Automatic driving decision generation method and device, equipment and storage medium

Abstract

The invention provides an automatic driving decision generation method and device, equipment and a storage medium. The automatic driving decision generation method comprises the steps that multiple pieces of current environment information collected by multiple pieces of sensing equipment at the current moment are acquired; judging whether target historical data matched with the multiple pieces of current environment information exists or not; if the target historical data exists, determining a virtual neurotransmitter concentration corresponding to each sensing device according to a historical weight, an original weight and a reliability coefficient corresponding to each sensing device in the target historical data; and generating an automatic driving decision according to the weights of the plurality of sensing devices. According to the method, the adjustment process of the neurotransmitters in the human brain is simulated, the weights of different sensing devices are adjusted through the virtual neurotransmitters with different concentrations, and the automatic driving decision is generated according to the weights of the multiple sensing devices. By simulating the flexibility of information processing of the human brain in a complex environment, the robustness and response speed of the automatic driving system in a variable environment and under the condition that sensing equipment is more and more diverse can be ensured.

Description

Technical Field

[0001] This application relates to the field of autonomous driving technology, specifically to an autonomous driving decision generation method, apparatus, device, and storage medium. Background Technology

[0002] With the rapid development of the automotive industry, more and more cars are equipped with autonomous driving technology. Current autonomous driving systems integrate multiple sensors, such as cameras, radar, and lidar, to recognize and understand the vehicle's surroundings. After obtaining perception information from these sensors, the autonomous driving system needs to determine which information is more important and make autonomous driving decisions based on that importance.

[0003] In related technologies, decision-making models typically use preset rules or simple algorithms to weight the sensor-collected information, and the decision-making model needs to be trained frequently.

[0004] While this decision-making model is effective for pre-trained scenarios, it often lacks sufficient adaptability and flexibility in extreme or unknown situations. Furthermore, as the number of sensors on vehicles increases and the types of perceived information become more diverse, static rules often fail to make decisions quickly.

[0005] It should be noted that the information disclosed in the background section of this application is intended only to enhance the understanding of the general background of this application, and should not be construed as an admission or in any way implying that the information constitutes prior art known to those skilled in the art. Summary of the Invention

[0006] In view of this, this application provides an autonomous driving decision generation method, apparatus, device and storage medium to solve the problem that the existing technology of generating decisions by using preset rules or simple algorithms often lacks sufficient adaptability and flexibility, and that static rules often cannot make decisions quickly when the types of perceived information are increasing.

[0007] In a first aspect, embodiments of this application provide an autonomous driving decision generation method, including:

[0008] Acquire multiple pieces of current environmental information collected by multiple sensing devices at the current moment;

[0009] Based on multiple pieces of current environmental information, the virtual neurotransmitter concentration corresponding to each of the sensing devices is determined, and the virtual neurotransmitter concentration is used to characterize the weight of the corresponding sensing device;

[0010] Autonomous driving decisions are generated based on the weights of the multiple sensing devices.

[0011] In this embodiment, based on research on the balance of neurotransmitters in the human brain, the regulation process of neurotransmitters in the human brain is simulated. Different concentrations of virtual neurotransmitters are used to adjust the weights of different sensing devices, and autonomous driving decisions are generated based on the weights of multiple sensing devices. It can be understood that by mimicking the flexibility of the human brain in information processing in complex environments, the robustness and reaction speed of the autonomous driving system can be guaranteed in changing environments and with an increasing variety of sensing devices.

[0012] In one possible implementation, determining the virtual neurotransmitter concentration corresponding to each of the sensing devices based on multiple pieces of current environmental information includes:

[0013] Determine whether there is target historical data that matches the multiple current environmental information, wherein the target historical data includes multiple historical environmental information and multiple historical weights corresponding to the multiple sensing devices;

[0014] If there is target historical data that matches multiple current environmental information, then the virtual neurotransmitter concentration corresponding to each sensing device is determined according to the historical weight corresponding to each sensing device in the target historical data.

[0015] In this embodiment, when determining the virtual neurotransmitter concentration corresponding to each sensing device, it is first determined whether there is target historical data that matches multiple current environmental information. If target historical data exists, the virtual neurotransmitter concentration of each sensing device is determined according to the historical weight corresponding to each sensing device in the target historical data. It can be understood that the target historical data includes multiple historical environmental information and multiple historical weights corresponding to multiple sensing devices. When the autonomous driving system obtains the current environmental information, it can directly adjust the virtual neurotransmitter concentration of each sensing device according to the historical weights in the target historical data, without needing to use complex weight adjustment rules to adjust the virtual neurotransmitter concentration of the sensing devices. Therefore, autonomous driving decisions can be generated more quickly.

[0016] In one possible implementation, determining the virtual neurotransmitter concentration for each sensing device based on its historical weight in the target historical data includes:

[0017] Based on the historical weight, original weight, and reliability coefficient of each sensing device in the target historical data, the virtual neurotransmitter concentration corresponding to each sensing device is determined.

