Vehicle obstacle avoidance control method, device and vehicle

Abstract

The application discloses a vehicle obstacle avoidance control method and device, wherein the method comprises: in the case that a vehicle running state parameter meets a low-speed running condition, predicting a vehicle running track according to environmental perception data of the vehicle and the vehicle running state parameter; constructing a vehicle contour envelope for each vehicle track point in the vehicle running track to generate a corresponding vehicle contour envelope sequence; detecting a target vehicle contour envelope which first spatially overlaps with an obstacle from the vehicle contour envelope sequence, and determining an obstacle collision time based on a target track prediction point time corresponding to the target polygon; and performing a vehicle obstacle avoidance collaborative control operation according to the obstacle collision time and accelerator pedal state information of the vehicle. Thus, false alarms and false braking in a low-speed working condition are significantly reduced, and the identification sensitivity of an active safety system to a real collision risk is improved.

Description

Technical Field

[0001] This application relates to the field of driver assistance systems, and in particular to vehicle obstacle avoidance control methods, devices and vehicles. Background Technology

[0002] With the increase in car ownership, traffic accidents in low-speed urban driving scenarios, such as collisions and scrapes, are becoming more frequent. Although these accidents are usually not fatal, they have a significant impact on driving safety and user experience due to their high probability of occurrence, high cumulative repair costs, and tendency to cause traffic congestion.

[0003] Currently, active safety technologies commonly used in automobiles, such as Autonomous Emergency Braking (AEB), employ sensors to monitor the surrounding environment and issue warnings or automatically brake when potential hazards arise. However, these technologies and their core decision-making models are largely optimized for scenarios such as highways, and their risk assessment logic is relatively straightforward, primarily based on pre-defined and simplified dynamic relationship determination models.

[0004] However, when this logic, which is mainly applicable to high-speed scenarios, is directly applied to low-speed environments with complex road conditions and variable driving intentions (such as following other vehicles, starting, and parking), its inherent limitations are exposed and amplified. Due to the highly complex interaction between the vehicle and its surrounding environment under low-speed conditions, the simplified judgment model described above struggles to accurately distinguish between real dangers and normal driving operations, resulting in frequent discrepancies between the system's judgment and the driver's actual intentions and real risks.

[0005] On the one hand, unnecessary interventions occur during normal acceleration, following, or navigating narrow areas, with frequent false alarms and braking severely disrupting driving and reducing user trust. On the other hand, when facing genuine emergency collision risks, the decision-making model may be too simplistic and fail to comprehensively consider key risk factors such as driver error, leading to insufficient reaction or delayed braking. Ultimately, this not only affects user experience and driving safety but also limits the further development and application of low-speed active safety technologies. Summary of the Invention

[0006] This application provides a vehicle obstacle avoidance control method, system, device, storage medium, and program product to at least solve one of the above-mentioned technical problems.

[0007] In a first aspect, embodiments of this application provide a vehicle obstacle avoidance control method, comprising: predicting a vehicle trajectory based on environmental perception data and vehicle trajectory parameters when the vehicle's driving state parameters are detected to meet low-speed driving conditions; constructing vehicle contour envelopes for multiple vehicle trajectory points in the vehicle trajectory to generate corresponding vehicle contour envelope sequences; the vehicle contour envelope sequences are combined in the order of trajectory point prediction times; detecting the target vehicle contour envelope that first spatially overlaps with an obstacle from the vehicle contour envelope sequence, and determining the obstacle collision time based on the target trajectory prediction point time corresponding to the target polygon; and performing vehicle obstacle avoidance cooperative control operations based on the obstacle collision time and the vehicle's accelerator pedal state information.

[0008] Secondly, embodiments of this application provide a vehicle, including: an environmental perception sensor for collecting environmental perception data around the vehicle; and a vehicle controller for executing any of the vehicle obstacle avoidance control methods described above.

[0009] Thirdly, embodiments of this application provide a storage medium storing one or more programs including execution instructions, which can be read and executed by electronic devices (including but not limited to computers, servers, or network devices) to perform any of the vehicle obstacle avoidance control methods described above.

[0010] Fourthly, a computer device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform any of the vehicle obstacle avoidance control methods described above in this application.

[0011] Fifthly, embodiments of this application also provide a computer program product, the computer program product including a computer program stored on a storage medium, the computer program including program instructions, which, when executed by a computer, cause the computer to execute any of the above-described vehicle obstacle avoidance control methods.

[0012] The beneficial effects of the embodiments of this application are as follows:

[0013] By introducing trajectory prediction based on environmental perception and driving status when the vehicle meets low-speed driving conditions, and constructing a temporally ordered envelope sequence of vehicle contour envelopes for multiple trajectory points, continuous and geometric modeling of the vehicle's future spatial occupancy is achieved. This allows for precise identification of the moment when the vehicle contour first spatially overlaps with an obstacle and determination of the collision time, elevating risk assessment in low-speed scenarios from a rough judgment based on distance or speed to a physical collision prediction based on the evolution of the actual contour. Furthermore, the system incorporates accelerator pedal status to determine the driver's actual intention, enabling timely intervention when there is an imminent collision risk and the driver's actions are inconsistent with safety requirements, while suppressing unnecessary braking triggers under normal low-speed operation. This not only significantly reduces false alarms and false braking in low-speed conditions and improves the sensitivity of the active safety system to identifying real collision risks, but also enables more driver-intended coordinated control, balancing safety, comfort, and system reliability. Attached Figure Description

[0014] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 A flowchart illustrating an example of a vehicle obstacle avoidance control method according to an embodiment of this application is shown;

[0016] Figure 2 A flowchart illustrating an example of performing vehicle obstacle avoidance cooperative control operation based on obstacle collision time and vehicle accelerator pedal state information according to an embodiment of this application is shown.

[0017] Figure 3 This diagram illustrates the speed-time relationship of a vehicle during the driver's braking operation.

[0018] Figure 4 A flowchart illustrating another example of performing vehicle obstacle avoidance cooperative control operation based on obstacle collision time and vehicle accelerator pedal state information according to an embodiment of this application is shown.

[0019] Figure 5 A flowchart illustrating an example of determining obstacle collision time using contour envelope according to an embodiment of this application is shown.

[0020] Figure 6 A flowchart illustrating an example of a vehicle obstacle avoidance control method according to an embodiment of this application is shown;

[0021] Figure 7This is a schematic diagram illustrating an example of the collision field point when a typical vehicle is turning in a low-speed scenario.

[0022] Figure 8 A structural block diagram of an example vehicle according to an embodiment of this application is shown;

[0023] Figure 9 This is a schematic diagram of the structure of an embodiment of the electronic device of this application. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other.

[0025] It should also be noted that, in this document, the terms "comprising" or "including" include not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0026] It should be noted that with the continuous increase in vehicle ownership, minor traffic accidents such as collisions and scrapes in urban low-speed driving scenarios are also showing a significant upward trend. Statistics show that accidents related to low-speed scenarios (below 20km / h) account for more than 30% of all traffic accidents, and constitute the vast majority in urban driving situations. Although such accidents usually do not cause serious personal injury, they are frequent, have high cumulative maintenance costs, and can easily cause traffic congestion. Furthermore, if the accelerator pedal is accidentally pressed (e.g., the gas pedal or accelerator), they can have serious consequences (such as the risk of crashing into a crowd). Therefore, active safety capabilities such as low-speed collision avoidance and accelerator pedal misapplication protection are of great significance for improving driving safety.

