Intelligent driving method and device

Abstract

An intelligent driving method and device. The method comprises: acquiring a laser point cloud; and then when it is determined, on the basis of the laser point cloud, that there is a suspected obstacle on a road where a vehicle having an intelligent driving capability is located, and the vehicle is at a first position, on the basis of first information and second information, determining whether there is an obstacle on the road where the vehicle is located, wherein the first position is a position where the vehicle historically made erroneous braking on the basis of the laser point cloud, the first information is used for indicating the road condition of the road where the vehicle is located at the current moment, and the second information is used for indicating the road condition of the road where the vehicle is located when the vehicle historically made erroneous braking at the first position; and on the basis of whether there is an obstacle on the road where the vehicle is located, controlling the vehicle to perform intelligent driving, rather than directly performing braking when it is determined that there is a suspected obstacle on the current road, so that erroneous braking of the vehicle during intelligent driving is avoided, thereby improving the intelligent driving safety and the intelligent driving experience of a user.

Description

A smart driving method and device Technical Field

[0001] This application relates to the field of intelligent driving technology, and in particular to an intelligent driving method and device. Background Technology

[0002] Intelligent driving (referred to as intelligent driving) refers to the use of technologies such as artificial intelligence, sensors, and communication and positioning to help people drive vehicles. It has the ability to perceive the surrounding environment, make decisions, and perform driving tasks.

[0003] During intelligent driving, laser point clouds contain point cloud noise due to the material properties of the object being detected (such as highly reflective materials), weather, and environmental factors (such as water mist and dust). This point cloud noise may be mistaken for the presence of obstacles (or suspected obstacles), which may cause vehicles with intelligent driving capabilities to brake incorrectly, affecting intelligent driving safety and the user's intelligent driving experience.

[0004] Summary of the Invention

[0005] This application provides an intelligent driving method and device to prevent vehicles from braking accidentally during intelligent driving, thereby improving intelligent driving safety and the user's intelligent driving experience.

[0006] Firstly, this application provides an intelligent driving method, which can be applied to a device, a module of the device (such as a processor, processing unit, chip, circuit, etc.), or a system corresponding to the device. The device can be a terminal device (such as an in-vehicle device). Based on this, the method includes: acquiring laser point clouds; then, based on the laser point clouds, determining that a suspected obstacle exists on the road where a vehicle with intelligent driving capabilities is located, and when the vehicle is at a first position, determining whether an obstacle exists on the road where the vehicle is located based on first information and second information; wherein, the first position is the position where the vehicle historically mis-brakes based on the laser point clouds, the first information is used to indicate the road conditions on the road where the vehicle is located at the current moment, and the second information is used to indicate the road conditions on the road where the vehicle was located when it historically mis-brakes at the first position. Afterwards, based on whether an obstacle exists on the road where the vehicle is located, controlling the vehicle to perform intelligent driving.

[0007] In the above method, the current moment refers to the moment when the vehicle is in the first position and a suspected obstacle is determined to exist on the road where the vehicle is located. That is, the first information indicates the road conditions when the vehicle is currently in the first position. The second information indicates the road conditions when, historically, an obstacle was mistakenly identified as existing on the road where the vehicle was located in the first position, but in reality, there was no obstacle on the road corresponding to the first position. In other words, the second information indicates that there is no obstacle on the road corresponding to the first position. It can be understood that both the first and second information are associated with the first position.

[0008] Based on this, at the same location (i.e., the first location), when it is determined from the laser point cloud that there is a suspected obstacle on the current road, in order to avoid the vehicle braking accidentally again, the road conditions of the current road (i.e., the first information) are identified through the second information to determine whether there is an obstacle on the current road, and then the vehicle is controlled to perform intelligent driving, instead of braking directly when it is determined that there is a suspected obstacle on the current road. This avoids the vehicle braking accidentally during intelligent driving, improves intelligent driving safety and the user's intelligent driving experience.

[0009] In one possible implementation, determining the presence of a suspected obstacle on the road where a vehicle with autonomous driving capabilities is located based on laser point clouds includes: projecting the laser point clouds onto an image corresponding to the current road; if it is determined that a high-reflectivity point in the laser point cloud is projected onto the image but not onto the current road, and a low-reflectivity point in the laser point cloud is projected onto the image but onto the current road, then it is determined that a suspected obstacle exists on the road where the vehicle is located; wherein, a high-reflectivity point is a point in the laser point cloud with an intensity greater than or equal to a first threshold, and a low-reflectivity point is a point in the laser point cloud with an intensity less than or equal to a second threshold, wherein the second threshold is less than the first threshold.

[0010] In the above implementation, it can be understood that the vehicle's accidental braking due to a suspected obstacle is caused by the projection of low-reflectivity points in the laser point cloud onto the image corresponding to the current road, which is located on the current road. In other words, the factor causing the accidental braking lies in the low-reflectivity points in the laser point cloud, not the high-reflectivity points, where the low-reflectivity points are considered point cloud noise. Therefore, if it is determined that the projection of a high-reflectivity point in the laser point cloud onto the image is located on the current road, it indicates that an obstacle exists on the road where the vehicle is located. At this point, the vehicle can be braked to ensure the safety of intelligent driving.

[0011] In one possible implementation, determining whether there is an obstacle on the road where the vehicle is located based on the first information and the second information includes: determining that there is no obstacle on the road where the vehicle is located when the feature similarity between the first information and the second information is greater than or equal to a third threshold; conversely, determining that there is an obstacle on the road where the vehicle is located when the feature similarity between the first information and the second information is less than the third threshold.

