Control method, device and equipment of autonomous vehicle and storage medium

Abstract

The present disclosure provides a control method and device of an autonomous vehicle, equipment and a storage medium, relates to the technical field of computers, in particular to the technical fields of autonomous driving, data processing, artificial intelligence, computer vision and the like. The specific implementation scheme is as follows: in the process of executing a target autonomous driving task, an emergency parking plan is acquired when a network interruption is detected; the effectiveness of a parking stop in the emergency parking plan is verified based on a sensor of the autonomous vehicle; and the emergency parking plan is executed to control the autonomous vehicle to drive to the parking stop when the effectiveness verification is passed.

Description

Technical Field

[0001] This disclosure relates to the field of computer technology, and in particular to the fields of autonomous driving, data processing, artificial intelligence, and computer vision. Background Technology

[0002] In the field of autonomous driving, network communication is a crucial support for precise vehicle control and safe operation. Autonomous vehicles can obtain cloud data through networks and combine it with the perception results from onboard sensors to achieve stable driving.

[0003] However, network interruptions can easily occur while vehicles are in motion due to road conditions, equipment failures, etc., affecting driving safety and parking planning during periods of network unavailability. Summary of the Invention

[0004] This disclosure provides a control method, apparatus, device, and storage medium for an autonomous vehicle.

[0005] According to one aspect of this disclosure, a control method for an autonomous vehicle is provided, comprising: If a network interruption is detected during the execution of the target autonomous driving task, an emergency parking plan is obtained. Based on the sensors of autonomous vehicles, the validity of parking points in emergency parking planning is verified; If the validity verification passes, an emergency parking plan is executed to control the autonomous vehicle to drive to the parking point.

[0006] According to another aspect of this disclosure, a control device for an autonomous vehicle is provided, comprising: The acquisition module is used to acquire emergency parking plans when a network interruption is detected during the execution of the target autonomous driving task. The verification module is used to verify the validity of parking points in emergency parking planning based on the sensors of autonomous vehicles. The first execution module is used to execute emergency parking planning if the validity verification passes, so as to control the autonomous vehicle to drive to the parking point.

[0007] According to another aspect of this disclosure, an electronic device is provided, comprising: At least one processor; and The memory is communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform any of the methods described in the present disclosure.

[0008] According to another aspect of this disclosure, a non-transitory computer-readable storage medium is provided storing computer instructions, wherein the computer instructions are used to cause the computer to perform any of the methods according to embodiments of this disclosure.

[0009] According to another aspect of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements any of the methods according to embodiments of this disclosure.

[0010] According to another aspect of this disclosure, a vehicle is provided, including the electronic equipment provided in this disclosure.

[0011] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this disclosure, nor is it intended to limit the scope of this disclosure. Other features of this disclosure will become readily apparent from the following description. Attached Figure Description

[0012] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein: Figure 1 This is a schematic flowchart of a control method for an autonomous vehicle according to an embodiment of the present disclosure; Figure 2 This is a schematic flowchart illustrating the process of updating parameters in a target autonomous driving task according to an embodiment of the present disclosure; Figure 3 This is a schematic diagram of a process for reading an emergency parking plan from a target storage space according to an embodiment of the present disclosure; Figure 4 This is a schematic diagram of a process for obtaining an emergency parking plan in the event of a network interruption detected according to an embodiment of the present disclosure; Figure 5 This is a schematic diagram of the overall architecture of a control method for an autonomous vehicle according to an embodiment of the present disclosure; Figure 6 This is a flowchart illustrating a control method for an autonomous vehicle in a network state of network quality degradation but without disconnection, according to an embodiment of the present disclosure. Figure 7 This is a flowchart illustrating a control method for an autonomous vehicle in a network state of predictable network outage, according to an embodiment of the present disclosure. Figure 8 This is a flowchart illustrating a control method for an autonomous vehicle in the event of a sudden network interruption, according to an embodiment of the present disclosure. Figure 9 This is a schematic diagram of a process for real-time monitoring of whether the network of an autonomous vehicle has been restored, according to an embodiment of the present disclosure. Figure 10This is a schematic diagram of the structure of a control device for an autonomous vehicle according to an embodiment of the present disclosure; Figure 11 This is a block diagram of an electronic device used to implement the control method for an autonomous vehicle according to embodiments of the present disclosure. Detailed Implementation

[0013] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.