[0027] In related technologies, many manufacturers implement low-speed AEB and accelerator pedal misoperation protection functions based on single or multiple sensing devices, and generally rely on Time-To-Collision (TTC) and Time-To-React (TTR) as triggering criteria. For example, TTC is usually calculated based on obstacle distance and relative speed, while TTR is an assessment indicator of the time available for the driver to take evasive action.

[0028] However, some technical approaches that directly calculate TTC or TTR based on vehicle longitudinal distance and relative speed are primarily designed for high-speed straight-line scenarios. In urban conditions such as low speeds, high-curvature steering, and lateral approach, the spatial relationships and motion patterns of vehicles are more complex. The calculated TTC / TTR results often deviate significantly from the actual collision time, leading to numerous false triggers or missed triggers in low-speed conditions. Furthermore, obstacle recognition based on a single sensing source (such as vision or ultrasound only) is susceptible to factors such as occlusion, blind spots, and material reflections, further amplifying the risk of misjudgment.

[0029] It should be understood that the above description of the relevant technologies is intended only to help the public better understand the inventive spirit and motivation of this application, and is not intended to limit this application. Furthermore, the technical solutions described in the above-mentioned relevant technologies are not prior art, and may also be undisclosed technical solutions, such as those under research or in the laboratory stage.

[0030] The technical solutions in this application, including the collection, storage, use, processing, transmission, provision, and disclosure of users' personal information, comply with relevant laws and regulations and do not violate public order and good morals.

[0031] Figure 1 A flowchart illustrating an example of a vehicle obstacle avoidance control method according to an embodiment of this application is shown.

[0032] Regarding the execution subject of the method in the embodiments of this application, it can be any controller or processor with computing or processing capabilities. In some examples, the method in the embodiments of this application can be integrated and configured in an electronic device or terminal through software, hardware or a combination of software and hardware, and the type of terminal or electronic device can be diverse, such as mobile phone, tablet computer, desktop computer or vehicle terminal, etc.

[0033] For example, the execution subject of the method in this application embodiment can be integrated into the vehicle obstacle avoidance control controller. Based on the characteristics of low-speed scenarios, the trajectory prediction, contour envelope construction and collision time calculation process are reorganized. In low-speed complex road conditions (such as congested following, queuing start, parking, narrow road meeting, etc.), a more granular evaluation method that is more in line with the actual vehicle motion characteristics can be adopted, which significantly reduces the frequency of false alarms and false braking, and improves the applicability and effectiveness of low-speed active safety control.

[0034] like Figure 1 As shown, in step S110, when the vehicle driving state parameters are detected to meet the low-speed driving conditions, the vehicle driving trajectory is predicted based on the vehicle's environmental perception data and the vehicle driving state parameters.

[0035] In some implementations, real-time driving status parameters, such as current vehicle speed, longitudinal acceleration, steering angle, steering wheel angular velocity, gear status, and throttle opening rate of change, are obtained from the vehicle's CAN network, powertrain controller, or chassis controller. Subsequently, based on a preset low-speed threshold (e.g., 0-20 km / h, calibrated according to the vehicle model), it can be determined whether the vehicle is in a low-speed condition. If the conditions are met, the trajectory prediction model can be activated.

[0036] Here, the trajectory prediction model can be diverse, such as a single-track or dual-track vehicle kinematic model, a short-time dynamic model, or a path extrapolation model integrating inertial navigation and vision / radar. It utilizes the vehicle's current speed, acceleration, turning angle, and prediction time domain length to generate trajectory points at multiple discrete future moments. During the prediction process, environmental perception data (from vision, millimeter-wave radar, ultrasonic sensors, lidar, etc.) can be integrated to confirm road boundaries, obstacle locations, and vehicle-accessible areas, ensuring the rationality of the trajectory prediction and preventing the generation of trajectory points inconsistent with the physical environment. Furthermore, the prediction time domain window can be customized or adjustable, for example, covering 1-3 seconds. Prediction points can also be generated at fixed time intervals (e.g., 50ms-100ms) to ensure the temporal continuity of the trajectory points and achieve accurate characterization of the vehicle's future path.

[0037] Regarding the details of data acquisition, in some examples of embodiments of this application, corresponding raw environmental perception data are collected based on each environmental perception sensor. The collected raw environmental perception data are then fused to generate an obstacle point cloud, which at least represents the spatial location information of the obstacles. The environmental perception sensors include one or more of the following: ultrasonic radars distributed in the forward, rearward, and lateral directions of the vehicle; millimeter-wave radars positioned in the forward and / or rearward directions of the vehicle; surround-view cameras; and lidar.

[0038] In some implementations, to improve the accuracy of environmental characterization during low-speed trajectory prediction, after acquiring vehicle driving state parameters and determining that the vehicle is in a low-speed condition, raw environmental perception data can be further collected based on multi-source environmental perception sensors. These include ultrasonic radars distributed in front of, behind, and to the sides of the vehicle, millimeter-wave radars positioned in front of and / or behind the vehicle, vehicle surround-view cameras, and lidar. Since different sensors have complementary characteristics in detection distance, angular resolution, texture perception, and speed measurement, the raw data collected by each sensor are first synchronized in time and transformed to a single coordinate system, enabling them to describe the spatial environment information at the same moment in the vehicle coordinate system. Subsequently, using feature association, point cloud fusion, and target matching, the detection results from different data sources are spatially fused to form an obstacle point cloud representing the true spatial location and shape of obstacles.

[0039] The obstacle point cloud generated through the multi-source fusion process not only contains the three-dimensional spatial location information of obstacles but also carries their size, shape, and category features, thus providing reliable environmental boundary constraints. With the help of the obstacle point cloud, the trajectory prediction model can more accurately identify passable areas around the vehicle, avoiding the generation of predicted paths that traverse obstacles or do not conform to the actual road structure, thereby significantly improving the rationality and safety of trajectory prediction. Especially in environments such as low-speed following, parking, and meeting oncoming traffic in narrow passages, where obstacles are close together and diverse in type, the fused point cloud, through the fusion processing of multi-source sensor data, can effectively reduce the potential for missed or false detections from a single sensor, achieving robust obstacle detection in complex environments and improving the actual perception capability of complex near-field environments.

[0040] In step S120, vehicle contour envelopes are constructed for multiple vehicle trajectory points in the vehicle's driving trajectory to generate corresponding vehicle contour envelope sequences. These vehicle contour envelope sequences are combined according to the order of the predicted time of the trajectory points.

[0041] It should be noted that the trajectory prediction yields the trajectory of the vehicle's center of mass, but the vehicle's collision risk is related to the actual outline of the vehicle body, rather than the center of mass itself. Therefore, the physical shape of the vehicle at each prediction moment is accurately modeled and a sequence of contour envelopes that evolves over time is formed.

[0042] Specifically, for each predicted trajectory point, based on the vehicle's external parameters (such as vehicle length, width, front and rear overhangs, and track width) and the vehicle's heading angle at that moment (calculated from the steering wheel angle), a vehicle contour envelope is constructed for that corresponding moment. The envelope can use a rectangular polygon model (suitable for most passenger vehicles), or, in advanced versions, an approximate convex polygon model can be used to make the vehicle's posture more accurate. For example, using the trajectory point as a coordinate reference, the coordinates are rotated according to the vehicle's heading angle, and the coordinates of the outer contour vertices are generated according to the vehicle's dimensions, forming a polygon representing the vehicle's boundary. Subsequently, the system arranges all envelopes in chronological order according to the trajectory time, forming a vehicle contour envelope sequence. This sequence fully describes the spatial form actually occupied by the vehicle in the next few hundred milliseconds to several seconds.