[0018] In this embodiment, when determining the virtual neurotransmitter concentration corresponding to each sensing device, the autonomous driving system not only refers to the historical weight of each sensing device in the target historical data, but also determines the virtual neurotransmitter concentration corresponding to each sensing device based on its original weight and reliability coefficient. It is understood that if the weight of each sensing device has been adjusted, referring to its original weight ensures that the adjustment of the virtual neurotransmitter concentration corresponding to each sensing device is based on the previous adjustment, thus obtaining a more accurate virtual neurotransmitter concentration. Furthermore, the reliability coefficient of each sensing device reflects the reliability of the current environmental information collected by each sensing device. Adjusting the virtual neurotransmitter concentration corresponding to each sensing device based on its reliability coefficient can prevent the autonomous driving system from making erroneous autonomous driving decisions.

[0019] In one possible implementation, determining the virtual neurotransmitter concentration for each sensing device based on its historical weight, original weight, and reliability coefficient in the target historical data includes:

[0020] The historical weight, original weight, and reliability coefficient corresponding to each sensing device in the target historical data are multiplied together to obtain the virtual neurotransmitter concentration corresponding to each sensing device.

[0021] In this embodiment, the historical weight, original weight, and reliability coefficient corresponding to each sensing device in the target historical data are multiplied to obtain the virtual neurotransmitter concentration for each sensing device. This allows for faster acquisition of the virtual neurotransmitter concentration for each sensing device.

[0022] In one possible implementation, the number of historical environmental information items in the target historical data that match multiple current environmental information items is greater than or equal to a preset number threshold.

[0023] In this embodiment, the current environmental information is abundant. To determine whether target historical data exists, the number of historical environmental information pieces matching the current environmental information can be compared to a preset threshold. If the number of historical environmental information pieces matching the current environmental information is greater than or equal to the preset threshold, then target historical data exists. It is understandable that if the condition for determining target historical data were that every piece of historical environmental information matched the current environmental information, it would be difficult to find the target historical data. Furthermore, a small number of historical environmental information pieces that do not match the current environmental information would not significantly impact driving decisions. Therefore, this embodiment ensures both the quantity and matching accuracy of target historical data.

[0024] In one possible implementation, the method further includes:

[0025] If no target historical data matches the multiple current environmental information, the virtual neurotransmitter concentration corresponding to each of the sensing devices is determined according to the default weight adjustment rules and the reliability of each sensing device.

[0026] In this embodiment, when no target historical data exists, the virtual neurotransmitter concentration corresponding to each sensing device is determined according to the default weight adjustment rule. It is understood that even if no target historical data is found, this embodiment can still generate an autonomous driving decision corresponding to the current environmental information, preventing situations where the system cannot be processed and ensuring the safety of autonomous driving.

[0027] In one possible implementation, the default weight adjustment rule includes:

[0028] When any of the current environmental information meets the preset conditions, the virtual neurotransmitter concentration of the sensing device associated with any of the current environmental information is increased.

[0029] In this embodiment, each piece of environmental information has corresponding preset conditions. When any piece of current environmental information meets the preset conditions, the virtual neurotransmitter concentration of the sensing device associated with that current environmental information is increased. For example, when the distance between an obstacle and the vehicle is less than a preset distance, the virtual neurotransmitter concentrations of the radar and camera are increased. It can be understood that when environmental information meets the preset conditions, it indicates that the environmental information is more important at the current moment. Therefore, increasing the virtual neurotransmitter concentration of the sensing device corresponding to this environmental information is beneficial for making corresponding autonomous driving decisions.

[0030] In one possible implementation, increasing the virtual neurotransmitter concentration of the sensing device associated with any of the current environmental information when any of the current environmental information meets a preset condition includes:

[0031] When any of the current environmental information meets the preset conditions, the virtual neurotransmitter concentration of the sensing device associated with any of the current environmental information is increased, and the virtual neurotransmitter concentration is less than or equal to the maximum value of the virtual neurotransmitter concentration.

[0032] It is understood that the virtual neurotransmitter concentration will not be increased indefinitely in this embodiment of the application. A maximum value of the virtual neurotransmitter concentration will be preset to avoid the virtual neurotransmitter concentration being too high and affecting the generation of autonomous driving decisions.

[0033] In one possible implementation, increasing the virtual neurotransmitter concentration of the sensing device associated with any of the current environmental information when any of the current environmental information meets a preset condition includes:

[0034] When any of the current environmental information meets the preset conditions, the concentration of the virtual neurotransmitter of the sensing device associated with any of the current environmental information is increased at a preset first rate.

[0035] In this embodiment, the virtual neurotransmitter concentration of the sensing device is increased at a preset first rate, and the first rate is not necessarily a fixed value. For example, as the distance between the obstacle and the vehicle decreases, the rate of increase of the virtual neurotransmitter concentration of the radar and camera increases. It can be understood that by pre-setting the rate of increase of the virtual neurotransmitter concentration, the virtual neurotransmitter concentration of the sensing device can reach the preset response threshold more quickly, thereby responding more quickly (e.g., braking).

[0036] In one possible implementation, the default weight adjustment rule includes:

[0037] When any of the current environmental information does not meet the preset conditions, the concentration of virtual neurotransmitters of the sensing device associated with any of the current environmental information is reduced.

[0038] In this embodiment, when any current environmental information does not meet preset conditions, the virtual neurotransmitter concentration of the sensing device related to that current environmental information is reduced. This step simulates the neurotransmitter recycling process in the human brain. When the stimulus of the environmental information disappears, the virtual neurotransmitter concentration of the sensing device related to that environmental information is reduced, causing the virtual neurotransmitter concentration of each sensing device to gradually return to the baseline level, thereby ensuring that the autonomous driving system does not overreact when there is no corresponding environmental information.