[0014] The terms “first,” “second,” etc., used in this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion, such as including a series of steps or units. A method, system, product, or apparatus is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or apparatuses.

[0015] It should be noted that, unless it is explicitly stated that there is a sequential order of execution between different operations, or that there is a sequential order of execution between different operations in terms of technical implementation, the execution order between multiple operations may not be significant, and multiple operations may be executed simultaneously.

[0016] In real-world driving scenarios, autonomous vehicles typically rely on onboard multi-source sensors, such as cameras, lidar, millimeter-wave radar, ultrasonic sensors, and IMUs (Inertial Measurement Units), to complete local environmental perception and decision-making. However, they also rely on 4G (4th-Generation Mobile Communication Technology) / 5G (5th-Generation Mobile Communication Technology) cellular networks and V2X (Vehicle to Everything) communication to achieve cloud-based collaborative perception, remote scheduling, task assignment, and operational monitoring.

[0017] In related technologies, the following handling methods are commonly used when a vehicle network malfunctions or experiences an interruption: (1) Exit the automatic driving mode and require manual intervention; (2) Trigger emergency braking, stop directly in the current lane, and wait for network restoration or rescue; (3) Drive along a preset fixed route (such as the route of an autonomous driving task) until the car stops, without considering the real-time status of the route and the stop point.

[0018] Among these, sudden braking caused by network interruption can turn a vehicle into a fixed obstacle on the road, which can easily lead to rear-end collisions and secondary accidents, especially in heavy traffic scenarios such as highways and urban main roads. The solution that requires manual intervention relies on the driver's real-time response ability, which is difficult to apply in fully autonomous driving scenarios and cannot respond to emergencies in a timely manner. The solution that follows a preset fixed route is also prone to failure in emergency response because it does not take into account whether the stop point is occupied, whether there are safety hazards in the surrounding area, or whether the route is impassable due to temporary construction or obstacles. It still has significant safety loopholes.

[0019] In view of this, the present disclosure provides a control method for an autonomous vehicle, which quickly obtains an emergency parking plan when the network is interrupted, and verifies the validity of the parking point based on on-board sensors, so that the autonomous vehicle can achieve safe and reliable emergency parking when the network quality is interrupted.

[0020] like Figure 1 The diagram shown is a flowchart illustrating the control method for an autonomous vehicle provided in this disclosure, including the following: S101, in the process of executing the target autonomous driving task, if a network interruption is detected, obtain an emergency parking plan.

[0021] The target autonomous driving task refers to the pre-set task that the autonomous vehicle needs to complete. The specific task content can be set according to the actual needs of the scenario, and this disclosure does not limit it. During the execution of the autonomous driving task, the autonomous vehicle will perform the autonomous driving task based on the perception of multiple source sensors.

[0022] A network outage indicates a failure in data transmission between the autonomous vehicle and an external communication network or cloud server, resulting in a broken transmission link and the inability to transmit data. For example, it may be impossible to obtain real-time dynamics of the autonomous driving mission's route from the network.

[0023] Emergency parking planning refers to the autonomous driving strategy generated to ensure driving safety in the event of a network outage. This emergency parking plan can include rationally planned safe parking spots and reference routes to those spots. Furthermore, while maintaining the current autonomous driving strategy, this emergency parking plan can, based on the vehicle's current location and surrounding environment information, use a pre-set algorithm model to plan safe parking spots and reference routes that can be executed offline in advance.

[0024] S102 verifies the validity of parking spots in emergency parking planning based on sensors from autonomous vehicles.

[0025] The designated parking spot refers to the geographical location specified in the emergency parking plan for the final parking of autonomous vehicles.

[0026] During implementation, autonomous vehicles can use their onboard sensors to acquire actual environmental data of the parking spot and its surroundings, and verify the effectiveness of the parking spot in the emergency parking plan in order to detect and evaluate whether the parking spot is suitable for autonomous vehicles to park in the current actual environment.

[0027] S103, if the validity verification passes, execute the emergency parking plan to control the autonomous vehicle to drive to the parking point.