[0043] In step S130, the target vehicle contour envelope that first spatially overlaps with the obstacle is detected from the vehicle contour envelope sequence, and the obstacle collision time is determined based on the target trajectory prediction point time corresponding to the target polygon.

[0044] It should be noted that the essence of a collision between a vehicle and an obstacle is the spatial overlap of their geometric boundaries. Therefore, only when the vehicle's outline envelope first enters the space occupied by the obstacle can this moment be confirmed as the earliest actual collision point. Thus, by using the envelope sequence traversal method to determine the first overlap, the moment of collision can be accurately located.

[0045] More specifically, geometric collision detection algorithms (such as convex polygon intersection detection) can be used to detect each envelope in the vehicle contour envelope sequence against the known obstacle contour one by one, determining whether there is an intersection between the vehicle envelope polygon and the obstacle envelope. If there is an overlap, the trajectory prediction time corresponding to that envelope is recorded; if it is the first overlap in the sequence, that time is confirmed as the obstacle collision time. Furthermore, the obstacle can be a stationary or low-speed moving target, and its position and shape are provided in real time by the vehicle environment perception system. If there is no overlap in the envelope sequence, the system determines that there is no collision risk for the vehicle in the prediction time domain. By detecting the first overlap based on multi-time envelopes, a collision time with actual physical meaning can be obtained.

[0046] In step S140, the vehicle obstacle avoidance cooperative control operation is performed based on the obstacle collision time and the vehicle's accelerator pedal status information.

[0047] It should be noted that in low-speed scenarios, driver intentions are complex, and situations such as accidentally pressing the accelerator or accelerating abnormally towards obstacles are common. Therefore, relying solely on collision time to determine braking trigger may lead to false alarms or interfere with driving. By incorporating accelerator pedal status, the system can introduce driver intention judgment into risk assessment, enabling the control strategy to accurately intervene in dangerous situations while suppressing unnecessary intervention during normal operation. It should be understood that the accelerator pedal can be either the accelerator pedal or the power pedal to support different types of vehicle acceleration control. In some examples described in this article, the description will be simplified by referring to the accelerator pedal, but it should be understood that it can also refer to the power pedal.

[0048] In some implementations, multiple factors can be considered, such as the duration of the obstacle collision (e.g., less than 3 seconds is considered high risk), the current throttle opening and its rate of change, and whether the driver is in a state of continuous acceleration, lightly pressing the accelerator, or releasing the accelerator. This allows the control system to execute multi-level obstacle avoidance cooperative control strategies. For example, for imminent collision risks and significant acceleration, rapid braking or traction limiting can be used; for low-risk normal fine-tuning behaviors, no active intervention or only a warning can be given; and in the absence of collision risk, the current vehicle dynamics are maintained. Furthermore, braking intensity can non-linearly increase as the collision time decreases in real time, thus balancing safety and comfort.

[0049] By integrating collision time with accelerator pedal state through the embodiments of this application, different scenarios such as normal driving, intentional approach, misoperation, and real danger can be effectively distinguished. This makes low-speed obstacle avoidance control no longer a singular approach, but a collaborative control method with scene understanding capabilities. As a result, the number of false braking events is significantly reduced, and the system's ability to respond promptly to emergency risks is improved, thereby enhancing driving comfort, driving safety, and user trust.

[0050] Figure 2 A flowchart illustrating an example of a vehicle obstacle avoidance cooperative control operation performed according to an embodiment of this application, based on obstacle collision time and vehicle accelerator pedal state information.

[0051] like Figure 2 As shown, in step S210, it is monitored whether the obstacle collision time exceeds a first collision time threshold. Here, the urgency of the current collision risk is assessed using the first collision time threshold to distinguish between low-risk and high-risk situations and determine whether emergency obstacle avoidance control is necessary.

[0052] Here, a reasonable first collision time threshold (e.g., 3 seconds) can be set based on different vehicles, driving environments, and safety standards. If the collision time is greater than this value, it indicates that the collision occurred a long time ago, and the risk between the vehicle and the obstacle can be considered low. If the collision time is less than this value, it means that the collision risk is imminent, and the system will proceed to the next step of risk assessment.

[0053] In step S221, when the obstacle collision time exceeds the first collision time threshold, the driving state is determined to be safe.

[0054] Once the system determines that the collision time with an obstacle exceeds the first collision time threshold, it indicates that the collision risk is relatively low, and no emergency obstacle avoidance maneuver is required. This effectively avoids unnecessary system intervention in low-risk scenarios, ensuring the driver maintains normal vehicle control while increasing driver trust in the active safety system and reducing the psychological burden of over-reliance on it.

[0055] In step S223, when the obstacle collision time does not exceed the first collision time threshold, it is identified whether the accelerator pedal state information matches the preset strong acceleration intention condition. The accelerator pedal state information includes accelerator pedal opening information and / or accelerator pedal change rate.

[0056] At low speeds, a driver's accelerator pedal operation can be affected by unclear intentions or misoperation. Especially when encountering obstacles, the driver may unintentionally increase the accelerator, causing the vehicle to approach the obstacle rapidly, thus increasing the risk of collision. Therefore, identifying the accelerator pedal's state information (including accelerator pedal opening information and rate of change) can help determine whether the driver intends to accelerate strongly and prevent accidental accelerator pedal presses.

[0057] In some implementations, the accelerator pedal opening (e.g., the specific position of the accelerator pedal) can be acquired in real time using onboard sensors (such as pedal sensors), and the rate of pedal change (i.e., the rate of change of accelerator pedal opening) can be calculated. Furthermore, based on years of driving data and vehicle dynamics models, the system can use machine learning or rules to set certain conditions for strong acceleration intent. For example, when the accelerator pedal opening exceeds a certain set threshold (e.g., above 70%) and the rate of change is rapid (e.g., greater than a certain value, such as a 30% change rate of opening), it can be determined that the driver has a strong acceleration intent.

[0058] In step S230, if the accelerator pedal status information is found to match the strong acceleration intention condition, the accelerator pedal mis-pressing protection function is activated to perform vehicle driving force output limiting control operation.

[0059] Here, upon recognizing a strong acceleration intent, control measures can be taken to prevent collisions caused by accidental accelerator pedal press. By limiting the vehicle's driving force output (such as limiting the throttle signal and adjusting engine torque output), the control system can effectively slow down the vehicle's acceleration process, thereby reducing the risk of collision.

[0060] By accurately identifying the accelerator pedal status and the driver's intentions, the system can effectively determine whether the driver has entered a dangerous area due to accidental acceleration. Once the system detects that the accelerator pedal status matches the conditions for strong acceleration intention, it immediately activates the accelerator pedal misoperation protection function. This not only prevents accidental collisions caused by misoperation but also enhances the system's ability to adapt to the driver's true intentions, making it more intelligent. For example, when the obstacle is very close and the collision time is short, the control force will be stronger; while when the collision risk is far away, the control force will be weaker.

