Method for Modifying the Perception Performance Limitations of the Expected Functional Safety of Autonomous Driving and Vehicle

Abstract

The present invention discloses a method for modifying the perception performance limitation of the expected function safety of an autonomous vehicle and a vehicle. The method includes: S1. For different scenarios, assign confidence levels to the required sensors; S2. Set the main sensor for each scenario; S3. Obtain the confidence levels of each sensor in the current scenario, and determine whether the confidence level of the main sensor is less than the assigned confidence level of the sensor in this scenario; if not, the main sensor remains unchanged; if so, execute step S4; S4. Determine whether the confidence level of the auxiliary sensor is less than the assigned confidence level of the sensor in this scenario; if not, use this auxiliary sensor as the main sensor; if so, execute step S5; S5. Assign weights to the main and auxiliary sensors, and determine whether the product of the confidence level of the auxiliary sensor and the weight of the auxiliary sensor is less than that of the main sensor; if not, use this auxiliary sensor as the main sensor; if so, the main sensor remains unchanged. The present invention can effectively alleviate the potential risks caused by the performance limitations of sensors in autonomous vehicles.

Description

Technical Field

[0001] The present invention relates to the field of expected functional safety, and particularly to a method for modifying the perception performance limitation of autonomous driving expected functional safety and a vehicle. Background Art

[0002] Nowadays, autonomous driving is developing towards a higher-level driving assistance system. In order to meet the safety requirements of road vehicles, the traditional functional safety standard ISO 26262 cannot cover the driving scenarios where accidents are caused by performance limitations. The new functional safety standard ISO 21448 is supplemented to make up for the parts that cannot be covered in ISO 26262. ISO 21448 defines SOTIF (Safety Of The Intended Functionality) as the harm caused by insufficient design of the intended function specification, performance limitations, or reasonably foreseeable misuses by humans.

[0003] The perception module is the "eye" of an autonomous vehicle, and sensors are greatly affected by the external environment. Once a problem occurs, the autonomous vehicle will surely lose control. How to modify the performance limitation of the perception module of an autonomous vehicle and ensure its effectiveness is very important, otherwise it will pose a serious safety hazard to autonomous vehicles.

[0004] Generally, for the sensor arrangement of an autonomous vehicle, in order to solve the consequences in case of failure, including the fusion scheme and redundancy scheme already considered in functional safety, they are all for dealing with and preventing the vehicle from being in a temporarily blind state after sensor failure. If the driver cannot take over in time, there will be potential dangers. For a higher-level driverless driving system, there are also always safety hazards. Summary of the Invention

[0005] Aiming at the problems existing in the prior art, the present invention provides a method for modifying the perception performance limitation of autonomous driving expected functional safety and a vehicle, so as to solve the problem of safety hazards existing in autonomous vehicles due to sensor performance limitations.

[0006] The present invention provides a method for modifying the perception performance limitation of autonomous driving expected functional safety, including the following steps:

[0007] S1. Assign confidence levels to the required sensors for different scenarios;

[0008] S2. Set the main sensor for each scenario, and the remaining sensors are auxiliary sensors;

[0009] S3. Obtain the confidence levels of each sensor in the current scenario, and determine whether the confidence level of the main sensor is less than the assigned confidence level of the sensor in this scenario; if not, the main sensor remains unchanged; if so, execute step S4;

[0010] S4. Determine whether the confidence level of the secondary sensor is less than the allocated confidence level of the sensor in this scenario; if not, use this secondary sensor as the primary sensor; if so, execute step S5;

[0011] S5. Allocate the weights of the primary sensor and the secondary sensor, and determine whether the product of the confidence level of the secondary sensor and its weight is less than the product of the confidence level of the primary sensor and its weight; if not, use this secondary sensor as the primary sensor; if so, the primary sensor remains unchanged.

[0012] Furthermore, set the sensor with the highest confidence level in each scenario as the primary sensor.

[0013] Furthermore, the weight of the primary sensor is greater than or equal to the weight of the secondary sensor.