[0028] In this embodiment, upon detecting a network outage, an emergency parking plan is obtained. This ensures that the autonomous vehicle has a clear response plan in the event of a network outage, avoiding the risk of the autonomous vehicle becoming a fixed obstacle due to stationary braking, and addressing the limitation of pre-set fixed routes in adapting to complex road conditions. By utilizing the vehicle's own sensors to verify the validity of the planned parking points, the safety and feasibility of the parking points are guaranteed, preventing new risks arising from unavailability due to changes in parking points. Executing the emergency parking plan after successful validity verification and controlling the vehicle to drive to the parking point reduces the adverse effects of potential network outages, enabling the autonomous vehicle to achieve relatively stable, reliable, and low-risk operation even in abnormal situations. Therefore, this disclosure implements a mechanism of pre-planning plus proximity verification, ensuring the safety and reliability of autonomous driving.

[0029] In this embodiment of the disclosure, the validity verification of the parking points in the emergency parking plan based on the sensors of the autonomous vehicle can be achieved through the following steps: Step A1: Within the set distance to the stop point, perform environmental modeling based on sensors to obtain environmental data; During implementation, the sensing area can be defined by extending a specific distance from the designated parking spot in the emergency parking plan towards the vehicle's current direction of travel. Then, onboard sensors collect environmental information about the parking spot and its surroundings within this area. Finally, this information is integrated to construct a three-dimensional environmental model containing details such as parking spot locations, obstacle distribution, and road boundaries.

[0030] Step A2: If the environmental data meets the following parking requirements, the parking spot validity verification is considered successful: (1) The stop was not occupied; In other words, the environmental data obtained through environmental modeling can demonstrate that the stop is not occupied by other vehicles, pedestrians, or obstacles. For example, if multi-source sensors detect other stationary vehicles in the stop area, or if the stop is blocked by a large object, then it means that the stop is occupied, and the validity check of the stop fails.

[0031] (2) The size of the parking spot allows for the parking of autonomous vehicles; Since different autonomous vehicles have different sizes, parking spots need sufficient space to accommodate them. Therefore, environmental data obtained through environmental modeling can indicate the size information of the parking spots, such as length, width, and height. If the size of the parking spot can accommodate the autonomous vehicle, the validity verification of the dimensions is confirmed.

[0032] (3) There are no obstructions on the reference path for the autonomous vehicle to reach the stop point; In other words, the environmental data obtained through environmental modeling can demonstrate that the reference path from the current location to the stop point has safe passage conditions. For example, there are no static obstacles such as construction barriers or traffic cones blocking the reference path, nor is there a risk of collision with dynamic obstacles such as vehicles changing lanes or pedestrians crossing the road. Under these circumstances, the validity verification of the reference path is confirmed to be successful.

[0033] During implementation, relevant information, such as the location and size of the docking point and information on surrounding obstacles, can be extracted from the environmental data obtained from environmental modeling to facilitate relevant verification.

[0034] (4) The process of the autonomous vehicle moving to the parking point satisfies dynamic constraints; Dynamic constraints refer to the limitations that autonomous vehicles must have in terms of motion parameters during the transition from their current state to a target state, ensuring that these parameters remain within the limits allowed by the vehicle's physical performance and safe operation.

[0035] Therefore, if the environmental data obtained through environmental modeling can show that the vehicle's trajectory and operation will not exceed the dynamic constraints during its movement from the current location to the stopping point, the validity verification of the stopping point is confirmed.

[0036] The dynamic constraints include at least one of the following: a. The curvature of the reference path of the autonomous vehicle is not greater than the curvature threshold; Curvature is a physical quantity used to describe the degree of curvature of a curve. In the context of autonomous vehicles, the curvature of the reference path indicates the degree of curvature along the path from the current location to the stop point. A larger curvature indicates a more curved path; a smaller curvature indicates a path closer to a straight line.

[0037] The curvature threshold is a preset maximum allowable value based on the physical characteristics and safety standards of autonomous vehicles. If the curvature of the reference path exceeds the curvature threshold, the autonomous vehicle may not be able to complete a turn safely. For example, an autonomous vehicle traveling at high speed may experience dangerous situations such as rollover due to excessive centrifugal force.

[0038] Therefore, when validating the stop point, it is ensured that the curvature at any point on the reference path from the current location to the stop point does not exceed a curvature threshold. This ensures that the autonomous vehicle can safely turn along the reference path during operation, avoiding safety issues caused by excessive curvature of the reference path.

[0039] b. The deceleration of the autonomous vehicle does not exceed the deceleration threshold.