[0061] Therefore, by activating the accelerator pedal misoperation protection function, the system ensures that the vehicle does not approach obstacles excessively under strong acceleration intentions, reducing the probability of a collision. Through intelligent speed limit control, the system can gently intervene in driver operation, avoiding sudden braking intervention and providing a smooth and comfortable driving experience while ensuring safety.

[0062] In some examples of embodiments of this application, in order to improve the accuracy of collision risk judgment in the scenario of accidental accelerator pedal depressing, the system further determines the longitudinal power link response delay based on the accelerator pedal state information, and calculates the first collision reaction time TTR_AMAP for triggering the accidental depressing protection function by combining the obstacle collision time and the actual braking process time of the vehicle.

[0063] Figure 3The diagram illustrates the speed-time relationship of a vehicle during a driver's braking operation, where V represents the vehicle speed, T is the time axis, V0 is the current speed, V1 is the response speed after pedal operation, and t1-t3 correspond to the segmented response points of the acceleration phase, the constant speed phase, and the braking phase, respectively.

[0064] First, determine the longitudinal power link response delay corresponding to the accelerator pedal status information.

[0065] When a vehicle receives input from the accelerator pedal, it does not immediately exhibit a change in acceleration; instead, it requires a response process through the longitudinal power chain, including the engine, motor, and transmission. Therefore, the impact of accelerator pedal operation on the vehicle's actual speed inevitably involves a chain response delay t1. To accurately predict the speed evolution trend of a vehicle in the event of accidental accelerator pedal input, this delay should first be determined and incorporated into the collision time calculation.

[0066] Specifically, the system monitors the pedal opening and pedal change rate in real time, obtaining precise input through the pedal sensor. Based on the link response modeling calibrated for different vehicle models, the power link delay t1 is obtained from calibration tests of different vehicle models. It may include throttle signal filtering delay, engine / motor torque build-up time, transmission hydraulic pressure build-up time, or tire longitudinal force build-up process, etc. For example, the measured delay is generally tens to hundreds of milliseconds.

[0067] For example, based on the vehicle's calibrated throttle opening-acceleration mapping table, the target acceleration a1 of the vehicle at the current pedal opening is obtained. By combining the above parameters, it can be determined that pedal operation will cause the vehicle to continuously accelerate from V0 to V1 within the time period t1. By unifying the modeling of accelerator pedal state, powertrain characteristics, and acceleration mapping relationship, a realistic and effective basis for vehicle dynamic response is provided for collision time prediction. Especially in the case of accidental throttle pressing leading to short-term rapid acceleration, the real-time performance and accuracy of risk assessment can be significantly improved.

[0068] The first collision reaction time is calculated based on the obstacle collision time, the longitudinal power link response delay, and the braking process time.

[0069] It should be noted that traditional TTC calculations are based solely on the speed-distance relationship, without considering acceleration changes caused by throttle input, or factors such as link delay and braking start delay, and therefore cannot accurately cover scenarios where the accelerator is accidentally pressed.

[0070] like Figure 3The speed-time relationship shown is divided into three phases: acceleration, constant speed, and braking. During acceleration, the acceleration is a1 with a duration of t1. Although the driver has released the accelerator or begun preparing to brake, there is a non-negligible response delay in the longitudinal power chain consisting of the engine / motor, transmission, and tires. During this delay phase, the vehicle maintains its original driving force and continues to accelerate or maintain power output, causing the vehicle speed to increase.

[0071] During the constant speed phase, the driver has not yet applied effective braking, and the vehicle speed remains relatively stable for a short period of time. This represents the effective time window in which the driver can make a braking response, and it is also a key segment that the system needs to consider when calculating the collision response time. Its duration is t2.

[0072] During the braking phase, the acceleration is a3 and the braking process time (or driving reaction time) is t3. Once the driver's braking operation is truly effective, the vehicle enters the braking process phase. At this time, the vehicle decelerates with an acceleration of a3 (such as the maximum braking acceleration) until it comes to a complete stop or avoids contact with the obstacle.

[0073] During the acceleration phase, from acceleration to braking, after the pedal is pressed, the vehicle maintains acceleration a1 within the power link delay time t1.

[0074] According to the kinematic formula, the speed increases to:

[0075] Equation (1)

[0076] The speed increases from V0 to V1.

[0077] During the constant speed phase, the vehicle enters a brief constant speed phase after reaching V1 (corresponding to...). Figure 3 The intermediate speed plateau segment is used to solve for the middle segment t2 in the target collision time window TTR_AMAP.

[0078] Let s1 be the displacement during the acceleration phase:

[0079] Equation (2)

[0080] Let s3 be the displacement required during the braking phase (a3 be the braking acceleration):

[0081] Equation (3)

[0082] In the formula, The relative speed of the obstacle can be obtained in real time through millimeter-wave radar, lidar, or vision-radar fusion algorithms. For example, the speed of the obstacle detected by the sensor is differentially calculated with the current speed of the vehicle to obtain the real-time relative speed of the obstacle relative to the vehicle.

[0083] Remaining distance of the obstacle The distance to the three stages is satisfied:

[0084] Equation (4)

[0085] Uniform speed phase time:

[0086] Equation (5)

[0087] During the braking phase, the vehicle brakes at a deceleration of a3 until a collision is avoided or the vehicle comes to a complete stop. The final AMAP reaction time TTR_AMAP is... This is the effective time window that the driver or active safety system truly has to identify risks and make braking decisions.

[0088] Taking into account the throttle-acceleration mapping relationship, actuator / power link response delay, and subsequent braking capability, a time margin can be reserved before the latest braking decision moment to avoid a collision, based on the current pedal opening and vehicle status.

[0089] If the reaction time to the first collision is less than the first reaction time threshold, the accelerator pedal mis-pressing protection function is activated.

[0090] When TTR_AMAP is detected to be less than the preset first reaction time threshold, it can be determined that the driver's available reaction time is insufficient. Combined with the strong acceleration intention condition, the accelerator pedal mis-pressing protection is triggered. If the driver continues to apply high throttle input, it is highly likely to be a throttle mis-pressing behavior. At this time, the driving force output should be limited, and the acceleration ability should be immediately limited to avoid further risks, thereby ensuring a safe distance between the vehicle and the obstacle.

[0091] Figure 4 A flowchart illustrating another example of performing vehicle obstacle avoidance cooperative control operation based on obstacle collision time and vehicle accelerator pedal state information according to an embodiment of this application is shown.

[0092] In step S410, if it is found that the accelerator pedal state information does not match the accelerator pedal strong acceleration intention condition, the second collision reaction time is calculated based on the obstacle collision time and the braking process time.

[0093] Here, when the driver's accelerator pedal state information does not meet the conditions for strong acceleration intention (e.g., the accelerator pedal opening is low and the rate of change is slow), it can be assumed that the driver does not currently have a significant active acceleration demand. In this case, the future motion trend of the vehicle will not be significantly deviated due to the obvious acceleration change caused by the throttle.

[0094] Therefore, in such scenarios, there is no need to adopt a braking model that considers throttle-acceleration mapping and actuator response delay. Instead, we can directly revert to a simpler and more conservative collision time model, namely the traditional TTR (Time To Collision Reaction), which estimates the time window in which the driver or system must begin braking to avoid a collision based on the vehicle's current speed, relative speed to the obstacle, and the braking system's response time.