[0012] In the above implementation, the first information and the second information indicate the road conditions of the vehicle at the same location (i.e., the first position) but at different times (the first information corresponds to the current time, and the second information corresponds to the time when the mis-braking occurred historically). Therefore, it can be understood that if the feature similarity between the first and second information is greater than or equal to the third threshold, it means that the first and second information have similar features, or even identical features. This is because the second information indicates the road conditions when the vehicle historically mis-braked at the first position (i.e., the second information indicates that the road conditions were free of obstacles). Therefore, it can be determined that the road where the vehicle is located, as indicated by the first information, is free of obstacles, thus enabling the detection of obstacles on the road where the vehicle is located, preventing the vehicle from mis-braking during intelligent driving, and improving intelligent driving safety and the user's intelligent driving experience. Conversely, if the feature similarity between the first and second information is less than the third threshold, it means that the features of the first and second information are dissimilar. That is, it can be determined that the road where the vehicle is located, as indicated by the first information, is free of obstacles. In this case, the vehicle can be braked to ensure the safety of intelligent driving.

[0013] In one possible implementation, the first information and the second information are image information, with the first information corresponding to a first image and the second information corresponding to a second image. To this end, the feature similarity between the first information and the second information is determined, including: determining the feature similarity between the low-reflection projection region in the laser point cloud located on the road where the vehicle is located in the first image and the corresponding region in the second image.

[0014] In one possible implementation, the first information and the second information are image information, with the first information corresponding to a first image and the second information corresponding to a second image. To this end, determining the feature similarity between the first information and the second information includes: determining the feature similarity between the region corresponding to the road where the vehicle is located in the first image and the region corresponding to the road where the vehicle is located in the second image.

[0015] In the above implementation, considering that the vehicle braked incorrectly because there was a suspected obstacle on the road where the vehicle was located, the feature similarity is calculated only for the area of ​​the road where the vehicle is located in the image. This can ensure the accuracy of detection and reduce the amount of data required for feature similarity calculation, thus reducing the computational latency.

[0016] In one possible implementation, after determining that there are no obstacles on the road where the vehicle is located, the method further includes: using the first information as the second information.

[0017] In the above implementation, the first information is equivalent to the latest detection result for the first position, which in turn updates the second information to improve the accuracy of subsequent detections.

[0018] In one possible implementation, controlling the vehicle to perform intelligent driving based on whether there is an obstacle on the road includes: controlling the vehicle to continue driving and / or issuing an alarm when it is determined that there is no obstacle on the road; and controlling the vehicle to brake when it is determined that there is an obstacle on the road.

[0019] In the above implementation, a warning is issued when it is determined that there are no obstacles on the current road to prompt the user to drive cautiously, thereby improving intelligent driving safety. Furthermore, when it is determined that there are no obstacles on the current road, direct braking is not initiated to avoid accidental braking during intelligent driving, which could affect intelligent driving safety and the user's intelligent driving experience. When an obstacle is determined to exist on the road, the vehicle's brakes are controlled to ensure intelligent driving safety.

[0020] Secondly, an apparatus is provided, comprising an acquisition module and a processing module; wherein the acquisition module is used to acquire laser point clouds; the processing module is used to determine whether an obstacle exists on the road where a vehicle with intelligent driving capabilities is located, based on first information and second information, when it is determined from the laser point clouds that a suspected obstacle exists on the road and the vehicle is located at a first position; wherein the first position is a position where the vehicle historically made an accidental braking based on the laser point clouds, the first information is used to indicate the road conditions on the road where the vehicle is located at the current moment, and the second information is used to indicate the road conditions on the road where the vehicle was located when the vehicle historically made an accidental braking at the first position; the processing module is further used to control the vehicle to perform intelligent driving based on whether an obstacle exists on the road.

[0021] In one possible implementation, the processing module is specifically used to: project the laser point cloud onto an image corresponding to the road where the vehicle is located; if it is determined that the projection of a high-reflectivity point in the laser point cloud onto the image is not located on the road where the vehicle is located, and the projection of a low-reflectivity point in the laser point cloud onto the image is located on the road where the vehicle is located, then it is determined that there is a suspected obstacle on the road where the vehicle is located; wherein, the high-reflectivity point is a point in the laser point cloud with an intensity greater than or equal to a first threshold, and the low-reflectivity point is a point in the laser point cloud with an intensity less than or equal to a second threshold, the second threshold being less than the first threshold.

[0022] In one possible implementation, the processing module is specifically used to: determine that there are no obstacles on the road where the vehicle is located when the feature similarity between the first information and the second information is greater than or equal to a third threshold; and determine that there are obstacles on the road where the vehicle is located when the feature similarity between the first information and the second information is less than the third threshold.

[0023] In one possible implementation, the first information and the second information are image information, the first information corresponds to a first image, and the second information corresponds to a second image; the processing module is specifically used to: determine the feature similarity between the low reflection projection region in the laser point cloud located on the road where the vehicle is located in the first image and the corresponding region in the second image.

[0024] In one possible implementation, the first information and the second information are image information, the first information corresponds to a first image, and the second information corresponds to a second image; the processing module is specifically used to: determine the feature similarity between the region corresponding to the road where the vehicle is located in the first image and the region corresponding to the road where the vehicle is located in the second image.

[0025] In one possible implementation, after determining that there are no obstacles on the road where the vehicle is located, the processing module is further configured to: use the first information as the second information.

[0026] In one possible implementation, the processing module is specifically used to: control the vehicle to continue driving and / or issue an alarm when it is determined that there is no obstacle on the road; and control the vehicle to brake when it is determined that there is an obstacle on the road.