[0040] Deceleration refers to the rate at which a vehicle's speed decreases per unit of time. For example, a vehicle experiences greater deceleration during emergency braking, while deceleration is relatively smaller during normal slowdown and stopping.

[0041] The deceleration threshold is a preset maximum allowable deceleration value. During the vehicle's journey from its current location to the stopping point, its deceleration is controlled to not exceed this threshold. This ensures that the autonomous vehicle operates smoothly and safely when decelerating and stopping, avoiding situations such as sudden braking.

[0042] In this embodiment, the curvature of the reference path is no greater than a curvature threshold, ensuring that the autonomous vehicle will not experience steering difficulties, loss of control, or collisions with surrounding objects due to an excessively curved path while driving to the parking point, thus ensuring the smoothness and operability of the autonomous vehicle's trajectory. The deceleration is no greater than a deceleration threshold, ensuring the stability and safety of the autonomous vehicle's driving. These dynamic constraints further enhance the feasibility and reliability of emergency parking planning, enabling the autonomous vehicle to park more safely and stably at the parking point in emergency situations such as network outages.

[0043] The above parking requirements ensure that the autonomous vehicle can park stably and safely at the designated parking location. In addition, in this embodiment of the disclosure, to further improve the rationality of the parking location, the parking requirements may also include the following condition (5).

[0044] (5) The location of the stop is compliant.

[0045] This involves verifying the compliance of parking spots. Environmental data obtained through environmental modeling can indicate whether the location of a parking spot falls within a no-parking zone. For example, the geographical location of a parking spot must comply with traffic regulations and safe driving standards, and must not be located in any area defined as a no-stopping zone by laws, regulations, or traffic signs. If a parking spot is located in a no-parking zone, the validity verification of the parking spot will fail.

[0046] During implementation, no-parking semantic rules can be set to verify whether parking spots are compliant. For example, no-parking semantic rules can be set to check whether parking spots are located in areas explicitly prohibited by traffic regulations and laws, whether they are on overpasses, inside tunnels, etc. By using no-parking semantic rules, non-compliant locations can be quickly filtered out, improving the efficiency and accuracy of parking spot location compliance verification.

[0047] In this embodiment, by using sensors to model the environment and acquire environmental data within a set distance from the parking point, the actual environmental conditions around the parking point can be accurately grasped. Parking requirements are set from multiple dimensions, including confirming that the parking point is unoccupied, of suitable size, with an unobstructed reference path, meeting dynamic constraints, and that the parking point's location is compliant, allowing for a comprehensive assessment of the parking point's availability and safety. If the environmental data meets the above parking requirements, the verification is passed, preventing autonomous vehicles from entering dangerous situations due to inappropriate parking point selection. This provides a reliable and safe parking location for autonomous vehicles in emergency situations such as network outages.

[0048] In this embodiment of the disclosure, while validating the parking points in the emergency parking plan, the autonomous driving task continues to be executed, and error restrictions are added to the autonomous driving task to improve driving safety. For example, the driving safety of the autonomous vehicle can be improved by updating the parameters in the target autonomous driving task. In implementation, it can be done as follows: Figure 2 As shown, it includes the following: S201, In the event of a network interruption, the vehicle speed and safe following distance in the autonomous driving task are updated based on preset control rules to obtain an updated autonomous driving task; wherein the updated vehicle speed is less than the original vehicle speed and the updated safe following distance is greater than the original safe following distance.

[0049] In other words, if a network interruption is detected, the autonomous vehicle will not immediately stop or lose control, but will update the vehicle speed and safe distance in the autonomous driving task through preset control rules.

[0050] During implementation, the vehicle speed in the autonomous driving task can be updated based on the current road type (e.g., highway, urban expressway, ordinary urban road), current vehicle speed and acceleration status, current lane speed limit information, network quality level, and vehicle autonomous driving safety configuration parameters. The minimum vehicle speed determined by the aforementioned factors can be selected as the upper limit of the vehicle speed in the autonomous driving task. The safe following distance is updated based on the current vehicle speed, current vehicle longitudinal control mode, network quality level, vehicle braking capacity, and road adhesion conditions.

[0051] The updated vehicle speed is lower than the previous speed. By reducing the speed, autonomous vehicles have more time to respond to unexpected situations, reducing the danger caused by missing information. At the same time, the updated safe following distance is greater than the previous safe following distance. Increasing the safe following distance provides autonomous vehicles with more reaction and braking space. Even if the vehicle in front suddenly slows down or an obstacle appears, the autonomous vehicle has enough distance to avoid a collision.