[0039] In one possible implementation, reducing the virtual neurotransmitter concentration of the sensing device associated with any of the current environmental information when any of the current environmental information does not meet the preset conditions includes:

[0040] When any of the current environmental information does not meet the preset conditions, the concentration of the virtual neurotransmitter of the sensing device associated with any of the current environmental information is reduced, and the concentration of the virtual neurotransmitter is greater than or equal to the initial concentration of the virtual neurotransmitter.

[0041] It is understood that the virtual neurotransmitter concentration will not be reduced indefinitely in this embodiment of the application. Instead, an initial concentration of the virtual neurotransmitter concentration will be preset to avoid the virtual neurotransmitter concentration being too low and affecting the generation of autonomous driving decisions.

[0042] In one possible implementation, reducing the virtual neurotransmitter concentration of the sensing device associated with any of the current environmental information when any of the current environmental information does not meet the preset conditions includes:

[0043] When any of the current environmental information does not meet the preset conditions, the concentration of the virtual neurotransmitter of the sensing device associated with any of the current environmental information is reduced at a preset second rate.

[0044] In this embodiment, the virtual neurotransmitter concentration of the sensing device is reduced at a preset second rate to avoid the virtual neurotransmitter concentration of one sensing device decreasing too quickly, which would result in a relatively high virtual neurotransmitter concentration in other sensing devices, thereby affecting the generation of autonomous driving decisions.

[0045] In one possible implementation, the default weight adjustment rule includes:

[0046] Based on the various pieces of current environmental information, the current application scenario is determined;

[0047] Adjust the virtual neurotransmitter concentration of the sensing device corresponding to the current application scenario.

[0048] Understandably, the importance of environmental information collected by each sensing device varies in different application scenarios. By determining the current application scenario, the virtual neurotransmitter concentration of the sensing devices corresponding to the current application scenario can be adaptively adjusted, enabling the autonomous driving system to make more accurate autonomous driving decisions more quickly.

[0049] In one possible implementation, after generating an autonomous driving decision based on the weights of the plurality of said sensing devices, the method further includes:

[0050] If the autonomous driving decision meets expectations, then the current perception information and weights corresponding to the multiple perception devices are saved as historical data.

[0051] In this embodiment, a feedback adjustment mechanism is introduced. If the autonomous driving decision meets expectations, multiple current perception information and multiple weights corresponding to multiple perception devices are saved as historical data. It can be understood that saving only the autonomous driving decisions that meet expectations ensures that the results of historical driving decisions in the historical data all meet expectations, thereby achieving a positive feedback effect.

[0052] Secondly, embodiments of this application provide an autonomous driving decision generation device, comprising:

[0053] The current environment information acquisition module is used to acquire multiple current environment information collected by multiple sensing devices at the current moment;

[0054] The judgment module is used to determine whether there is target historical data that matches the multiple current environmental information, wherein the target historical data includes multiple historical environmental information and multiple historical weights corresponding to the multiple sensing devices;

[0055] The virtual neurotransmitter concentration determination module is used to determine the virtual neurotransmitter concentration corresponding to each sensing device based on the historical weight, original weight, and reliability coefficient of each sensing device in the target historical data if there is target historical data that matches multiple current environmental information. The virtual neurotransmitter concentration is used to characterize the weight of the corresponding sensing device.

[0056] The autonomous driving decision generation module is used to generate autonomous driving decisions based on the weights of the multiple sensing devices.

[0057] Thirdly, embodiments of this application provide an electronic device, including:

[0058] processor;

[0059] Memory;

[0060] And a computer program, wherein the computer program is stored in the memory, the computer program including instructions that, when executed by the processor, cause the electronic device to perform the method described in any one of the first aspects.

[0061] Fourthly, embodiments of this application provide a computer-readable storage medium comprising a stored program, wherein, when the program is executed, it controls the device on which the computer-readable storage medium is located to perform the method described in any one of the first aspects.

[0062] It is understood that the autonomous driving decision generation device provided in the second aspect, the electronic device provided in the third aspect, and the computer-readable storage medium provided in the fourth aspect are all used to execute the autonomous driving decision generation method provided in this application. Therefore, the beneficial effects they can achieve can be referred to the beneficial effects in the corresponding methods, and will not be repeated here.

[0063] In this embodiment, based on research on the balance of neurotransmitters in the human brain, the regulation process of neurotransmitters in the human brain is simulated. Different concentrations of virtual neurotransmitters are used to adjust the weights of different sensing devices, and autonomous driving decisions are generated based on the weights of multiple sensing devices. It can be understood that by mimicking the flexibility of the human brain in information processing in complex environments, the robustness and reaction speed of the autonomous driving system can be guaranteed in changing environments and with an increasing variety of sensing devices. Attached Figure Description

[0064] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0065] Figure 1 A flowchart illustrating an autonomous driving decision generation method provided in an embodiment of this application;

[0066] Figure 2 A flowchart illustrating another autonomous driving decision generation method provided in this application embodiment;

[0067] Figure 3 This is a schematic diagram of the structure of an autonomous driving decision generation device provided in an embodiment of this application;

[0068] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0069] To better understand the technical solution of this application, the embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0070] It should be understood that the described embodiments are merely some, not all, of the embodiments in this application. All other embodiments obtained by those skilled in the art based on the embodiments in this application without inventive effort are within the scope of protection of this application.