[0027] Thirdly, a vehicle is provided, comprising an onboard detection device and means for performing the method described in the first aspect or any possible implementation thereof. Optionally, the vehicle further comprises an onboard image acquisition device and a positioning device. The laser device is used to generate a laser point cloud, the image acquisition device is used to acquire visual images, and the positioning device is used to determine the vehicle's position.

[0028] Fourthly, a computer device is provided, including a processor, a memory, a communication interface, and a bus. The processor, memory, and communication interface are connected via the bus and communicate with each other. The memory stores computer execution instructions. When the computer device is running, the processor executes the computer execution instructions in the memory to perform the operation steps of the method described in the first aspect or any possible implementation of the first aspect using the hardware resources in the computer device.

[0029] Fifthly, an apparatus is provided, comprising one or more processors and one or more memories; the one or more memories storing one or more computer programs, the one or more computer programs including instructions that, when executed by the one or more processors, cause the apparatus to perform operational steps of the method of the first aspect or any possible implementation thereof.

[0030] In a sixth aspect, a chip system is provided, the chip system comprising at least one chip and a memory, the at least one chip being configured to read and execute a program stored in the memory to implement the operational steps of the method described in the first aspect or any possible implementation thereof.

[0031] A seventh aspect provides a non-volatile computer-readable storage medium comprising a program that, when executed on a device, causes the device to perform the operational steps of the method described in the first aspect or any possible implementation thereof.

[0032] Eighthly, a computer program product is provided, which, when run on a device, causes the device to perform the operational steps of the method described in the first aspect or any possible implementation thereof.

[0033] Based on the implementation methods provided in the above aspects, this application can be further combined to provide more implementation methods. Attached Figure Description

[0034] Figure 1 is a schematic diagram of an intelligent driving scenario provided by an embodiment of this application;

[0035] Figure 2 is a schematic diagram of laser point cloud expansion provided in an embodiment of this application;

[0036] Figure 3 is a flowchart illustrating an intelligent driving method provided in an embodiment of this application;

[0037] Figure 4 is a schematic diagram of a detection area provided in an embodiment of this application;

[0038] Figure 5 is a schematic diagram of the structure of a device provided in an embodiment of this application;

[0039] Figure 6 is a schematic diagram of the structure of a device provided in an embodiment of this application. Detailed Implementation

[0040] In recent years, with the development of artificial intelligence and sensor technology, autonomous driving has gradually become a major driving force for innovation in the automotive industry. Autonomous driving (which can be abbreviated as intelligent driving) refers to the use of artificial intelligence, sensors, and communication and positioning technologies to assist humans in driving vehicles, enabling them to perceive their surroundings, make decisions, and execute driving tasks.

[0041] In relevant technical solutions, vehicles with intelligent driving capabilities (hereinafter referred to as intelligent driving vehicles) include multiple on-board devices (such as on-board detection devices, on-board image acquisition devices, and on-board computing devices). The on-board detection devices include lidar and other related sensors and processors. These devices can generate point clouds (or laser point clouds, point cloud data) based on lidar, and can also acquire information about obstacles around the vehicle based on the emission and reflection of millimeter-wave radar and ultrasonic radar. The on-board image acquisition devices include cameras and other related processors, and can acquire visual images. The on-board computing devices include processors, memory, and other related components. These devices are equipped with an operating system and applications. Based on this, the on-board computing devices can perceive the surrounding environment of the intelligent driving vehicle based on this information (such as laser point clouds and visual images), determine whether there are obstacles, pedestrians, or vehicles around the intelligent driving vehicle, and thus decide whether the intelligent driving vehicle can avoid obstacles and various risks on the road, helping people drive the vehicle more safely and conveniently, achieving intelligent driving.

[0042] In real-world scenarios, laser point clouds may exhibit high reflectivity, or point cloud noise, due to factors such as the material properties of the object being detected, weather, and environmental conditions (e.g., water mist and dust). High reflectivity typically manifests as high-reflectivity "ghosting" or "expansion." For example, if the object being detected is made of a highly reflective material, it will cause the laser point cloud to expand, creating point cloud noise around the object. High-reflectivity materials refer to materials with high reflectivity. Taking high-reflectivity "expansion" as an example, point cloud noise refers to the point cloud surrounding the object, not reflecting the actual point cloud of the object. In other words, point cloud noise can be understood as the "expanded" portion of the object. This "expanded" portion is not located at the actual location of the object, which may cause the onboard computing equipment of an autonomous vehicle to mistakenly perceive an obstacle at the location of the "expanded" portion. For example, if the target being detected is a road stud located on the road, its actual height (e.g., less than the radius of a wheel) generally does not affect vehicle movement. However, when the laser point cloud corresponding to the stud expands, the onboard computing device may perceive the expanded portion as indicating that the stud's height (e.g., greater than the wheel radius) affects vehicle movement. In this case, the onboard computing device interprets the location of the expanded portion as an obstacle. Based on this, when the onboard computing device determines the presence of an obstacle, it decides that the autonomous vehicle should brake due to this obstacle (this braking can be understood as a false braking). Therefore, point cloud noise may be a cause of false braking in autonomous driving.

[0043] Figure 1 is a schematic diagram of an intelligent driving scenario provided by an embodiment of this application. The vehicle in Figure 1 has intelligent driving capabilities and is in intelligent driving mode. A road sign is set on the right side of the road. Assuming that the road sign is made of a highly reflective material, the laser point cloud generated by the laser device based on the vehicle is shown in Figure 2. Figure 2 is a schematic diagram of laser point cloud expansion provided by an embodiment of this application. Referring to Figure 2, the area corresponding to the road sign is a high-reflectivity point area, the surrounding area corresponding to the road sign is a low-reflectivity point area, and the point cloud on both sides of the road is normal point cloud data. Among them, the low-reflectivity points belong to point cloud noise.