[0052] It should be noted that in the preset control rules, the longer the network interruption lasts and the higher the risk of network quality, the lower the speed of the autonomous vehicle after the update, and the greater the safe distance maintained.

[0053] S202, controls the autonomous vehicle to continue performing the updated autonomous driving task.

[0054] S203: If the validity verification of the stop point passes, stop executing the updated autonomous driving task.

[0055] Based on the above, once the validity of the designated stop is verified, the autonomous vehicle can smoothly transition from the updated autonomous driving task's driving path to the emergency parking planning path, thereby controlling the autonomous vehicle to drive to the stop.

[0056] In this embodiment, upon detecting a network interruption, the autonomous driving task is updated based on preset control rules, reducing vehicle speed and increasing the safe following distance, effectively mitigating the risks associated with network interruptions. Lower speeds allow the autonomous vehicle more reaction time to handle various emergencies, while a larger safe following distance provides a buffer, reducing the likelihood of collisions and allowing the vehicle to continue executing the updated autonomous driving task in a relatively safe state. Once the validity verification of the parking point in the emergency parking plan is successful, the updated autonomous driving task is stopped, and the vehicle is guided to a safe parking point, avoiding further uncertainties that might arise from prolonged operation during a network interruption and ensuring the vehicle can ultimately park smoothly and safely.

[0057] In this embodiment of the disclosure, emergency parking plans can be flexibly obtained by sensing the specific type of network interruption. Specifically, obtaining emergency parking plans when a network interruption is detected can be achieved based on the following steps: Step B1: Identify the type of network outage based on the network quality level; Network quality rating is a quantitative description of the current state of a network. Different network quality ratings reflect different levels of network health or condition.

[0058] Step B2: Obtain emergency parking plan based on the acquisition method corresponding to the interruption type of network interruption.

[0059] In other words, different strategies need to be adopted to obtain emergency parking plans for different types of network outages, so as to ensure that effective planning solutions can be obtained as much as possible in various situations.

[0060] In this embodiment of the disclosure, by first identifying the specific type of network interruption based on the network quality level, and then using the corresponding acquisition method according to different interruption types to obtain emergency parking plans, it is possible to more accurately and efficiently adapt to the current fault scenario when the network is abnormal, quickly generate an appropriate parking plan, improve the reliability and timeliness of emergency parking planning, and ensure the safety and stability of vehicle parking control in network interruption scenarios.

[0061] In this embodiment, the network quality of the autonomous vehicle can be periodically tracked and evaluated, and a network quality level can be given accordingly. This allows for the identification of whether the network outage is sudden or predictable based on the tracked network quality level in the event of an outage, and different emergency parking planning strategies can be obtained for different outage types. To improve the safety of the autonomous vehicle, the network quality can be continuously monitored during continuous driving, and the network status can be graded based on observable indicators. In implementation, identifying the type of network outage based on the network quality level can be achieved based on the following steps: Step C1: In the event of a network outage, obtain the currently stored network quality level; the network quality level is periodically evaluated and stored. In the event of a network outage, the network quality level, as last assessed before the outage, is immediately retrieved from the local cache.

[0062] During implementation, the network quality level is periodically assessed and stored. In each period, network quality-related indicators such as RSRP (Reference Signal Received Power), SINR (Signal to Interference plus Noise Ratio), latency, uplink and downlink rates, packet loss rate, heartbeat timeouts, and V2X link status are collected and calculated.

[0063] Subsequently, the aforementioned network metrics can be normalized in multiple dimensions and combined with preset weights to calculate a comprehensive network quality score. The weights of each metric can be dynamically configured based on vehicle operating scenarios, road types, or historical statistical data. Alternatively, a trained neural network model can adaptively determine the weights of each metric based on the network metrics and environmental data modeled by sensors.

[0064] In other embodiments, the weights of each indicator can also be obtained through historical operational data statistics, network quality models, or rule configuration methods. This solution is not limited to specific weight values ​​or calculation methods.