[0071] The terminology used in the embodiments of this application is for the purpose of describing particular embodiments only and is not intended to be limiting of this application. The singular forms “a,” “the,” and “the” used in the embodiments of this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise.

[0072] It should be understood that the term "and / or" used in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0073] With the rapid development of the automotive industry, more and more cars are equipped with autonomous driving technology. Current autonomous driving systems integrate multiple sensors, such as cameras, radar, and lidar, to recognize and understand the vehicle's surroundings. After obtaining perception information from these sensors, the autonomous driving system needs to determine which information is more important and make autonomous driving decisions based on that importance.

[0074] In related technologies, decision-making models typically use preset rules or simple algorithms to weight the sensor-collected information, and the decision-making model needs to be trained frequently.

[0075] While this decision-making model is effective for pre-trained scenarios, it often lacks sufficient adaptability and flexibility in extreme or unknown situations. Furthermore, as the number of sensors on vehicles increases and the types of perceived information become more diverse, static rules often fail to make decisions quickly.

[0076] To address the aforementioned problems, this application provides an autonomous driving decision generation method, apparatus, device, and storage medium. Based on research into the balance of neurotransmitters in the human brain, it simulates the neurotransmitter regulation process in the human brain, adjusts the weights of different sensing devices by using virtual neurotransmitters of varying concentrations, and generates autonomous driving decisions based on the weights of multiple sensing devices. It is understood that by mimicking the flexibility of the human brain in information processing in complex environments, the robustness and reaction speed of the autonomous driving system can be ensured under changing environments and with an increasing variety of sensing devices. The following detailed description, in conjunction with the accompanying drawings and specific embodiments, further illustrates these points.

[0077] See Figure 1 This is a flowchart illustrating an autonomous driving decision generation method provided in an embodiment of this application. Figure 1 As shown, the method mainly includes the following steps.

[0078] Step S101: Acquire multiple current environmental information collected by multiple sensing devices at the current moment.

[0079] Specifically, the autonomous driving decision generation device acquires multiple pieces of current environmental information collected by multiple sensing devices at the current moment. Because there are many types and numbers of sensing devices, the current environmental information collected by these devices is extensive, including some information that is useless to the autonomous driving decision generation device. Therefore, in one possible implementation, after acquiring the multiple pieces of current environmental information collected by multiple sensing devices at the current moment, the autonomous driving decision generation device extracts key elements from the current environmental information, such as pedestrians, vehicles, and traffic signs.

[0080] Step S102: Determine the virtual neurotransmitter concentration corresponding to each sensing device based on multiple current environmental information.

[0081] Specifically, the autonomous driving decision-making generation device determines the virtual neurotransmitter concentration corresponding to each sensing device based on multiple current environmental information. The virtual neurotransmitter concentration for each sensing device represents its weight. In essence, the autonomous driving decision-making generation device can mimic the working mechanism of neurotransmitters in the human brain, dynamically adjusting the weights of sensing devices based on current environmental information.

[0082] In one possible implementation, the storage medium of the autonomous driving decision generation device stores a large amount of historical data. This historical data is collected by the autonomous driving decision generation device during daily operation and stored in the vehicle's local storage or a cloud server, forming a historical database. The historical data can include various types of information, such as sensor readings, historical driving decisions, historical weights corresponding to each sensor, historical environmental information, and time stamps. The historical data can be used to train and optimize the autonomous driving decision generation algorithm. In one possible implementation, timestamps are used to represent the specific time of the historical data. Of course, those skilled in the art can store other information in the historical data according to actual needs; this application does not impose specific limitations on this.

[0083] If the autonomous driving decision generation device dynamically adjusts the virtual neurotransmitter concentration of each sensing device based on the current environmental information each time it obtains such information, the computational power required to generate autonomous driving decisions would be substantial. Therefore, in this embodiment, when target historical data matching the current environmental information is stored in the historical data, the virtual neurotransmitter concentration of each sensing device is determined based on its historical weight within the target historical data. In one possible implementation, the virtual neurotransmitter concentration of each sensing device can be directly determined according to its historical weight, i.e., the virtual neurotransmitter concentration of each sensing device can be adjusted to be the same as its historical weight in the target historical data.