[0014] The present invention also provides a vehicle that adopts the above method for modifying the performance limitations of the expected functional safety perception of autonomous driving.

[0015] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0016] The present invention can effectively alleviate the potential risks caused by the performance limitations of sensors in autonomous driving vehicles. Description of the Drawings

[0017] Figure 1 It is a flowchart of the method for modifying the performance limitations of the expected functional safety perception of autonomous driving in the embodiments of the present invention. Detailed Embodiments

[0018] In order to make the objectives, technical solutions and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0019] The present invention takes into account the influence of the partial failure of the perception module of autonomous driving vehicles caused by external environmental factors, and provides a method and a vehicle for modifying the performance limitations of the expected functional safety perception module of autonomous driving, so as to alleviate the potential hazards caused by the failure of the perception module due to external environmental factors.

[0020] The present invention provides a method for modifying the performance limitations of the expected functional safety perception of autonomous driving, as Figure 1 shown, including the following steps:

[0021] S1. Allocate confidence levels to the required sensors for different scenarios;

[0022] S2. Set the main sensor for each scenario, and the remaining sensors are secondary sensors;

[0023] S3. Obtain the confidence levels of the sensors in the current scenario, and determine whether the confidence level of the main sensor is less than the configured confidence level of the sensor in this scenario; if not, the main sensor remains unchanged; if so, execute step S4;

[0024] S4. Determine whether the confidence level of the secondary sensor is less than the configured confidence level of the sensor in this scenario; if not, use this secondary sensor as the main sensor; if so, execute step S5;

[0025] S5. Allocate the weights of the main sensor and the secondary sensor, and determine whether the product of the confidence level of the secondary sensor and the weight of the secondary sensor is less than the product of the confidence level of the main sensor and the weight of the main sensor; if not, use this secondary sensor as the main sensor; if so, the main sensor remains unchanged.

[0026] Furthermore, set the sensor with the highest confidence level in each scenario as the main sensor.

[0027] Furthermore, the weight of the main sensor is greater than or equal to the weight of the secondary sensor.

[0028] The embodiments of the present invention assume a sensing arrangement scheme for an autonomous vehicle, with 1 front camera, 1 front lidar, and 1 front millimeter-wave radar in the forward direction, and 4 corner millimeter-wave radars on the sides. The specific parameters of each sensor are shown in Table 1:

[0029] Table 1 Sensor Parameters

[0030] Sensor Category Quantity Functional Description Detection Range Field of View Angle Front Camera 1 Detection Target Information, Lane Line Information 130-160m 120° Front Lidar 1 Distance to the Detection Target Vehicle 120-200m 100° Front Millimeter Wave Radar 1 Detection Target Information 160-200m 90° Corner Millimeter Wave Radar 4 Detection of Adjacent Lane Target Information 60-70m 120°

[0031] Based on the above sensor arrangement (assumed) scheme, for these sensors, following the process of expected functional safety, first obtain the impact caused by the non-expected behavior of the whole vehicle from the non-expected behavior of the whole vehicle, analyze its main reasons, obtain the limitations of sensing perception, that is, sensor performance defects, and the external environmental impacts under this performance defect. Table 2 is an example of sensor performance limitations and triggering conditions.

[0032] Table 2 Sensor Performance Limitations and Triggering Conditions

[0033]

[0034]

[0035] According to the classification of the use of automobile sensors, they can be divided into forward sensors and lateral sensors. Forward sensors are used to sense forward targets, including front cameras, front lidars, and front millimeter-wave radars. Lateral sensors are used to sense lateral targets, that is, targets in adjacent lanes (left and right), including 4 corner millimeter-wave radars.

[0036] For lateral sensing, functional safety uses a redundant solution for the lateral millimeter-wave radar. The millimeter-wave radar itself is less affected by the external environment. In addition to its own parameter limitations, it is only sensitive to the temperature of the external environment. Even if the millimeter-wave radar fails, the vehicle only needs to jump to the MRC mode, such as: keeping the lane unchanged and limping home at a low speed (assuming the driver has not taken over the vehicle), etc.