[0044] Based on Figures 1 and 2, it can be seen that the low-reflection area is located on the road (or the road where the vehicle is located). In this case, the vehicle's onboard computing equipment determines the presence of an obstacle on the road based on the laser point cloud and then decides to avoid it. For example, if the vehicle is close to the detected obstacle, it brakes; if the vehicle is far from the detected obstacle, it maneuvers around it. However, there is actually no obstacle on the road, meaning the obstacle avoidance maneuver is unnecessary. Furthermore, due to inertia, such braking (even sudden braking) and turning can cause discomfort for the user and even pose a safety hazard, affecting intelligent driving safety and the user's intelligent driving experience.

[0045] Therefore, this application provides an intelligent driving method to determine whether the laser detection result at the current moment is accurate based at least on historical false braking data, so as to avoid false braking caused by point cloud noise during intelligent driving, thereby improving intelligent driving safety and user intelligent driving experience.

[0046] The following explanations are provided for some of the technologies and terms involved in the embodiments of this application.

[0047] (1) First position

[0048] The first location refers to the location where a vehicle historically mis-braked during intelligent driving based on laser point clouds. Referring to the descriptions in Figures 1 and 2 above, factors contributing to mis-braking include laser point cloud dilation. For example, scenarios involving laser point cloud dilation may include road signs made of highly reflective materials, road studs made of highly reflective materials, and so on.

[0049] In one possible implementation, historically, during intelligent driving, when the vehicle determines, based on laser point clouds, that there is a suspected obstacle on the road where the vehicle is located, it brakes. The braking acceleration is determined to be greater than or equal to a threshold (e.g., braking acceleration -4 m / s²). 2 The threshold is 3m / s 2If, at this point, the absolute value of the vehicle's braking acceleration is determined to be 4 (greater than the threshold 3), and the trajectory of the vehicle after this braking includes the suspected obstacle (or the location of the suspected obstacle is on the trajectory of the vehicle after this emergency braking), then the starting braking position (not the stopping position) is taken as the first position. This threshold can be a preset value based on experience. It can be understood that braking acceleration is generally negative; when the absolute value of the vehicle's braking acceleration is greater than the threshold, this braking can be considered an emergency braking. If the location of the suspected obstacle is on the trajectory of the vehicle after this emergency braking, it means that the vehicle can pass over the location of the suspected obstacle, and this braking may be a false braking caused by point cloud noise.

[0050] In one possible implementation, after determining the first position, second information corresponding to the first position is recorded; that is, the second information is associated with the first position. The second information indicates the road conditions when the vehicle was historically misjudged and braked at the first position. This can be understood as the road conditions at that time indicating there were no obstacles on the road where the vehicle was located. In other words, the second information indicates the road conditions when, historically, the vehicle was misjudged to be on a road with obstacles at the first position, but in reality, there were no obstacles on the road corresponding to the first position.

[0051] Optionally, the second information can be image information (including but not limited to visual images acquired by an image acquisition device, and information corresponding to a high-precision map). For example, the second information corresponds to a second image, which is a visual image acquired by the vehicle-mounted image acquisition device at the first location (as shown in Figure 1). It can be understood that the second image includes the road where the vehicle is located; for example, the second image includes the road ahead of the vehicle as shown in Figure 1. Therefore, the second image (i.e., the second information) can indicate that there were no obstacles on the road where the vehicle was located at the first location during the vehicle's intelligent driving process in the past. For example, the second image is the image shown in Figure 1, which can indicate that there were no obstacles on the road ahead of the vehicle at the first location.

[0052] Optionally, the second information can also be point cloud information, such as the laser point cloud of the vehicle at its first position in history (as shown in Figure 2). It can be understood that the vehicle's onboard computing device may have misjudged the presence of an obstacle on the road where the vehicle is located based on the laser point cloud at the first position, but in reality there is no obstacle on the road. Therefore, the laser point cloud features corresponding to the first position can be used to indicate that there is no obstacle on the road.

[0053] (2) First Information

[0054] The first information is used to indicate the road conditions of the road where the vehicle is located at the current moment. The current moment refers to the time when the vehicle is in its first position and a suspected obstacle is identified on the road where the vehicle is located; that is, the first information is associated with the first position. In other words, the first information indicates the road conditions of the road where the vehicle is located when it is in its first position. Therefore, it can be understood that the first information is generated when the vehicle is in its first position and a suspected obstacle is identified on the road where the vehicle is located based on the current laser point cloud. The first position can be determined based on the vehicle's onboard positioning device.

[0055] Optionally, the first information can be image information or point cloud information. For example, the first information corresponds to a first image, which is a visual image captured by the vehicle's camera at the current moment, or the first information is the laser point cloud of the vehicle at the current moment. It can be understood that the first information can reflect the road conditions in front of the vehicle (in the vehicle's direction of travel) at the current moment.

[0056] In summary, the difference between the first and second information lies in the following: the first information indicates the road conditions at the current moment (i.e., when the current laser point cloud indicates a suspected obstacle on the road where the vehicle is located, and the vehicle is in the first position), and the road conditions on the road where the vehicle is located at the first position. The second information refers to the road conditions at a historical moment (i.e., when it is determined that the vehicle braked incorrectly), and the road conditions on the road where the vehicle was located at the first position. In other words, the first and second information correspond to different times; the first information corresponds to the current moment, and the second information corresponds to a historical moment.