[0084] However, in this embodiment, the target historical data refers to historical data where the number of historical environmental information matching multiple current environmental information is greater than or equal to a preset threshold. Therefore, some historical environmental information in the target historical data may not match the current environmental information. If the virtual neurotransmitter concentration corresponding to each sensing device is only adjusted to be the same as the historical weight corresponding to each sensing device in the target historical data, it may lead to a large adjustment error in the virtual neurotransmitter concentration corresponding to some sensing devices. Therefore, in this embodiment, the virtual neurotransmitter concentration corresponding to each sensing device is determined based on the historical weight, original weight, and reliability coefficient corresponding to each sensing device in the target historical data. It can be understood that if the weight of each sensing device has been adjusted, referring to the original weight of each sensing device can ensure that the adjustment of the virtual neurotransmitter concentration corresponding to each sensing device is based on the previous adjustment, thereby obtaining a more accurate virtual neurotransmitter concentration. Among them, the reliability of the sensing device is used to characterize the priority of the environmental information collected by each sensing device in different scenarios. For example, when visibility is limited at night, radar and infrared sensors have higher reliability. Furthermore, the reliability coefficient of each sensing device can be pre-set before the autonomous driving decision generation device generates autonomous driving decisions. For example, if it is pre-set that radar and infrared sensors are more reliable when visibility is limited at night, then when the current environmental information indicates limited visibility, the weight of radar and infrared sensors is increased, and the weight of cameras is decreased, in order to better detect and analyze obstacle information in dark places and at a distance. Of course, the autonomous driving decision generation device can also summarize the reliability of each sensing device in different scenarios based on multiple historical data. For example, if historical data shows that radar is more accurate than cameras in rainy weather and high traffic density, then when the current environmental information indicates rainy weather and high traffic density, the weight of radar is increased.

[0085] Specifically, the historical weight, original weight, and reliability coefficient corresponding to each sensing device in the target historical data are multiplied to obtain the virtual neurotransmitter concentration for each sensing device. It can be understood that obtaining the virtual neurotransmitter concentration for each sensing device through multiplication allows for faster acquisition of the virtual neurotransmitter concentration, thereby improving the speed of generating autonomous driving decisions and ensuring the safety and real-time performance of autonomous driving.

[0086] In one possible implementation, if no target historical data exists that matches multiple current environmental information sets, the virtual neurotransmitter concentration for each sensing device is determined based on a default weight adjustment rule and the reliability of each sensing device. This can be understood as the autonomous driving decision-making generation device mimicking the working mechanism of neurotransmitters in the human brain, dynamically adjusting the weights of sensing devices based on current environmental information and the reliability of each device.

[0087] Specifically, the default weight adjustment rule includes increasing the concentration of virtual neurotransmitters from sensing devices related to any given current environmental information when such information meets preset conditions. For example, taking "obstacle location" as an example, the human brain typically releases adrenaline when dealing with emergencies, enabling individuals to react quickly. In this embodiment, when the distance between the obstacle and the vehicle is less than a first preset distance, the autonomous driving decision generation device begins to release virtual neurotransmitters similar to adrenaline. These virtual neurotransmitters are used to increase the weight of sensing devices (e.g., radar and cameras) that facilitate obstacle avoidance by the vehicle. In one possible implementation, the first preset distance is 15 meters. It is understood that the autonomous driving decision generation device intelligently adjusts the priorities of various sensing devices by analyzing the relative distance information between the obstacle and the vehicle in real time, mimicking the rapid response of the human brain in emergency situations.

[0088] Of course, the concentration of virtual neurotransmitters cannot increase indefinitely. Therefore, in one possible implementation, when any current environmental information meets a preset condition, the concentration of virtual neurotransmitters in the sensing devices related to that current environmental information is increased, and the concentration of virtual neurotransmitters is less than or equal to the maximum value of the virtual neurotransmitter concentration. For example, taking "obstacle location" as an example, when the distance between the obstacle and the vehicle is less than a second preset distance, the concentration of virtual neurotransmitters released by the autonomous driving decision generation device reaches its maximum value. At this time, the weight enhancement factor of the corresponding sensing device caused by the virtual neurotransmitters reaches its maximum value, thereby ensuring that obstacle avoidance behavior is executed quickly. In one possible implementation, the second preset distance is 5 meters. When the distance between the obstacle and the vehicle is less than 5 meters, the weight enhancement factor of the virtual neurotransmitters for the corresponding sensing device reaches 2.0.

[0089] In practical applications, the rate of increase of virtual neurotransmitter concentration needs to be preset. In this embodiment, when any current environmental information meets a preset condition, the virtual neurotransmitter concentration of the sensing device related to that current environmental information increases at a preset first rate. Alternatively, in one possible implementation, the first rate is a fixed value. When the current environmental information obtained by the autonomous driving decision generation device indicates that the current situation is not an emergency, the corresponding virtual neurotransmitter concentration can be set to increase at a fixed rate, thereby ensuring that the weights of the sensing devices increase stably, and thus ensuring that the autonomous driving decision generation device generates autonomous driving decisions that are more stable and consistent with the current situation. In another possible implementation, the first rate increases according to preset conditions. When the current environmental information obtained by the autonomous driving decision generation device indicates that the current situation is an emergency, the corresponding rate of increase of the virtual neurotransmitter concentration (the first rate) can be set to increase according to preset conditions. It can be understood that as the urgency of the current situation increases, the first rate also increases, and the corresponding rate of increase of the virtual neurotransmitter concentration increases faster, thereby enabling the weights of the corresponding sensing devices to reach the expected value more quickly, and thus generating the corresponding autonomous driving decision more rapidly.