[0065] Next, the network quality score is mapped to the corresponding network quality level. In some embodiments, the specific determination logic is as follows: if the network quality score is greater than the normal network quality threshold, it is determined to be low risk, i.e., the network quality is normal; if the network quality score is greater than the network quality warning threshold but less than the normal network quality threshold, it is determined to be medium risk, hereinafter referred to as the first target level; if the network quality score is less than the network quality warning threshold, it is determined to be high risk, hereinafter referred to as the second target level.

[0066] Finally, the network quality level obtained from each assessment is stored in the target storage space of the autonomous vehicle. In practice, medium-risk and high-risk levels can be stored with the same label. However, medium-risk and high-risk levels will trigger different operations. For example, a medium-risk level will trigger the advance planning of a safe parking plan and store it in the target storage space to comprehensively utilize network information to improve the accuracy of the safe parking plan and maximize the reliability and accuracy of emergency parking planning. In the case of a high-risk level, not only will a safe parking plan be planned in real time, but restrictions on autonomous driving tasks (such as speed and distance limits) will also be triggered simultaneously to ensure driving safety. Specifically, in practice, under medium-risk and high-risk levels, the autonomous vehicle will periodically or in real time update the safe parking plan to ensure that feasible and reliable safe parking plans can be obtained in the event of a network outage.

[0067] Step C2: The currently stored risk level is used to indicate the network quality degradation situation. The interruption type of the network outage is determined and marked as a predictable outage. If the risk level read before the network outage is medium or high, it indicates that the network outage was preceded by warning. Therefore, this type of network outage can be marked as a predictable outage.

[0068] Step C3: The currently stored risk level is used to indicate that, under normal network quality conditions, the interruption type of the network outage is determined and marked as a sudden interruption.

[0069] If the risk level read before the network outage is low risk (i.e., the network is normal), it indicates that the network outage was a sudden and unexpected event. Therefore, the outage type can be marked as a sudden outage.

[0070] In this embodiment, the network quality level is periodically assessed and stored, enabling continuous tracking and recording of potential network risks. When a network outage occurs, the currently stored risk level is retrieved, and based on this level, it is determined whether network quality has deteriorated. The outage type is marked as either predictable or sudden, allowing for targeted measures to be taken based on the nature of the outage, thereby improving the safety of autonomous vehicles.

[0071] In this embodiment of the disclosure, the emergency parking plan is obtained based on the method of obtaining the interruption type corresponding to the network interruption, which can be implemented as follows: In the case of a network interruption of the predictable interruption type, read the emergency parking plan from the target storage space.

[0072] It should be noted that autonomous vehicles all possess the ability to generate emergency parking plans in real time solely based on their own sensors, serving as a fundamental guarantee for the safety of autonomous vehicle operation. However, to further enhance the safety of autonomous vehicles during operation, the corresponding emergency parking plans can be retrieved from the target storage space based on network status. This allows for timely and rapid acquisition of emergency parking plans and avoids safety hazards caused by the limited perception range of autonomous vehicle sensors.

[0073] In this embodiment of the disclosure, when the network interruption type is a predictable interruption, reading the emergency parking plan from the target storage space ensures that even in the event of network problems, the autonomous vehicle can still quickly execute the emergency parking operation according to the pre-stored plan. This minimizes the problem of not being able to obtain a reasonable emergency parking plan due to network interruption, effectively reduces the safety risks caused by network failures, and improves the stability and reliability of autonomous vehicle operation.

[0074] In the case of a predictable network outage, a safe parking plan is generated and stored in the target storage space, such as... Figure 3 As shown, it can be implemented as follows: S301, if the network quality level of the autonomous vehicle is at the first target level, select at least one candidate stopping point that meets the preset requirements within a first distance range on the map; the first distance range is greater than the distance threshold.

[0075] Based on the above, the first target level can be understood as the medium-risk level corresponding to a network quality score that is greater than the network quality warning threshold but less than the normal network quality threshold.

[0076] The first distance range is an area defined by a high-precision map and is greater than a pre-set distance threshold. This ensures that autonomous vehicles have sufficient buffer time for decision-making and planning, assuming the network quality level is at the first target level. At the first target level, more information can be obtained based on network conditions, such as long-distance high-precision maps and real-time traffic conditions along the driving route, all of which contribute to the rational and accurate planning of parking spots.

[0077] During implementation, the top N candidate parking spots can be selected within a first distance range based on their parking quality scores. In other words, at medium-risk levels, multiple candidate parking spots, both near and far, can be planned in advance, providing multiple safe parking options in case of network outages and improving the safety and reliability of autonomous driving.