[0057] One possible implementation involves an onboard detection device generating a laser point cloud in real time, and an onboard image acquisition device acquiring visual images in real time. The onboard computing device projects the generated laser point cloud onto the acquired visual image. If it is determined that the projection of a high-reflectivity point in the laser point cloud onto the visual image is not located on the road where the vehicle is located, and the projection of a low-reflectivity point in the laser point cloud onto the visual image is located on the road where the vehicle is located, then it is determined that a suspected obstacle exists on the road where the vehicle is located. Here, high-reflectivity points are points in the laser point cloud with an intensity greater than or equal to a first threshold, and low-reflectivity points are points in the laser point cloud with an intensity less than or equal to a second threshold, where the second threshold is less than the first threshold. The point cloud intensity range is generally 0-255, and the first and second thresholds can be preset values, such as the first threshold being 200 and the second threshold being 100.

[0058] As can be understood, laser point cloud projection onto an image refers to projecting point cloud data in three-dimensional space onto the imaging plane of a camera to generate a two-dimensional image or perform other related calculations and analyses. For example, embodiments of this application can project laser point clouds onto an image using the following method: Obtain camera intrinsic parameters (such as focal length and principal point coordinates). Then, transform the laser point cloud coordinates from the laser coordinate system (three-dimensional) to the camera coordinate system (three-dimensional) to obtain three-dimensional points in the camera coordinate system. Next, use the camera intrinsic parameters to convert the three-dimensional points in the camera coordinate system into normalized image coordinates. Finally, map the normalized image coordinates to actual pixel coordinates, thereby realizing the projection of the laser point cloud onto the image.

[0059] One possible implementation involves the vehicle-mounted detection device generating a laser point cloud in real time. When the vehicle-mounted computing device identifies a candidate point cloud region in the laser point cloud with a number of high reflective points greater than or equal to a threshold for the number of high reflective points, it projects the point cloud in the candidate point cloud region onto a visual image acquired by the vehicle-mounted image acquisition device. If it is determined that the projection of the high reflective points in the candidate point cloud region onto the visual image is not located on the road where the vehicle is located, and the projection of the low reflective points in the candidate point cloud region onto the visual image is located on the road where the vehicle is located, then it is determined that there is a suspected obstacle on the road where the vehicle is located.

[0060] In the above implementation, projecting only the point cloud in the candidate point cloud region onto the visual image can reduce the computational load and latency. Considering that the vehicle's erroneous braking is caused by the presence of a suspected obstacle on the road where the vehicle is located (i.e., the projection of the low-reflectivity point on the visual image lies on the road where the vehicle is located), the presence of a suspected obstacle on the road where the vehicle is located is only determined when the projection of the low-reflectivity point on the visual image lies on the road where the vehicle is located, thus ensuring detection accuracy.

[0061] Optionally, considering the limited number of high-reflectivity points due to weather and environmental factors (such as water mist and dust), a threshold for the number of high-reflectivity points can be set to reduce subsequent computational load and save computing resources. For example, if the onboard computing device determines that the number of high-reflectivity points projected onto the road where the vehicle is located is greater than or equal to the threshold, then the projection of the high-reflectivity points onto the image is determined to be on the road where the vehicle is located; otherwise, the projection of the high-reflectivity points onto the visual image is determined not to be on the road where the vehicle is located. The threshold can be a preset value based on experience, such as 5. Similarly, if the onboard computing device determines that the number of low-reflectivity points projected onto the road where the vehicle is located is greater than or equal to the threshold, then the projection of the low-reflectivity points onto the visual image is determined to be on the road where the vehicle is located; otherwise, the projection of the low-reflectivity points onto the visual image is determined not to be on the road where the vehicle is located.

[0062] Based on the above description, Figure 3 is a flowchart illustrating an intelligent driving method provided in an embodiment of this application. This process is applied to a device, or a module of the device (such as a processor, processing unit, chip, or circuit), or a system corresponding to the device. The device can be a terminal device (such as an in-vehicle device). This process can also be applied to vehicles capable of intelligent driving. For example, the vehicle may be equipped with a driving system, such as an Advanced Driving System (ADS). Optionally, the ADS level installed in the vehicle can be at least one of the following: Level 2 (combined driving assistance), Level 3 (conditional automated driving), Level 4 (highly automated driving), or Level 5 (fully automated driving).

[0063] As shown in Figure 3, the method may include the following steps:

[0064] Step 301: Obtain laser point cloud.

[0065] In this step, the laser point cloud can be generated in real time by the vehicle-mounted detection device.

[0066] Step 302: Determine whether there are any suspected obstacles on the road where the vehicle with intelligent driving capability is located based on the laser point cloud. If so, proceed to step 303; otherwise, return to step 301 above.

[0067] In this step, the method for identifying suspected obstacles can be found in the description of the first piece of information above.

[0068] Step 303: Determine if the vehicle is in the first position. If yes, proceed to step 304; otherwise, return to step 301 above. The first position is the historical location where the vehicle mistakenly braked based on laser point clouds.

[0069] In this step, the description of the first position can be found in the above introduction.

[0070] One possible implementation is to determine the vehicle's location based on positioning technology. If the vehicle's location is within a preset range corresponding to the first location, then the vehicle is determined to be in the first location; otherwise, the vehicle is determined not to be in the first location. The preset range can be a value preset based on experience, such as a circular area with a radius of 3 meters centered on the first location.

[0071] Step 304: Determine whether there are obstacles on the road where the vehicle is located based on the first information and the second information, then proceed to step 305; otherwise, apply the brakes. The first information indicates the road conditions at the current moment, and the second information indicates the road conditions when the vehicle mistakenly braked at the first position in the past.