[0090] In one possible implementation, if the virtual neurotransmitter concentration remains at a high level, it may cause the autonomous driving decision-making device to overreact in the absence of an emergency. Therefore, in this embodiment, when any current environmental information does not meet preset conditions, the concentration of virtual neurotransmitters in the sensing devices associated with that current environmental information is reduced. For example, taking "obstacle location" as an example, when the distance between the obstacle and the vehicle is greater than or equal to a first preset distance, the autonomous driving decision-making device begins to reduce the concentration of virtual neurotransmitters similar to adrenaline. In one possible implementation, the first preset distance is 15 meters. It is understood that through this dynamic virtual neurotransmitter recycling mechanism, the autonomous driving decision-making device can mimic the human brain's adaptability to constantly changing environments, thereby generating more flexible and more environmentally relevant autonomous driving decisions.

[0091] Of course, the concentration of virtual neurotransmitters cannot decrease indefinitely. Therefore, in one possible implementation, when any current environmental information does not meet preset conditions, the concentration of virtual neurotransmitters in the sensing devices related to that current environmental information is reduced, and the concentration of virtual neurotransmitters is greater than or equal to the initial concentration. Before the autonomous driving decision-making device releases virtual neurotransmitters, it records the initial concentration of each virtual neurotransmitter. This initial concentration is used to characterize the default weight of each sensing device when there is no specific environmental information stimulus. When the concentration of virtual neurotransmitters decreases, it is controlled to be greater than or equal to the initial concentration, thereby avoiding the weight of any sensing device being too low.

[0092] In practical applications, the rate of decrease in virtual neurotransmitter concentration needs to be preset. In this embodiment, when any current environmental information does not meet the preset conditions, the virtual neurotransmitter concentration of the sensing device related to any current environmental information is reduced at a preset second rate. Of course, in one possible implementation, the second rate is a fixed value. When the first rate corresponding to the virtual neurotransmitter is a fixed value, the virtual neurotransmitter concentration is set to decrease at a fixed rate, thereby ensuring a stable reduction in the weight of the sensing device. In another possible implementation, the second rate decreases according to preset conditions. When the first rate corresponding to the virtual neurotransmitter increases according to preset conditions, the second rate is set to decrease according to preset conditions. It can be understood that when the stimulus of the environmental information disappears, because the virtual neurotransmitter concentration is too high at this time, in order to quickly restore the virtual neurotransmitter concentration to its initial concentration, the second rate can be relatively large, and as the virtual neurotransmitter concentration decreases, the second rate can adaptively decrease.

[0093] Furthermore, the types of neurotransmitters released in the human brain differ depending on the situation. For example, in emergency situations (e.g., encountering a sudden obstacle), the brain rapidly releases hormones such as adrenaline to enhance attention and react quickly to the emergency. When performing complex tasks, the release of neurotransmitters such as dopamine may increase, thereby helping the brain improve cognitive function and decision-making ability. Following these brain activities, in this embodiment, the autonomous driving decision generation device can adjust the type and concentration of virtual neurotransmitters under specific circumstances, thereby adjusting the priority of the sensing devices. In one possible implementation, the autonomous driving decision generation device determines the current application scenario based on multiple current environmental information; and adjusts the concentration of virtual neurotransmitters in the sensing devices corresponding to the current application scenario. Specifically, when the autonomous driving decision generation device determines that the current application scenario is an emergency based on the current environmental information, it adjusts the concentration of virtual neurotransmitters in the sensing devices corresponding to that emergency; when the autonomous driving decision generation device determines that the current application scenario is a non-emergency based on the current environmental information, it adjusts the concentration of virtual neurotransmitters in the sensing devices corresponding to that non-emergency. For example, if the current environmental information includes a fast-moving object (such as a child running towards the street), the autonomous driving decision-making device can determine that the current application scenario is an emergency based on the current environmental information. In this case, the autonomous driving decision-making device will immediately increase the concentration of a virtual neurotransmitter similar to adrenaline, increasing the weight of perception devices (such as radar and cameras) that are more important for determining the position and speed of moving objects. If the current environmental information indicates that there is traffic congestion, the autonomous driving decision-making device will increase the concentration of a virtual neurotransmitter similar to serotonin, increasing the weight of perception devices used to monitor vehicle behavior (such as convoy driving). It is understood that serotonin helps process stress and anxiety in the human brain; therefore, a serotonin-like virtual neurotransmitter can help the autonomous driving decision-making device handle dense traffic situations more smoothly.

[0094] Step S103: Generate autonomous driving decisions based on the weights of multiple sensing devices.

[0095] Specifically, when the current environmental information detected by the sensing devices is static, an autonomous driving decision is generated directly based on the virtual neurotransmitter concentration corresponding to each sensing device; when the current environmental information detected by the sensing devices is dynamic, a weighted average of the current environmental information is calculated based on the weights of multiple sensing devices, and an autonomous driving decision is generated based on the weighted average of the current environmental information.

[0096] In addition, in order to improve the self-learning ability of the autonomous driving decision generation algorithm, a feedback-based learning mechanism is introduced in the embodiments of this application, which allows the autonomous driving decision generation device to optimize itself based on past experience.

[0097] Specifically, if the autonomous driving decision meets expectations, the current perception information and weights corresponding to multiple perception devices are saved as historical data. Conversely, if the autonomous driving decision does not meet expectations, the current perception information and weights corresponding to multiple perception devices are deleted or saved as data to be optimized. In this embodiment, after the autonomous driving decision generation device executes the autonomous driving decision, it determines whether the decision meets expectations based on the actual results (e.g., obstacle avoidance success rate and comfort assessment). It is understood that if the vehicle successfully avoids obstacles and the comfort level is high, the autonomous driving decision meets expectations; if the vehicle fails to avoid obstacles or the comfort level is low, the autonomous driving decision does not meet expectations.

