraint information to the downstream algorithm module, and the executor related to the algorithm module executes the degradation strategy. At the same time, the algorithm module can autonomously optimize and adjust the constraint information of the global manager according to the current environmental status detected and its own strategy, but its adjustment threshold cannot exceed the upper constraint. For example: the perception algorithm module can adjust the proportion according to the real-time perceived environment (for example: 60% for vision, 20% for laser, and 20% for millimeter wave), but the proportion of a single sensor cannot be increased to 95% during the adjustment process (due to the global parameter adjuster constraint).

[0109] The algorithm module reports the current execution results (functions used and related parameters) to the task management at a fixed frequency, and the global manager performs verification and inspection. If the execution results of a certain module cannot meet the demand constraints for a long time, a strategy to exit the intelligent driving is implemented to ensure the consistency and security of the system.

[0110] The core logic and main innovation of this solution are: the autonomous driving system combines multiple data sources (including various types of sensor data and manually set information) to summarize and judge the risk of autonomous driving, and implements different levels of intelligent driving function degradation and response strategies accordingly. Since different types of sensors have different sensitivities to external interference, it is difficult to quantify them with a unified indicator. Therefore, the environmental interference identification is first determined by each sensor independently, and then the results are input into the state management module. The state management module conducts unified arbitration to determine the severity of the external environment, and outputs different intelligent driving function degradation levels to the algorithm module to ensure safe autonomous driving even in harsh environmental conditions.

[0111] The difference between this technical solution and the prior art is that:

[0112] (1) The existing technology mainly relies on a single type of sensor to judge the environmental conditions (such as laser radar or rain sensor). This type of solution is relatively limited in terms of the perception of environmental types and is only effective for specific types of weather (such as rainy days and snowy days). In this case, when faced with more complex and diverse environmental conditions, this solution adopts a dual sensor judgment method of camera and laser radar. The camera visual perception is trained by machine learning through a large number of samples of different weather and environments (such as rainy days, foggy days, snowy days, and wet ground). It can output a variety of complex environmental type combinations. At the same time, the change in the camera's object detection distance performance can feedback the visibility level and severity of the environment, providing a reference for the degree of degradation. Laser radar measures the distance by emitting lasers and receiving the beam reflected by the object, thereby generating a dense three-dimensional point cloud. In rainy and foggy weather, the laser beam is affected by absorption, reflection, and refraction to produce noise interference, resulting in a large number of noise points in the three-dimensional point cloud, which enables the laser radar algorithm to judge the current environmental interference conditions and output the severity of the impact on the sensor, which is used to determine the level of vehicle degradation. At the same time, the impact of sensor missed recognition and no recognition is reduced, and manual operation is introduced to manually correct the environment and degradation. Combining the above data, the system's perception performance in different environments can be greatly improved.

[0113] (2) The existing technology is limited in identifying weather types or cannot distinguish different weather types, resulting in a single means of downgrading in bad weather, and it is unable to follow up on different weather conditions to formulate different strategies. This solution can identify different types of weather and environments that are not conducive to driving through the camera, and can freely combine different downgrading strategies according to different environments, including but not limited to deceleration, increasing the following distance, and limiting acceleration and deceleration. While ensuring safety, it improves user experience and commuting efficiency.

[0114] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, segment or portion of code that includes one or more executable instructions for implementing the steps of a specific logical function or process, and the scope of the preferred embodiments of the present invention includes alternative implementations in which functions may not be performed in the order shown or discussed, including performing functions in a substantially simultaneous manner or in the reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present invention belong.

[0115] It should be understood that each part of the present invention can be implemented by hardware, software, firmware or a combination thereof. In the above embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution device.

[0116] A person of ordinary skill in the art may understand that all or part of the steps of the method for implementing the above-mentioned embodiment may be completed by instructing the relevant hardware through a program, and the program may be stored in a computer-readable storage medium, which, when executed, includes one of the steps of the method embodiment or a combination thereof.

[0117] The storage medium mentioned above can be a read-only memory, a magnetic disk or an optical disk, etc.

[0118] In the description of this specification, the description with reference to the terms "an embodiment", "some embodiments", "example", "specific example", or "embodiment" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0119] Although the embodiments of the present invention have been shown and described above, it is to be understood that the above embodiments are exemplary and are not to be construed as limitations of the present invention. A person skilled in the art may change, modify, replace and vary the above embodiments within the scope of the present invention.