[0095] Specifically, the following real-time data can be obtained: current vehicle speed The relative velocity of the obstacle The initial distance between the vehicle and the obstacle The response time of the vehicle's braking system (braking process time) Maximum available deceleration .

[0096] For example, the total braking distance required to avoid a collision at maximum deceleration can be calculated first.

[0097] Equation (6)

[0098] Then calculate the total collision reaction time based on the remaining distance:

[0099] Equation (7)

[0100] Braking process time Subtracting from the TTR, the final collision reaction time is obtained as follows:

[0101] Equation (8)

[0102] in, This indicates the second collision reaction time.

[0103] Therefore, by adopting a second collision reaction time, excessively complex AMAP calculations are avoided when there is no strong acceleration intention, thus improving the system's execution efficiency and obtaining more stable and conservative risk assessment results for low-speed scenarios. This allows for an accurate measurement of the driver's effective reaction time under the current pedal state.

[0104] In step S420, the second collision reaction time is compared with a second reaction time threshold and a third reaction time threshold, wherein the second reaction time threshold is greater than the third reaction time threshold.

[0105] Here, in order to achieve graded safety intervention, two response thresholds are preset, namely a larger second reaction time threshold. and a smaller third reaction time threshold ,and These two thresholds are used to classify different levels of collision urgency into three levels.

[0106] In step S431, if the second collision reaction time is detected to be less than the second reaction time threshold and greater than or equal to the third reaction time threshold, a vehicle collision warning operation is performed.

[0107] Here, the warning methods are diverse, including audible and visual alarms (buzzer + instrument panel prompts), HUD display of the collision path, seat vibration, or steering wheel vibration, etc. This significantly improves the driver's risk perception ability without mandatory intervention, avoids comfort and experience issues caused by accidental braking, and greatly reduces the probability of collisions in medium-risk scenarios.

[0108] In some examples of embodiments of this application, the orientation of the collision point is obtained when there is spatial overlap between the target vehicle's contour envelope and the obstacle. Specifically, after detecting the first spatial overlap with the obstacle based on the vehicle contour envelope sequence, the orientation information of the collision point is further determined according to the specific location of the overlapping area in the vehicle coordinate system. For example, it is identified whether the danger comes from the front, front left corner, front right corner, side, or rear area of ​​the vehicle, thereby accurately reflecting the spatial direction in which the future collision is most likely to occur.

[0109] Furthermore, the system executes vehicle collision warning operations based on the location of the collision point. Specifically, after obtaining the location of the collision point, the system can perform targeted collision warning operations based on that location, enabling the driver to perceive the direction of danger in the shortest possible time. For example, when the collision point is located at the front left corner of the vehicle, a highlight can be displayed in the left area of ​​the instrument panel or HUD, or the left side of the seat can be vibrated; when the risk comes from the side, a warning can be issued through the corresponding door panel indicator or the steering wheel side vibration module. Thus, by aligning the warning signal spatially with the direction of danger, the driver can quickly locate the source of risk without additional judgment and take appropriate deceleration or evasive action, thereby further enhancing system safety and user experience.

[0110] In step S433, if the second collision reaction time is detected to be less than the third reaction time threshold, a vehicle driving force output limiting control operation is performed.

[0111] For example, when When the risk level is determined to be low, no intervention is required; when When the risk level is determined to be medium risk, an early warning is triggered; when When the risk level is determined to be high, the driving force limitation is triggered.

[0112] In this embodiment, a dual-threshold grading mechanism is used to distinguish between situations where warnings are avoidable and situations where braking is unavoidable, thereby improving the system's response accuracy to different levels of danger and avoiding the need for forced intervention to handle all risks, thus enhancing driving comfort and user acceptance.

[0113] Figure 5 A flowchart illustrating an example of determining obstacle collision time using contour envelope according to an embodiment of this application is shown.

[0114] In step S510, the time of the adjacent trajectory prediction point before the time of the target trajectory prediction point is determined according to the vehicle contour envelope sequence.

[0115] It should be noted that in the vehicle contour envelope sequence, the target trajectory prediction point is the moment when the spatial overlap between the vehicle contour envelope and the obstacle is first detected. However, the temporal resolution of trajectory prediction is usually determined by a fixed sampling step size (e.g., 50 ms, 100 ms), which may not be sufficient to accurately characterize the actual time point at which the vehicle and obstacle come into contact. Therefore, the nearest neighboring prediction point before the target time can be found from the original trajectory point sequence to perform more refined temporal interpolation in between.

[0116] Specifically, the time index of the target prediction point is read from the vehicle contour envelope sequence, for example, t. k Then take the next adjacent predicted point t. k-1 The adjacent prediction point ensures that the vehicle has not yet touched the obstacle at that moment, therefore at t k-1 With t k There must exist a spatial transition interval between the vehicle contour envelopes, where the envelopes change from "non-overlapping" to "overlapping," which serves as the starting point for subsequent iterative interpolation.

[0117] In step S520, at least one time interpolation process is performed on the target trajectory prediction point time and the adjacent trajectory prediction point time to obtain at least one corresponding interpolated prediction time.

[0118] Specifically, in t k-1 and t kDuring this period, the vehicle's contour envelope state changes from safe to collision-prone. To improve the temporal accuracy of the collision time, interpolation is performed on this time period, further refining the originally coarse sampling time. The interpolated predicted time can refine the temporal granularity of the envelope change, making the collision detection closer to the actual occurrence time. For example, linear interpolation, cubic spline interpolation, etc., can be used. Furthermore, each interpolated time corresponds to a more precise predicted position and attitude of the vehicle in the future. The number of interpolated times can be flexibly adjusted according to the needs of the scenario, and one or more iterations of interpolation can be performed. Thus, trajectory prediction is upgraded from coarse-grained temporal resolution to fine-grained temporal resolution, thereby improving the accuracy of collision time estimation and significantly enhancing the time sensitivity during low-speed driving.

[0119] In step S530, for each sequentially arranged interpolation prediction time, an interpolated vehicle contour envelope corresponding to the interpolation prediction time is constructed, and it is detected whether the interpolated vehicle contour spatially overlaps with the obstacle.

[0120] It should be noted that the vehicle's future attitude and position change continuously over time. Even if the changes in the vehicle and obstacle states are not significant between the original prediction points, the vehicle outlines may overlap at small time steps. Therefore, the vehicle outline envelope can be reconstructed at each interpolation time point, and a polygon intersection algorithm can be used to determine whether it spatially overlaps with obstacles individually. Based on dynamic fine-grained detection, the accuracy of collision time prediction in low-speed scenarios can be improved.

[0121] In step S540, the first interpolation prediction time when the first spatial overlap with the obstacle is detected in each interpolated vehicle contour envelope is used to determine the obstacle collision time.

[0122] Because the interpolated predicted time has a higher temporal resolution, the interpolated time at which the vehicle contour envelope first overlaps with the obstacle can represent a more accurate approximation of the actual collision time, and the system uses this as the final obstacle collision time.

[0123] Specifically, in the time series of the interpolation contour envelope, the first interpolation moment containing overlapping states is found and marked as the first interpolation prediction moment. This moment is used as the obstacle collision time predicted in this case, which significantly improves the accuracy of collision time judgment in low-speed and low-dynamic scenarios, avoids misjudgment or delayed judgment due to insufficient time resolution, and improves the execution reliability and user experience of the entire low-speed active safety system.