[0098] In this embodiment, based on research on the balance of neurotransmitters in the human brain, the regulation process of neurotransmitters in the human brain is simulated. Different concentrations of virtual neurotransmitters are used to adjust the weights of different sensing devices, and autonomous driving decisions are generated based on the weights of multiple sensing devices. It can be understood that by mimicking the flexibility of the human brain in information processing in complex environments, the robustness and reaction speed of the autonomous driving system can be guaranteed in changing environments and with an increasing variety of sensing devices.

[0099] Corresponding to the above embodiments, this application also provides a flowchart of another autonomous driving decision generation method.

[0100] See Figure 2 This is a flowchart illustrating another autonomous driving decision generation method provided in this application embodiment. Figure 2 As shown, it mainly includes the following steps.

[0101] Step S201: Acquire multiple current environmental information collected by multiple sensing devices at the current moment.

[0102] Step S202: Determine whether there is target historical data that matches multiple current environmental information. If yes, proceed to step S203; otherwise, proceed to step S204.

[0103] Step S203: Multiply the historical weight, original weight and reliability coefficient of each sensing device in the target historical data to obtain the virtual neurotransmitter concentration of each sensing device.

[0104] Step S204: Determine the virtual neurotransmitter concentration for each sensing device based on the default weight adjustment rules and the reliability of each sensing device.

[0105] Step S205: Generate an autonomous driving decision based on the weights of multiple sensing devices.

[0106] Step S206: If the autonomous driving decision meets expectations, save the multiple current perception information and multiple weights corresponding to multiple perception devices as historical data.

[0107] For specific details regarding the embodiments of this application, please refer to the above. Figure 1 For the sake of brevity, the descriptions in the embodiments will not be repeated here.

[0108] Corresponding to the above embodiments, this application also provides an autonomous driving decision generation device.

[0109] See Figure 3 This is a schematic diagram of the structure of an autonomous driving decision generation device provided in an embodiment of this application. Figure 3 As shown, the autonomous driving decision generation device includes: a current environment information acquisition module 301, a judgment module 302, a virtual neurotransmitter concentration determination module 303, and an autonomous driving decision generation module 304. These components communicate through one or more buses. Those skilled in the art will understand that the server structure shown in the figure does not constitute a limitation on the embodiments of the present invention; it can be a bus topology, a star topology, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0110] Among them, the current environment information acquisition module 301 is used to acquire multiple current environment information collected by multiple sensing devices at the current moment;

[0111] The judgment module 302 is used to determine whether there is target historical data that matches the multiple current environmental information, wherein the target historical data includes multiple historical environmental information and multiple historical weights corresponding to the multiple sensing devices;

[0112] The virtual neurotransmitter concentration determination module 303 is used to determine the virtual neurotransmitter concentration corresponding to each sensing device based on the historical weight, original weight and reliability coefficient of each sensing device in the target historical data if there is target historical data that matches multiple current environmental information. The virtual neurotransmitter concentration is used to characterize the weight of the corresponding sensing device.

[0113] The autonomous driving decision generation module 304 is used to generate autonomous driving decisions based on the weights of the multiple sensing devices.

[0114] Corresponding to the above embodiments, this application also provides an electronic device.

[0115] See Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 4As shown, the electronic device 400 may include a processor 401, a memory 402, and a communication unit 403. These components communicate via one or more buses. Those skilled in the art will understand that the structure of the electronic device shown in the figure does not constitute a limitation on the embodiments of this application. It may be a bus topology or a star topology, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0116] The communication unit 403 is used to establish a communication channel, enabling the electronic device to communicate with other devices. It can receive user data sent by other devices or send user data to other devices.

[0117] The processor 401 serves as the control center of the electronic device, connecting various parts of the device via interfaces and lines. It executes software programs, instructions, and / or modules stored in the memory 402, and calls data stored in the memory to perform various functions and / or process data. The processor may be composed of integrated circuits (ICs), such as a single packaged IC or multiple packaged ICs with the same or different functions connected together. For example, the processor 401 may consist only of a central processing unit (CPU). In this embodiment, the CPU may have a single processing core or include multiple processing cores.

[0118] The memory 402 is used to store the execution instructions of the processor 401. The memory 402 can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk or optical disk.

[0119] When the execution instructions in memory 402 are executed by processor 401, the electronic device 400 is able to perform operations. Figure 1 Some or all of the steps in the illustrated embodiments.

[0120] In a specific implementation, this application embodiment also provides a computer storage medium, wherein the computer storage medium may store a program, and when the program is executed, it may include some or all of the steps of the simulation scene generation method provided in various embodiments of this application. The storage medium may be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0121] In this application embodiment, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent the existence of A alone, the simultaneous existence of A and B, or the existence of B alone. A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, and c can represent: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0122] Those skilled in the art will recognize that the units and algorithm steps described in the embodiments disclosed herein can be implemented using electronic hardware, computer software, or a combination of electronic hardware and software. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0123] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0124] In the several embodiments provided in this application, any function, if implemented as a software functional unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0125] The same or similar parts between the various embodiments in this specification can be referred to mutually. In particular, the device embodiments and terminal embodiments are basically similar to the method embodiments, so the description is relatively simple, and the relevant parts can be referred to the description in the method embodiments.