[0072] In one possible implementation, before executing step 304, it can be determined whether there are obstacles on the road where the vehicle is located based on the visual image acquired at the current moment. If the visual image identifies an obstacle on the road, step 304 is not required, and the vehicle can be braked directly, ensuring timely braking, the safety of intelligent driving, and avoiding subsequent invalid calculations. Conversely, if the visual image identifies no obstacle on the road, step 304 is executed.

[0073] In one possible implementation, if the feature similarity between the first and second pieces of information is greater than or equal to a third threshold, it is determined that there are no obstacles on the road where the vehicle is located; conversely, if the feature similarity between the first and second pieces of information is less than the third threshold, it is determined that there are obstacles on the road where the vehicle is located. The third threshold can be a preset value, such as 95%. It can be understood that the larger the third threshold, the greater the detection accuracy and the higher the safety of intelligent driving.

[0074] In the above implementation method, the feature similarity between the first information and the second information can be determined by at least one of the following methods.

[0075] Method 1: The first information and the second information are image information (as shown in Figure 1), where the first information corresponds to the first image and the second information corresponds to the second image. Based on this, the image similarity between the low-reflectivity projection region (or simply the detection region) in the laser point cloud located on the road where the vehicle is located in the first image and the corresponding region in the second image (i.e., the region in the second image that is at the same position as the detection region) is calculated, and this image similarity is used as the feature similarity between the first information and the second information. Figure 4 is a schematic diagram of a detection region provided in an embodiment of this application. As shown in Figure 4, the detection region is located on the road where the vehicle is located and is associated with the low-reflectivity points in the laser point cloud. Optionally, image similarity can be implemented based on grayscale matching, template matching, or deep learning algorithms.

[0076] Method 2: The first and second information are image information, with the first information corresponding to the first image and the second information corresponding to the second image. Based on this, the image similarity between the region corresponding to the road where the vehicle is located in the first image (e.g., the region of the road where the vehicle is located in the image) and the region corresponding to the road where the vehicle is located in the second image is calculated, and this image similarity is used as the feature similarity between the first and second information.

[0077] Method 3: The first and second information are image information, with the first information corresponding to the first image and the second information corresponding to the second image. Based on this, the image similarity between the first and second images is directly calculated, and this image similarity is used as the feature similarity between the first and second information.

[0078] Method 4: The first and second information are point cloud information. The first information corresponds to the first laser point cloud (i.e., the laser point cloud at the current moment), and the second information corresponds to the second laser point cloud (i.e., the laser point cloud when the vehicle was in its first position historically). Based on this, the point cloud feature similarity between the first and second laser point clouds can be directly calculated, and this point cloud feature similarity can be used as the feature similarity between the first and second information. Optionally, the point cloud feature similarity can be implemented based on the Euclidean distance algorithm or the Hausdorff distance algorithm.

[0079] Of the methods 1-4 described above, methods 1 and 2 calculate feature similarity only for the area (or partial information) of the road where the vehicle is located in the image. This ensures detection accuracy while reducing the amount of data required for feature similarity calculation and thus reducing computational latency. Methods 3 and 4 calculate feature similarity using all information, which further enhances detection accuracy.

[0080] Step 305: Issue an alarm when it is determined that there are no obstacles on the road where the vehicle is located.

[0081] This process can be applied to scenarios where the user is in the vehicle, or to scenarios where the user is not in the vehicle (i.e., remote intelligent driving based on terminal devices). Therefore, the alarm methods may include, but are not limited to, at least one of the following: issuing an audible alarm through the in-vehicle device; sending an instruction message to the terminal device, which indicates that a suspected obstacle has been detected on the road where the vehicle is located based on laser detection, thereby alerting the user that the laser detection is being interfered with and urging them to drive with caution.

[0082] In one possible implementation, after determining that there are no obstacles on the road where the vehicle is located, the first information is used as the second information. This is equivalent to updating the second information to improve the accuracy of subsequent detection.

[0083] In summary, since the first and second information indicate the same location (i.e., the first location) but different road conditions at different times, when the vehicle is located at the location where accidental braking occurred (i.e., the first location), if the laser point cloud indicates a suspected obstacle on the road where the vehicle is located, the road conditions of the road at the current time (i.e., the first information) can be used to identify the road conditions of the road where the vehicle is located at the current time, ensuring the accuracy of the detection. Furthermore, if it is determined that there is no obstacle on the road where the vehicle is located, a warning is issued to prompt the user to drive cautiously rather than brake directly, thereby preventing accidental braking during intelligent driving and improving intelligent driving safety and the user's intelligent driving experience.

[0084] The intelligent driving method provided according to the embodiments of this application has been described in detail above based on Figure 3. The apparatus for performing the above method according to the embodiments of this application will be described below with reference to Figures 5 and 6.

[0085] Figure 5 is a schematic diagram of an intelligent driving device provided in an embodiment of this application. This device 500 can be used to implement the above-described intelligent driving method, and therefore can also achieve the beneficial effects of the above-described method embodiments.

[0086] As shown in Figure 5, the device 500 includes an acquisition module 510 and a processing module 520. The acquisition module 510 is used to acquire laser point clouds. The processing module 520 is used to determine whether there is an obstacle on the road when it is determined from the laser point clouds that there is a suspected obstacle on the road where a vehicle with intelligent driving capability is located, and the vehicle is located at a first position. The first position is the position where the vehicle accidentally braked based on the laser point clouds in the past. The first information is used to indicate the road conditions of the road where the vehicle is located at the current moment, and the second information is used to indicate the road conditions of the road where the vehicle was located when the vehicle accidentally braked at the first position in the past. The processing module 520 is also used to control the vehicle to perform intelligent driving based on whether there is an obstacle on the road.