[0120] The method, device, processor and computer-readable storage medium thereof for implementing degradation processing of the autonomous driving function in severe weather and environment of the present invention are adopted. In view of the limitations of the perception of surrounding environmental conditions in severe weather and the singleness of the degradation strategy, various types of weather and surrounding environmental conditions that are not conducive to driving can be identified, and the impact of the vehicle's perception performance can be judged based on multi-sensor data to reduce the situation of degradation upon entering the weather conditions, thereby formulating different degradation strategies according to the characteristics of different weather and environmental conditions.

[0121] In this specification, the present invention has been described with reference to specific embodiments thereof. However, it is apparent that various modifications and variations may be made without departing from the spirit and scope of the present invention. Therefore, the specification and drawings should be regarded as illustrative rather than restrictive.

Claims

1. A method for implementing automatic driving function degradation processing in severe weather and environment, characterized in that: The method comprises the following steps: (1) Identify the state of the surrounding environment through the perception system module installed on the driving vehicle, and obtain the risks of the currently affected sensors and the risks of the weather conditions; (2) Inputting the judgment result into the degradation status management module to evaluate the severity of the external environment and obtain the risk level of the current vehicle environment; (3) The system implements a corresponding downgrade strategy for the vehicle according to the downgrade level result output by the downgrade status management module, and performs corresponding autonomous driving decision execution.

2. The method for implementing automatic driving function degradation processing in severe weather and environment according to claim 1, characterized in that: The sensor risk assessment process is performed as follows: (a1) Configure the basic weight sensor_weights(SW) of each visual sensor in the perception function, where SW is the weight of each sensor; (a2) Set the sensor degradation level risk parameter sensor_status_risk (SSR), the sensor weather impact factor weather_impact (WI), and the risk severity of weather factors weather_risk (WR) according to the actual situation. SSR represents the degree of limitation of each sensor through machine learning scene recognition, WI represents the impact of each sensor in different weather environments, and WR represents the risk severity of each weather impact on the system. (a3) Calculate the comprehensive risk of each sensor according to the following formula: Among them, i represents the defined weather type, j represents the defined sensor type, N and M represent the number of defined weather types and sensor types respectively. sensor Normalize and take the value [0,1].

3. The method for implementing automatic driving function degradation processing in severe weather and environment according to claim 1, characterized in that: Weather condition risk assessment is processed as follows: (b1) Define the basic risk weights of weather factors weather_weights (WW), where WW is the weight of each weather factor; (b2) Define the risk severity of each weather factor weather_risk(WR) according to the actual situation; (b3) Calculate the comprehensive risk of current weather conditions according to the following formula: Where i represents the defined weather type and N is the number of defined weather types.

4. The method for implementing automatic driving function degradation processing in severe weather and environment according to claim 1, characterized in that: The perception system module combines the degree of influence of the camera, lidar and millimeter-wave radar and the effect of fusion imaging, and performs judgment and processing inside the module to obtain the degree of environmental influence after multi-sensor fusion.

5. The method for implementing automatic driving function degradation in severe weather and environment according to claim 4, characterized in that: The degradation status management module outputs the corresponding perceived function impact level light according to the acquired environmental impact level, including: Normal, Degraded, Partially Failed or Failed; and calculates the system comprehensive risk level R in the following way: system : R system =α·R weather +β·R sensor Among them, α is the weight of weather condition risk, R weather is the weather status risk level, β is the weight of sensor status risk, R sensor It is the risk level of sensor status.

6. The method for implementing automatic driving function degradation in severe weather and environment according to claim 5, characterized in that: The step (3) is specifically as follows: The system performs corresponding function and parameter adjustments through the global function controller and the global parameter adjuster according to the degradation level results output by the degradation status management module, so as to implement different autonomous driving degradation strategies.

7. The method for implementing automatic driving function degradation processing in bad weather and environment according to any one of claims 1 to 6, characterized in that: The method further comprises: Manually correct the currently acquired sensor results, including: environment type, environment severity and whether to enforce manual operation. Enforcement means that manual setting can be selected as reference or highest priority processing.

8. A device for implementing automatic driving function degradation processing in severe weather and environment, characterized in that: The device comprises: a processor configured to execute computer executable instructions; A memory storing one or more computer executable instructions, wherein when the computer executable instructions are executed by the processor, the steps of the method for implementing the degradation processing of the automatic driving function in the severe weather and environment described in any one of claims 1 to 7 are implemented.

9. A processor for implementing automatic driving function degradation processing in severe weather and environment, characterized in that: The processor is configured to execute computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the method for implementing automatic driving function degradation processing in severe weather and environment described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and the computer program can be executed by a processor to implement the steps of the method for implementing automatic driving function degradation processing in severe weather and environment as described in any one of claims 1 to 7.

Images