[0124] By introducing interpolation between adjacent time points, constructing interpolated contour envelopes, and detecting overlaps one by one, the system achieves refined processing of the temporal dimension based on the original trajectory prediction model. This enables the system to obtain the true collision time between vehicles and obstacles with higher accuracy. Therefore, it eliminates the need for high frame rate sensors, achieving temporal refinement through interpolation at the algorithm level. This approach offers high adaptability and cost-effectiveness, enabling refined collision prediction and safety control in low-speed traffic environments at low cost.

[0125] In some examples of embodiments of this application, the time interval between the first interpolation prediction time and the adjacent second interpolation prediction time before the first interpolation prediction time is less than a preset time interval threshold.

[0126] Specifically, at the first interpolation prediction time when the first spatial overlap occurs... The system can then further determine the interpolation time from its previous interpolation prediction time. Whether the time interval between them is less than a preset time interval threshold. For example, when When the time interval is less than a threshold (e.g., 5 ms, 10 ms), the current interpolation accuracy is considered sufficient to characterize the transition between states that have never been touched and those that have been touched, and there is no need to continue performing more intensive interpolation. This effectively avoids continuing interpolation operations when sufficient time accuracy has been achieved, reducing unnecessary computational overhead.

[0127] Alternatively or additionally, the time interpolation process employs median interpolation based on a time interval bisection.

[0128] Specifically, when the time span between the first and second interpolation prediction times is large, a median interpolation method based on a time interval bisection can be used to further refine the collision time. This is achieved by taking the median time between the two interpolation times (e.g., ...). The median moment is used as the new interpolation prediction moment. The corresponding interpolated vehicle contour envelope is then constructed, and it is determined whether it spatially overlaps with an obstacle. If the contour at the median moment still does not overlap with an obstacle, this median is taken as the new "non-collision moment," and continues to form a new interpolation interval with the subsequent "collision moment." If the contour at the median moment has already overlapped, this median is taken as the new "collision moment," and continues to form a new interpolation interval with the preceding "non-collision moment." Through this binary search iterative process, until the time interval between adjacent interpolation moments meets a preset time interval threshold, the median moment at this point can be used as a more accurate collision time.

[0129] Figure 6A flowchart of an example of a vehicle obstacle avoidance control method according to an embodiment of this application is shown, which describes the complete decision-making process of the system in low-speed conditions, based on multi-sensor perception fusion, vehicle trajectory prediction, collision time calculation and throttle / brake coordinated control, to achieve omnidirectional collision avoidance and throttle misapplication protection.

[0130] like Figure 6 As shown, the system first determines whether the vehicle meets the functional activation conditions, such as whether the vehicle speed is in a low-speed range (e.g., below 20 km / h), whether the gear is in drive / reverse / neutral, whether the driver's seatbelt is fastened, and whether various environmental perception sensors are in normal working condition, including health checks of sensors such as surround-view cameras, millimeter-wave radar, ultrasonic radar, and lidar. When the working conditions are met, the system collects environmental data around the vehicle based on multiple types of sensors such as ultrasonic radar, millimeter-wave radar, surround-view cameras, and lidar, and generates a unified obstacle point cloud through a multi-source perception fusion strategy. This obstacle point cloud can contain information such as the position, height, reflection characteristics, and type of obstacles, thus comprehensively representing the obstacle characteristics in all directions around the vehicle. This enables accurate near-range perception of static, dynamic, and lateral obstacles around the vehicle, and is particularly suitable for lateral distance mapping problems in turning scenarios.

[0131] After obtaining the obstacle point cloud, the omnidirectional collision detection module combines the vehicle's current speed, longitudinal acceleration, steering wheel angle, and other driving status information to make a short-term prediction of the vehicle's future trajectory, and generates a time-series vehicle polygonal contour envelope based on the predicted trajectory. Subsequently, the system performs omnidirectional collision detection on the vehicle contour envelope and obstacle point cloud using a binary search approach or other efficient spatial collision detection algorithms to identify obstacles that may collide with the vehicle and determine the obstacle with the minimum collision time and the corresponding collision point location. When the time-to-collision (TTC) is greater than a threshold of 1 (e.g., 3 seconds), it indicates that the vehicle does not pose an imminent risk, and the system does not intervene; however, when the TTC is less than a threshold of 1, the system enters the risk assessment and control strategy execution phase.

[0132] During the risk assessment phase, the system first identifies whether the driver has engaged the accelerator pedal abnormally. For example, when the accelerator pedal opening exceeds a set threshold or the rate of change of the accelerator pedal opening exceeds a preset threshold, the system predicts the vehicle's longitudinal acceleration based on the accelerator-acceleration mapping table and, combined with actuator response delay, calculates the collision reaction time TTR_AMAP, which takes into account the impact of the accelerator pedal. If TTR_AMAP is less than the preset threshold, meaning that the latest possible braking time for a collision under continuous accelerator pedal application has been compressed, the system identifies it as a potential accelerator pedal misapplication scenario and promptly activates the accelerator pedal misapplication protection function, including limiting drive force output, engine torque cut-off, and, if necessary, mild braking intervention.

[0133] If the driver's accelerator pedal input is normal, the system assesses the risk level based on a generally defined collision reaction time (TTR). The calculation of TTR comprehensively considers factors such as the vehicle's current speed, acceleration, braking system response time, and actuator delay. When the TTR is less than threshold 2 but not lower than threshold 3, the system triggers a directional collision warning, including audible and visual alarms, steering wheel or seat vibration alerts, and displays an image of the obstacle at the corresponding collision location on the in-vehicle display or 360° panoramic imaging interface, guiding the driver to take active braking or evasive action. When the TTR is less than threshold 3, the system immediately performs emergency braking and power cut-off operations to minimize or avoid a collision.

[0134] During vehicle start-up, due to potential driver inattention and closer proximity to obstacles, the system will prioritize suppressing the accelerator signal when a collision risk is detected. This is to prevent accidental accelerator pedal presses during low-speed start-up that could lead to a collision or injury to vulnerable road users. During the low-speed driving phase after vehicle start-up, if a collision risk is detected, the system will apply precise braking based on collision time and reaction time, achieving proactive safety control by intervening before a collision.

[0135] As a result, the system can comprehensively realize functions such as all-way collision avoidance warning, emergency braking intervention, and accelerator pedal misoperation protection under conditions such as low-speed start, reversing, narrow road passage, and urban congestion, achieving multiple active safety capabilities and effectively improving the vehicle's safety, reliability, and user experience in low-speed complex environments.

[0136] Figure 7 This diagram illustrates an example of the impact point of a typical vehicle turning in a low-speed scenario.

[0137] In this application embodiment, an omnidirectional collision point and collision time calculation method is proposed. Based on obstacle point cloud generated by fusion perception information and vehicle predicted trajectory, it can accurately obtain the real collision point position and collision time (TTC) between the vehicle and each obstacle under complex working conditions such as low speed and high curvature turning. It effectively overcomes the distance distortion problem that is easy to occur when traditional TTC calculation is based solely on distance or trajectory projection distance, and avoids missed collision control due to overestimation of TTC or false triggering of low-speed AEB due to underestimation of TTC.

[0138] Specifically, the fused multi-sensor perception results are transmitted to the decision-making and planning layer in the form of point clouds. Corresponding obstacle polygons are generated for each cluster of obstacle point clouds, and collision correlation is verified by combining information such as obstacle speed and drivable area to screen target obstacles that may collide with the vehicle.