[0087] Both the acquisition module 510 and the processing module 520 can be implemented in software or in hardware. For example, the implementation of the acquisition module 510 will be described below. Similarly, the implementation of the processing module 520 can refer to the implementation of the acquisition module 510.

[0088] As an example of a software functional unit, module 510 may include code running on a computing instance. The computing instance may include at least one of a physical host (computing device), a virtual machine, or a container. Further, the aforementioned computing instance may be one or more. For example, module 510 may include code running on multiple hosts / virtual machines / containers. It should be noted that the multiple hosts / virtual machines / containers used to run the code may be distributed within the same region or in different regions. Further, the multiple hosts / virtual machines / containers used to run the code may be distributed within the same availability zone (AZ) or in different AZs, each AZ including one or more geographically proximate data centers. Typically, a region may include multiple AZs.

[0089] Similarly, multiple hosts / virtual machines / containers used to run this code can be distributed within the same Virtual Private Cloud (VPC) or across multiple VPCs. Typically, a VPC is set up within a region. Communication between two VPCs within the same region, as well as between VPCs in different regions, requires a communication gateway to be set up within each VPC to enable interconnection between VPCs.

[0090] As an example of a hardware functional unit, the acquisition module 510 may include at least one computing device, such as a server. Alternatively, the acquisition module 510 may be implemented using a central processing unit (CPU), an application-specific integrated circuit (ASIC), or a programmable logic device (PLD). The PLD may be implemented using a complex programmable logical device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), a data processing unit (DPU), a neural network processing unit (NPU), a system-on-chip (SoC), an offload card, an accelerator card, or any combination thereof.

[0091] The multiple computing devices included in the acquisition module 510 can be distributed in the same region or in different regions. Similarly, the multiple computing devices included in the acquisition module 510 can be distributed in the same Availability Zone (AZ) or in different AZs. Likewise, the multiple computing devices included in the acquisition module 510 can be distributed in the same Virtual Private Cloud (VPC) or in multiple VPCs. These multiple computing devices can be any combination of computing devices such as servers, ASICs, PLDs, CPLDs, FPGAs, GALs, DPUs, NPUs, SoCs, offloading cards, and accelerator cards.

[0092] It should be noted that, in other embodiments, the acquisition module 510 and the processing module 520 can be used in any step of the above-described intelligent driving method.

[0093] In one possible implementation, the processing module 520 is specifically used to: project the laser point cloud onto an image corresponding to the road where the vehicle is located; if it is determined that the projection of a high-reflectivity point in the laser point cloud onto the image is not located on the road where the vehicle is located, and the projection of a low-reflectivity point in the laser point cloud onto the image is located on the road where the vehicle is located, then it is determined that there is a suspected obstacle on the road where the vehicle is located; wherein, the high-reflectivity point is a point in the laser point cloud with an intensity greater than or equal to a first threshold, and the low-reflectivity point is a point in the laser point cloud with an intensity less than or equal to a second threshold, the second threshold being less than the first threshold.

[0094] In one possible implementation, the processing module 520 is specifically used to: determine that there are no obstacles on the road where the vehicle is located when the feature similarity between the first information and the second information is greater than or equal to a third threshold; and determine that there are obstacles on the road where the vehicle is located when the feature similarity between the first information and the second information is less than the third threshold.

[0095] In one possible implementation, the first information and the second information are image information, the first information corresponds to a first image, and the second information corresponds to a second image; the processing module 520 is specifically used to: determine the feature similarity between the low reflection projection region in the laser point cloud located on the road where the vehicle is located in the first image and the corresponding region in the second image.

[0096] In one possible implementation, the first information and the second information are image information, the first information corresponds to a first image, and the second information corresponds to a second image; the processing module 520 is specifically used to: determine the feature similarity between the region corresponding to the road where the vehicle is located in the first image and the region corresponding to the road where the vehicle is located in the second image.

[0097] In one possible implementation, after determining that there are no obstacles on the road where the vehicle is located, the processing module 520 is further configured to: use the first information as the second information.

[0098] In one possible implementation, the processing module 520 is specifically used to: control the vehicle to continue driving and / or issue an alarm when it is determined that there is no obstacle on the road; and control the vehicle to brake when it is determined that there is an obstacle on the road.

[0099] Based on the above embodiments, this application also provides a device that can implement the methods in the above embodiments and has the functions of device 500. Referring to FIG6, the device 600 includes: a transceiver 601, a processor 602, and a memory 603. The transceiver 601, the processor 602, and the memory 603 are interconnected.

[0100] Optionally, the transceiver 601, the processor 602, and the memory 603 are interconnected via a bus 604. The bus 604 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus. The bus can be divided into an address bus, a data bus, and a control bus. For ease of illustration, only one thick line is used in Figure 6, but this does not indicate that there is only one bus or one type of bus.

[0101] The transceiver 601 is used to receive and send signals to enable communication with other devices.

[0102] The function of the processor 602 can be referred to the description in the above embodiments, and will not be repeated here.

[0103] The processor 602 can be a central processing unit (CPU), a network processor (NP), or a combination of a CPU and an NP. The processor 602 may further include a hardware chip. This hardware chip can be an application-specific integrated circuit (ASIC), a programmable logic device (PLD), or a combination thereof. The PLD can be a complex programmable logic device (CPLD), a field-programmable gate array (FPGA), a generic array logic (GAL), or any combination thereof. The processor 602 can implement the above functions through hardware, or it can implement them by executing corresponding software. The steps of the method disclosed in the above embodiments of this application can be directly reflected as the processor 602 completing the execution, or as the hardware and software modules in the processor 602 combining to complete the execution.