[0078] S302: For each candidate parking spot, generate a corresponding safe parking plan, resulting in multiple safe parking plans.

[0079] For each selected candidate stop, the autonomous vehicle generates a corresponding safe parking plan based on its current location, driving status, traffic conditions, and other information. The safe parking plan typically includes information such as the reference path from the vehicle's current location to the stop, driving speed, and following distance, to ensure that the autonomous vehicle can safely and smoothly reach the stop.

[0080] During implementation, a safe parking plan can be generated for each candidate parking spot to obtain multiple safe parking plans.

[0081] S303 stores multiple safe parking plans in the target storage space; among them, the emergency parking plan is the safe parking plan that is closest to the autonomous vehicle among the multiple safe parking plans, or the safe parking plan with the highest quality.

[0082] Among them, selecting the safe parking plan corresponding to the candidate parking point closest to the current location of the autonomous vehicle as the emergency parking plan can enable the autonomous vehicle to reach the parking point as soon as possible when a network terminal appears, reducing the risks during the driving process.

[0083] It can also filter out the highest-quality safe parking plans from multiple safe parking plans as emergency parking plans. This allows autonomous vehicles to choose the optimal parking location and driving route when a network outage occurs, improving the safety, comfort, and compliance of the parking process.

[0084] The quality of each safe parking plan can be determined based on the parking quality score of the candidate parking spots it includes, which will be described later. Alternatively, it can be derived from the parking quality score and a pre-trained neural network model; this disclosure does not limit this approach.

[0085] In this embodiment of the disclosure, a corresponding safe parking plan is generated for each candidate parking point. This allows autonomous vehicles to prepare for emergencies in high network risk scenarios in advance. By selecting parking points within an appropriate range and formulating multiple parking plans, it can ensure that the vehicle can quickly call the optimal emergency parking solution when facing network problems. This ensures that the vehicle can park safely and orderly, reduces potential safety hazards caused by network risks, and improves the reliability and safety of autonomous driving.

[0086] S304, when the network quality level of the autonomous vehicle is the second target level, update the vehicle speed and safe distance in the autonomous driving task based on the preset control rules to obtain the updated autonomous driving task; execute the updated autonomous driving task.

[0087] Based on the above, the second target level can be understood as the high-risk level corresponding to a network quality score that is less than the network quality warning threshold.

[0088] Understandably, the network quality corresponding to the second target level is much lower than that corresponding to the first target level, which means that autonomous vehicles face a significantly increased risk of network outages.

[0089] During implementation, when the network quality level of the autonomous vehicle is at the second target level, not only will a reasonable safe parking plan be planned in real time and stored in the target storage space, but the speed and safe distance of the autonomous vehicle will also be limited based on preset control rules. After a formal network interruption, the autonomous vehicle will continue and strengthen the preset control rules, continuing to strictly constrain the driving status of the autonomous vehicle until the autonomous vehicle has completely completed the emergency stop.

[0090] In this embodiment of the disclosure, when the network quality level of the autonomous vehicle is the second target level, the vehicle speed and safe distance in the autonomous driving task are updated based on preset control rules. In the case of high network risk and interruption, this can provide more reaction time for the autonomous vehicle to deal with network problems, reduce safety hazards caused by network anomalies, and ensure that the vehicle can safely and orderly complete the emergency parking operation.

[0091] In some embodiments, emergency parking plans are obtained based on the acquisition method corresponding to the interruption type of the network interruption, such as... Figure 4 As shown, in the event of a sudden interruption, the autonomous vehicle will rely on its own sensors to perceive the environment and plan a reasonable emergency parking plan, including the following: S401, In the event of a sudden network interruption, the autonomous vehicle's surrounding environment is identified based on the sensors of the autonomous vehicle.

[0092] In the event of a sudden network outage, the system abandons reliance on maps and networks, relying solely on the autonomous vehicle's own sensors to collect and analyze real-time environmental information surrounding the vehicle.

[0093] S402, based on the surrounding environment of the autonomous vehicle, select candidate stopping points that meet preset requirements within a second distance range; the second distance range is less than or equal to a distance threshold.

[0094] The second distance range is a preset distance interval. When the second distance range is less than or equal to a distance threshold, it ensures that autonomous vehicles can find a suitable stopping point within a relatively short distance in the event of a sudden network interruption, reducing the risks associated with continuing to drive.