[0139] like Figure 7As shown in the diagram, the four obstacle points represent four typical collision scenarios common during low-speed vehicle travel. Specifically, obstacle point 1 is located in front of the vehicle but outside its travel domain, and there is no risk of collision if the vehicle continues on its current predicted trajectory. Obstacle point 2 is located within the vehicle's travel domain, and the collision point is directly in front of or behind the vehicle. Obstacle point 3 is located within the vehicle's feasible domain and at its side front; the vehicle will make contact with the obstacle from the side front direction when traveling on its current trajectory. Obstacle point 4 is located to the side of the vehicle, and the vehicle will collide with the side of the vehicle when traveling on its current trajectory. For obstacle points 3 and 4, which are collision points during lateral or turning maneuvers, if the traditional method of projecting the obstacle's position onto the predicted trajectory and then calculating the distance is used, at the moment shown in the diagram, the projected point may have already fallen inside the vehicle's outline, resulting in a significant deviation between the calculated TTC and the actual time of collision.

[0140] To further meet the requirements of precise control under low-speed, high-curvature conditions, an omnidirectional TTC calculation method based on vehicle polygon traversal and time interval bisection can also be adopted. For example... Figure 7 As shown, starting from the current moment, the trajectory points of the vehicle at each future moment are predicted according to the time step, and a corresponding vehicle polygon outline is generated at each predicted moment. Then, these vehicle polygons are traversed sequentially along the time. When the vehicle polygon at a certain moment is detected to contain the target obstacle point for the first time, it can be determined that there is a transition interval from the "non-collision" state to the "collision" state between that moment and the previous moment. Next, interpolation is performed by dividing the time interval between the two adjacent predicted moments to obtain a new intermediate trajectory point and regenerate the vehicle polygon. It is then determined whether the polygon at the intermediate moment contains the obstacle point. If the intermediate moment contains the obstacle point, it is taken as the new "collision endpoint" and divided with the previous "non-collision endpoint" again. If the intermediate moment still does not contain the obstacle point, it is taken as the new "non-collision endpoint" and divided with the subsequent "collision endpoint" again. By repeating the above binary traversal process, the time difference between the trajectory points corresponding to the vehicle polygon containing the obstacle point and the vehicle polygon not containing the obstacle point is less than a preset time threshold (e.g., 0.05 seconds), which means that it is considered to be close to the actual collision moment in time.

[0141] Based on this, the time corresponding to the median of the two aforementioned moments is taken and differen from the current moment to obtain the high-precision collision time (TTC) for the obstacle point. Since this method directly performs an omnidirectional traversal based on the spatial overlap between the complete vehicle outline and the obstacle's geometric position, rather than simply using forward distance or trajectory projection distance, it is uniformly applicable to forward, backward, and various lateral collision scenarios. Especially under conditions such as low vehicle speed, sharp turns, and obstacle avoidance, it significantly improves the accuracy and stability of TTC calculation, thus providing a more reliable time basis for subsequent warnings, braking interventions, and accelerator pedal misapplication protection.

[0142] In this application embodiment, to address the problems of distance distortion and high misjudgment risk in traditional distance or trajectory projection-based TTC calculations under low-speed, high-curvature conditions, a polygon traversal method for accurately calculating the omnidirectional collision point and collision time of a vehicle is proposed. Specifically, a vehicle polygon is generated from the predicted trajectory points of the vehicle and omnidirectional collision detection is performed. Then, methods such as binary approximation and trajectory densification are used to gradually narrow down the time interval of the collision, thereby obtaining a high-precision collision time TTC. It should be understood that the implementation scope of this application embodiment is not only applicable to the binary search method based on trajectory prediction at equal time intervals, but can also be extended to the approximation method based on equidistant trajectory points, or directly construct a high-density trajectory by densifying trajectory points and then sequentially detect collisions, thereby supporting multiple implementation paths under different accuracy requirements and computational resource conditions.

[0143] Furthermore, the embodiments of this application are not only applicable to static obstacles, but also capable of accurately calculating the collision time and distance of moving targets by synchronously updating the position of dynamic obstacles at each prediction moment during the traversal process, further enhancing the applicability of the system in real traffic scenarios. Through omnidirectional obstacle perception and collision prediction, the system can effectively cover visual blind spot scenarios such as low-speed vehicles, turning, maneuvering, and reversing, providing more comprehensive risk identification capabilities for passenger cars, commercial vehicles, and other types of vehicles.

[0144] Therefore, by combining higher-precision collision time estimation, low-speed collision avoidance and accelerator misoperation protection are designed in a coordinated manner, and a multi-layer protection system is constructed from early warning, active collision avoidance to accelerator misoperation protection. It can trigger corresponding braking control, accelerator suppression or directional warning according to different risk levels, thereby significantly reducing false triggering events, improving user experience, and greatly enhancing the active safety performance of the vehicle in low-speed complex environments.

[0145] Figure 8 A structural block diagram of an example vehicle according to an embodiment of this application is shown.

[0146] like Figure 8 As shown, vehicle 800 includes an environmental perception sensor 810 and a vehicle controller 820.

[0147] The environmental perception sensor 810 is used to collect environmental perception data around the vehicle.

[0148] Environmental perception data can be diverse and provided by various types of sensors in different forms. This includes near-range obstacle distance data from ultrasonic radars distributed in the vehicle's front, rear, and sides; relative obstacle velocity and distance information from millimeter-wave radar; semantic information from surround-view cameras; and three-dimensional point cloud data from lidar. By synchronizing and calibrating these multi-dimensional sensor data in time and space, the vehicle can be provided with environmental features such as obstacle position, shape, type, and motion state covering all four directions, providing reliable input for subsequent trajectory prediction and collision risk calculation.

[0149] The vehicle controller 820 is used to execute the vehicle obstacle avoidance control method as described in any of the above embodiments of this application.

[0150] In some implementations, the vehicle controller 820 is executed by other electronic control units with equivalent functions, such as the vehicle controller (VCU), autonomous driving domain controller, ADAS controller, etc., and can achieve the same or similar technical effects as described in this application. More details can be referred to or combined with the description in the method embodiments above, and will not be repeated here.

[0151] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of combined actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Secondly, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application. In the above embodiments, the descriptions of each embodiment have their own emphasis; for parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0152] In some embodiments, this application also provides a computer program product, the computer program product including a computer program stored on a non-volatile computer-readable storage medium, the computer program including program instructions, which, when executed by a computer, cause the computer to perform any of the above-described vehicle obstacle avoidance control methods.

[0153] In some embodiments, this application also provides an electronic device, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform a vehicle obstacle avoidance control method.

[0154] The apparatus described in the embodiments of this application can be used to execute the vehicle obstacle avoidance control method of this application, and accordingly achieve the technical effects achieved by the vehicle obstacle avoidance control method of this application, which will not be elaborated further here. In the embodiments of this application, the relevant functional modules can be implemented by a hardware processor.

[0155] Figure 9 This is a schematic diagram of the hardware structure of an electronic device for executing a vehicle obstacle avoidance control method according to another embodiment of this application, as shown below. Figure 9 As shown, the device includes:

[0156] One or more processors 910 and memory 920, Figure 9 Take the 910 processor as an example.