Claims

1. A method for generating autonomous driving decisions, characterized in that, include: Acquire multiple pieces of current environmental information collected by multiple sensing devices at the current moment; Determine whether there is target historical data that matches the multiple current environmental information, wherein the target historical data includes multiple historical environmental information and multiple historical weights corresponding to the multiple sensing devices; If there is target historical data that matches multiple current environmental information, then the virtual neurotransmitter concentration corresponding to each sensing device is determined according to the historical weight, original weight and reliability coefficient of each sensing device in the target historical data. The virtual neurotransmitter concentration is used to characterize the weight of the corresponding sensing device. Autonomous driving decisions are generated based on the weights of the multiple sensing devices.

2. The method according to claim 1, characterized in that, The step of determining the virtual neurotransmitter concentration for each sensing device based on the historical weight, original weight, and reliability coefficient of each sensing device in the target historical data includes: The historical weight, original weight, and reliability coefficient corresponding to each sensing device in the target historical data are multiplied together to obtain the virtual neurotransmitter concentration corresponding to each sensing device.

3. The method according to claim 1, characterized in that, The number of historical environmental information that matches multiple current environmental information in the target historical data is greater than or equal to a preset number threshold.

4. The method according to claim 1, characterized in that, The method further includes: If no target historical data matches the multiple current environmental information, the virtual neurotransmitter concentration corresponding to each of the sensing devices is determined according to the default weight adjustment rules and the reliability of each sensing device.

5. The method according to claim 4, characterized in that, The default weight adjustment rules include: When any of the current environmental information meets the preset conditions, the virtual neurotransmitter concentration of the sensing device associated with any of the current environmental information is increased.

6. The method according to claim 5, characterized in that, When any of the current environmental information meets a preset condition, the virtual neurotransmitter concentration of the sensing device associated with any of the current environmental information is increased, including: When any of the current environmental information meets the preset conditions, the virtual neurotransmitter concentration of the sensing device associated with any of the current environmental information is increased, and the virtual neurotransmitter concentration is less than or equal to the maximum value of the virtual neurotransmitter concentration.

7. The method according to claim 5, characterized in that, When any of the current environmental information meets a preset condition, the virtual neurotransmitter concentration of the sensing device associated with any of the current environmental information is increased, including: When any of the current environmental information meets the preset conditions, the concentration of the virtual neurotransmitter of the sensing device associated with any of the current environmental information is increased at a preset first rate.

8. The method according to claim 4, characterized in that, The default weight adjustment rules include: When any of the current environmental information does not meet the preset conditions, the concentration of the virtual neurotransmitter of the sensing device associated with any of the current environmental information is reduced.

9. The method according to claim 8, characterized in that, When any of the current environmental information does not meet the preset conditions, the virtual neurotransmitter concentration of the sensing device associated with any of the current environmental information is reduced, including: When any of the current environmental information does not meet the preset conditions, the concentration of the virtual neurotransmitter of the sensing device associated with any of the current environmental information is reduced, and the concentration of the virtual neurotransmitter is greater than or equal to the initial concentration of the virtual neurotransmitter.

10. The method according to claim 8, characterized in that, When any of the current environmental information does not meet the preset conditions, the virtual neurotransmitter concentration of the sensing device associated with any of the current environmental information is reduced, including: When any of the current environmental information does not meet the preset conditions, the concentration of the virtual neurotransmitter of the sensing device associated with any of the current environmental information is reduced at a preset second rate.

11. The method according to claim 4, characterized in that, The default weight adjustment rules include: Based on the various pieces of current environmental information, the current application scenario is determined; Adjust the virtual neurotransmitter concentration of the sensing device corresponding to the current application scenario.

12. The method according to claim 1, characterized in that, After generating an autonomous driving decision based on the weights of the multiple sensing devices, the method further includes: If the autonomous driving decision meets expectations, then the current perception information and weights corresponding to the multiple perception devices are saved as historical data.

13. An autonomous driving decision generation device, characterized in that, include: The current environment information acquisition module is used to acquire multiple current environment information collected by multiple sensing devices at the current moment; The judgment module is used to determine whether there is target historical data that matches the multiple current environmental information, wherein the target historical data includes multiple historical environmental information and multiple historical weights corresponding to the multiple sensing devices; The virtual neurotransmitter concentration determination module is used to determine the virtual neurotransmitter concentration corresponding to each sensing device based on the historical weight, original weight, and reliability coefficient of each sensing device in the target historical data if there is target historical data that matches multiple current environmental information. The virtual neurotransmitter concentration is used to characterize the weight of the corresponding sensing device. The autonomous driving decision generation module is used to generate autonomous driving decisions based on the weights of the multiple sensing devices.

14. An electronic device, characterized in that, include: processor; Memory; And a computer program, wherein the computer program is stored in the memory, the computer program including instructions that, when executed by the processor, cause the electronic device to perform the method of any one of claims 1 to 12.

15. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes a stored program, wherein, when the program is executed, it controls the device on which the computer-readable storage medium is located to perform the method of any one of claims 1 to 12.

Images