[0037] For forward sensing, including front camera, front lidar, and front millimeter-wave radar, the three types of sensors have different effects on different external environmental conditions, that is, they have different confidence levels in different external environments. The details are as follows:

[0038] 1. The forward binocular camera is set based on the dual-camera deflection principle. It can identify the attributes of obstacles and calculate the distance of obstacles. It has the characteristics of low cost, high recognition and long detection distance. However, it has high requirements for light intensity and cannot be used at night, in rain / snow / fog and other bad weather environments.

[0039] 2. The front laser radar is used to detect obstacles around the vehicle. It has a wide detection range, long detection distance, and high accuracy, but its penetration is not strong, and haze / rain / snow can be easily mistakenly detected as obstacles.

[0040] 3. The front millimeter-wave radar is a millimeter-wave radar device based on the Doppler effect. It has the characteristics of low cost, dynamic tracking, and long detection distance. However, it has many noise points, low resolution, and a small range angle. It is easy to misdetect or miss static obstacles, such as misidentifying speed brakes and height limit poles as obstacles.

[0041] The present invention determines whether the sensor is suitable for use in certain scenarios and sets percentages for different scenarios to define its confidence. For example, for the following environments, the front camera, front laser radar, and front millimeter wave radar have different confidences, as shown in Table 3 below:

[0042] Table 3 Confidence of sensors in different scenarios

[0043]

[0044]

[0045] The present invention assumes that the initial value of the sensor fusion scheme is that the front camera is the main sensor, and the front lidar / front millimeter-wave radar are the auxiliary sensors. The main and auxiliary sensors simultaneously detect the information of the target object, and the data during fusion uses the main sensor as the main information for vehicle execution reference. Define the weights of the main and auxiliary sensors, and define the weight of the main sensor as k 主 , and the weight of the auxiliary sensor is k 辅 .

[0046] Due to the influence of external environmental factors, the confidence data of the sensors will constantly change, that is, the limitations of sensor performance. Therefore, if the following two conditions are met, the order of the main and auxiliary sensors will be changed:

[0047] 1. When the confidence of the main sensor is not greater than its confidence in this scenario, and the confidence of any auxiliary sensor is greater than its confidence in this scenario, the auxiliary sensor with the highest confidence will be ranked as the main sensor;

[0048] 2. When the confidence of the main sensor is not greater than its confidence in this scenario, and the confidence of the auxiliary sensor is also not greater than its confidence in this scenario, if the product of the confidence of the auxiliary sensor and its weight is greater than the product of the main sensor, then this auxiliary sensor will be ranked as the main sensor.

[0049] This can alleviate the limitations of sensor performance and improve the safety of autonomous vehicles.

[0050] Those skilled in the art can easily understand that the above is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent replacements, and improvements made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for modifying the safety perception performance limitations of expected functions in autonomous driving, characterized in that, It includes the following steps: S1. Assign confidence levels to the required sensors for different scenarios; S2. Set the main sensor for each scenario, and the remaining sensors are auxiliary sensors; among them, the sensor with the highest confidence level in each scenario is set as the main sensor; S3. Obtain the confidence levels of each sensor in the current scenario, and determine whether the confidence level of the main sensor is less than the assigned confidence level of this sensor in this scenario; if not, the main sensor remains unchanged; if so, execute step S4; S4. Determine whether the confidence level of the auxiliary sensor is less than the assigned confidence level of this sensor in this scenario; if not, use this auxiliary sensor as the main sensor; if so, execute step S5; S5. Assign weights to the main sensor and the auxiliary sensor, and determine whether the product of the confidence level of the auxiliary sensor and the weight of the auxiliary sensor is less than the product of the confidence level of the main sensor and the weight of the main sensor; if not, use this auxiliary sensor as the main sensor; if so, the main sensor remains unchanged; among them, the weight of the main sensor is greater than or equal to the weight of the auxiliary sensor.

2. A vehicle, characterized in that, The vehicle adopts the method for modifying the perception performance limitation of the expected functional safety of autonomous driving described in claim 1.

Images