[0157] The device for implementing the vehicle obstacle avoidance control method may also include an input device 930 and an output device 940.

[0158] The processor 910, memory 920, input device 930, and output device 940 can be connected via a bus or other means. Figure 9 Taking the example of a connection between China and Israel via a bus.

[0159] The memory 920, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules corresponding to the vehicle obstacle avoidance control method in the embodiments of this application. The processor 910 executes various functional applications and data processing of the server by running the non-volatile software programs, instructions, and modules stored in the memory 920, thereby implementing the vehicle obstacle avoidance control method of the above-described method embodiments.

[0160] The memory 920 may include a program storage area and a data storage area, wherein the program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the device, etc. Furthermore, the memory 920 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 920 may optionally include memory remotely located relative to the processor 910, and these remote memories may be connected to the device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0161] Input device 930 can receive input digital or character information and generate signals related to user settings and function control of the device. Output device 940 may include display devices such as a display screen.

[0162] The one or more modules are stored in the memory 920, and when executed by the one or more processors 910, they execute the vehicle obstacle avoidance control method in any of the above method embodiments.

[0163] The above-described product can perform the methods provided in the embodiments of this application, and has the corresponding functional modules and beneficial effects for performing the methods. Technical details not described in detail in this embodiment can be found in the methods provided in the embodiments of this application.

[0164] The electronic devices in this application embodiments exist in various forms, including but not limited to:

[0165] (1) Mobile communication devices: These devices are characterized by their mobile communication capabilities and primarily aim to provide voice and data communication. These terminals include: smartphones (e.g., iPhones), multimedia phones, feature phones, and low-end phones, etc.

[0166] (2) Ultra-mobile personal computer devices: These devices fall under the category of personal computers, possessing computing and processing capabilities, and generally also have mobile internet access features. These terminals include PDAs, MIDs, and UMPCs, such as the iPad.

[0167] (3) Portable entertainment devices: These devices can display and play multimedia content. This category includes audio and video players (such as iPods), handheld game consoles, e-book readers, as well as smart toys and portable car navigation devices.

[0168] (4) Server: A device that provides computing services. The components of a server include a processor, hard disk, memory, system bus, etc. Servers are similar to general computer architectures, but because they need to provide highly reliable services, they have higher requirements in terms of processing power, stability, reliability, security, scalability, and manageability.

[0169] (5) Other electronic devices with data interaction functions.

[0170] In some embodiments, this application also provides a mobile platform on which the computer device described in any embodiment of this application is installed. The mobile platform includes, but is not limited to, vehicles, tracked robots, bipedal robots, quadrupedal robots, etc., wherein the vehicle can be a passenger car, pickup truck, truck, etc. It should be noted that the above are merely examples, and this application does not limit the specific form of the mobile platform.

[0171] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0172] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0173] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A vehicle obstacle avoidance control method, comprising: When the vehicle's driving status parameters are detected to meet the low-speed driving conditions, the vehicle's driving trajectory is predicted based on the vehicle's environmental perception data and vehicle driving status parameters. For each of the multiple vehicle trajectory points in the vehicle's driving trajectory, a vehicle contour envelope is constructed to generate a corresponding vehicle contour envelope sequence; the vehicle contour envelope sequence is combined according to the order of the trajectory point prediction time. The target vehicle contour envelope that first spatially overlaps with the obstacle is detected from the vehicle contour envelope sequence, and the obstacle collision time is determined based on the target trajectory prediction point time corresponding to the target vehicle contour envelope. Based on the obstacle collision time and the vehicle's accelerator pedal status information, the vehicle obstacle avoidance cooperative control operation is executed.

2. The method according to claim 1, wherein, The step of performing vehicle obstacle avoidance cooperative control operations based on the obstacle collision time and the vehicle's accelerator pedal state information includes: When the collision time with the obstacle exceeds the first collision time threshold, it is determined to be a safe driving state; When the obstacle collision time does not exceed the first collision time threshold, the accelerator pedal status information is identified as matching the preset strong acceleration intention condition; the accelerator pedal status information includes accelerator pedal opening information and / or accelerator pedal change rate. If the accelerator pedal status information is found to match the strong acceleration intention condition, the accelerator pedal mis-pressing protection function is activated to perform vehicle driving force output limiting control operation.

3. The method according to claim 2, wherein, The activation of the accelerator pedal mis-pressing protection function includes: Determine the longitudinal power link response delay corresponding to the accelerator pedal state information; The first collision reaction time is calculated based on the obstacle collision time, the longitudinal power link response delay, and the braking process time. If the reaction time to the first collision is less than the first reaction time threshold, the accelerator pedal mis-pressing protection function is activated.

4. The method according to claim 2, wherein, After identifying whether the accelerator pedal status information matches a preset strong acceleration intention condition, the method further includes: If it is found that the accelerator pedal state information does not match the strong acceleration intention condition of the accelerator pedal, the second collision reaction time is calculated based on the obstacle collision time and the braking process time. If the second collision reaction time is detected to be less than the second reaction time threshold and greater than or equal to the third reaction time threshold, a vehicle collision warning operation is performed; the second reaction time threshold is greater than the third reaction time threshold. If the second collision reaction time is detected to be less than the third reaction time threshold, a vehicle driving force output limiting control operation is performed.

5. The method according to claim 4, wherein, The execution of the vehicle collision warning operation includes: Obtain the location of the collision point when the target vehicle's outline envelope overlaps with the obstacle in space; Execute a vehicle collision warning operation based on the location of the collision point.

6. The method according to claim 1, wherein, The step of detecting the target vehicle contour envelope that first spatially overlaps with the obstacle from the vehicle contour envelope sequence, and determining the obstacle collision time based on the target trajectory prediction point time corresponding to the target vehicle contour envelope, includes: Based on the vehicle contour envelope sequence, determine the time of the adjacent trajectory prediction point before the time of the target trajectory prediction point; At least one time interpolation process is performed on the target trajectory prediction point time and the adjacent trajectory prediction point time to obtain at least one corresponding interpolated prediction time. For each of the sequentially arranged interpolation prediction times, an interpolated vehicle contour envelope corresponding to the interpolation prediction time is constructed, and it is detected whether the interpolated vehicle contour envelope spatially overlaps with the obstacle. The first interpolation prediction time is detected in the envelope of each interpolated vehicle profile, and the obstacle collision time is determined based on the first interpolation prediction time.

7. The method according to claim 6, wherein, The time interval between the first interpolation prediction time and the adjacent second interpolation prediction time before the first interpolation prediction time is less than a preset time interval threshold; And / or, the time interpolation process employs median interpolation based on time interval bisection.

8. The method according to claim 1, further comprising: Each environmental sensing sensor collects its corresponding raw environmental sensing data. The collected raw environmental perception data are fused together to generate obstacle point clouds; The obstacle point cloud is used at least to characterize the spatial location information of the obstacles; The environmental perception sensors include one or more of the following: ultrasonic radars distributed in the front, rear and sides of the vehicle, millimeter-wave radars set in the front and / or rear of the vehicle, surround-view cameras, and lidar.

9. A computer device, comprising a memory, a processor, and a computer program stored in the memory, wherein, The processor executes the computer program to implement the steps of the method according to any one of claims 1-8.

10. A vehicle comprising: Environmental perception sensors are used to collect environmental perception data around the vehicle; A vehicle controller for performing the steps of the method as described in any one of claims 1-8.

Images