[0104] The memory 603 is used to store program instructions and data. Specifically, the program instructions may include program code, which includes computer operation instructions. The memory 603 may include volatile memory, such as random access memory (RAM); it may also include non-volatile memory, such as at least one disk storage device, hard disk drive (HDD), or solid state drive (SSD). The memory 603 may also be any other medium capable of carrying or storing program code in the form of instructions or data structures and accessible by a computer; this application does not limit this. The processor 602 executes the program instructions stored in the memory 603 to implement the above functions, thereby implementing the method provided in the above embodiments.

[0105] Based on the above embodiments, this application also provides a computer-readable storage medium storing a computer program, which, when executed by a computer, causes the computer to perform the methods provided in the above embodiments.

[0106] Optionally, the aforementioned computer may include, but is not limited to, communication devices such as terminal devices and network devices.

[0107] The storage medium can be any available medium that a computer can access. For example, but not limited to, a computer-readable medium can include RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer.

[0108] Based on the above embodiments, this application provides a chip for reading a computer program stored in a memory to implement the method provided in the above embodiments. Optionally, the chip may include a processor coupled to the memory for reading the computer program stored in the memory to implement the method provided in the above embodiments. Optionally, the chip may further include components such as a memory, a communication interface, and a power supply module. The memory is used to store the computer program; the communication interface is used to receive and send data; and the power supply module is used to supply power to the processor.

[0109] Based on the above embodiments, this application also provides a chip system including a processor for supporting a computer device in implementing the above embodiments. In one possible design, the chip system further includes a memory for storing necessary programs and data of the computer device. This chip system may be composed of chips or may include chips and other discrete components.

[0110] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, optical storage) containing computer-usable program code.

[0111] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to this application. It should be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in one or more blocks of the flowchart illustrations and / or one or more blocks of the block diagrams.

[0112] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means that implement the functions specified in one or more flowcharts and / or one or more block diagrams.

[0113] These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, such that the instructions, which execute on the computer or other programmable apparatus, provide steps for implementing the functions specified in one or more flowcharts and / or one or more block diagrams.

[0114] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. An intelligent driving method, characterized in that, The method includes: A laser point cloud is acquired. When the laser point cloud indicates that there is a suspected obstacle on the road where the vehicle with intelligent driving capability is located, and the vehicle is located at a first position, the presence of an obstacle on the road is determined based on first information and second information. The first position is the location where the vehicle mistakenly braked based on the laser point cloud in the past. The first information is used to indicate the road conditions of the road where the vehicle is located at the current moment, and the second information is used to indicate the road conditions of the road where the vehicle was located when the vehicle mistakenly braked at the first position in the past. The vehicle is controlled to perform intelligent driving based on whether there are obstacles on the road.

2. The method as described in claim 1, characterized in that, The determination of a suspected obstacle on the road where the vehicle with intelligent driving capability is located based on the laser point cloud includes: The laser point cloud is projected onto the image corresponding to the road; If it is determined that the projection of a high-reflectivity point in the laser point cloud onto the image is not located on the road, and the projection of a low-reflectivity point in the laser point cloud onto the image is located on the road, then it is determined that there is a suspected obstacle on the road; wherein, the high-reflectivity point is a point in the laser point cloud with an intensity greater than or equal to a first threshold, and the low-reflectivity point is a point in the laser point cloud with an intensity less than or equal to a second threshold, wherein the second threshold is less than the first threshold.

3. The method as described in claim 1, characterized in that, Determining whether there are obstacles on the road based on the first information and the second information includes: When the feature similarity between the first information and the second information is determined to be greater than or equal to a third threshold, it is determined that there are no obstacles on the road. When the feature similarity between the first information and the second information is less than the third threshold, it is determined that there is an obstacle on the road.

4. The method as described in claim 3, characterized in that, The first information and the second information are image information, where the first information corresponds to a first image and the second information corresponds to a second image; Determining the feature similarity between the first information and the second information includes: Determine the feature similarity between the low-reflection projection region in the laser point cloud located on the road in the first image and the corresponding region in the second image.

5. The method as described in claim 3, characterized in that, The first information and the second information are image information, where the first information corresponds to a first image and the second information corresponds to a second image; Determining the feature similarity between the first information and the second information includes: Determine the feature similarity between the region corresponding to the road in the first image and the region corresponding to the road in the second image.

6. The method according to any one of claims 1-5, characterized in that, After determining that there are no obstacles on the road, the method further includes: Use the first information as the second information.

7. The method according to any one of claims 1-6, characterized in that, The method of controlling the vehicle to perform intelligent driving based on whether there are obstacles on the road includes: When it is determined that there are no obstacles on the road, the vehicle is controlled to continue driving, and / or an alarm is issued; When an obstacle is detected on the road, the vehicle is brought to its brakes.

8. An apparatus, characterized in that, The device includes: The acquisition module is used to acquire laser point clouds; The processing module is configured to determine whether an obstacle exists on the road where a vehicle with intelligent driving capabilities is located, based on first information and second information, when the laser point cloud indicates that a suspected obstacle exists on the road and the vehicle is located at a first position; wherein, the first position is the position where the vehicle mistakenly braked based on the laser point cloud in the past, the first information is used to indicate the current road conditions, and the second information is used to indicate that there is no obstacle on the road; and is also configured to issue an alarm when it is determined that there is no obstacle on the road where the vehicle is located.

9. A vehicle, characterized in that, It includes an onboard detection device and means for performing the method as described in any one of claims 1-7, wherein the onboard detection device is used to generate a laser point cloud.

10. A computer program product, characterized in that, When it is operated on the device, it causes the device to perform the method as described in any one of claims 1-7.

Images