[0095] S403 generates emergency parking plans based on candidate parking spots.

[0096] During implementation, the basic trajectory can be kept unchanged, and the current vehicle speed and driving direction can be maintained, but with stricter safety restrictions. For example, the minimum speed can be selected from road speed limits, set speeds, and emergency thresholds as the emergency speed limit, and the expected following distance can be significantly increased.

[0097] Simultaneously, a parking plan that relies solely on local environment modeling is launched in parallel in a background thread. Once the plan is successful, the vehicle immediately and smoothly switches to the driving trajectory corresponding to the parking plan.

[0098] It should be noted that if the emergency parking plan is not read from the target storage space, or if the validity verification of the parking point fails, an emergency parking plan will also be generated based on the above steps S401-S403.

[0099] In this embodiment, the autonomous vehicle's sensors identify the surrounding environment, enabling timely and accurate acquisition of the vehicle's surroundings even in situations where network connectivity is unavailable and external information cannot be relied upon. Selecting candidate parking points that meet preset requirements within a second distance range less than or equal to a distance threshold ensures the vehicle can find a suitable parking location within a reasonable distance. These preset requirements guarantee the quality and applicability of the parking points. Finally, an emergency parking plan is generated based on the candidate parking points, allowing the autonomous vehicle to quickly and scientifically plan a safe and reliable parking solution in the event of a sudden network outage, ensuring the safety of the vehicle and surrounding traffic and minimizing the adverse effects of network interruption.

[0100] In this embodiment of the disclosure, the preset requirements for candidate stops include: the candidate stop belongs to a stop set, and the stop set includes at least one of the following: a. Dedicated parking space; This refers to areas specifically designated for vehicle parking.

[0101] b. Hard shoulder; Located on the right edge of a highway or expressway, this area is for vehicles to temporarily stop in case of an emergency.

[0102] c. Roadside parking area; These are parking areas set up on both sides of the road, which usually have clear parking markings and signs.

[0103] d. Service area; These are comprehensive service areas set up along highways and other road sections, with dedicated parking areas for vehicles to park for extended periods.

[0104] e. Gas station; This means providing a certain amount of parking space so that vehicles can temporarily stop before and after refueling.

[0105] f. Emergency lane; This lane is designated for emergency vehicles to use in emergency situations. Under special circumstances, autonomous vehicles may also temporarily park in the emergency lane if they meet the relevant regulations.

[0106] g. Position of least obstruction; The location of least obstruction is the position within a third distance range of an autonomous vehicle that has the least impact on traffic. For example, in congested traffic, there may be some open corners or positions near an autonomous vehicle that do not affect the normal driving of other vehicles; these positions can be considered as locations of least obstruction.

[0107] In this embodiment of the disclosure, the candidate parking points are defined as a set of parking points including dedicated parking spaces, hard shoulders, roadside parking areas, service areas, gas stations, emergency lanes, and minimal obstruction locations. They can be selected from multiple types of locations, taking into account both regular parking areas and areas where temporary parking is possible in special circumstances. In particular, the setting of minimal obstruction locations allows vehicles to find a place with the least impact on traffic in emergency or special situations, thereby achieving safe, efficient, and practical parking planning in different scenarios.

[0108] In this embodiment, the quality of parking spots can be quantified, and high-quality parking spots can be selected to complete the safe parking plan from the current location of the autonomous vehicle to the parking spot. Therefore, the aforementioned preset requirement may also include: the parking quality score of the candidate parking spot meets the target condition.

[0109] The parking quality score is a comprehensive score calculated based on multiple parking indicators, used to measure whether a candidate parking spot is suitable for parking.

[0110] The target condition is a pre-defined standard used to determine whether candidate parking spots meet parking requirements. In practice, it can be flexibly set according to different application scenarios, including but not limited to: selecting the candidate parking spots with the highest parking quality score, parking quality scores higher than a specific threshold, or the top N ranked candidates as the final parking spots.

[0111] For each candidate stop, a parking quality score can be generated based on at least one of the following parking indicators: (1) Safety indicators, used to measure the safety level of parking at candidate stops; In practice, the safety of parking at candidate parking spots can be measured based on the distance from the candidate parking spot to the nearest obstacle, the width of the available parking area, and whether the parking area is within the line of sight.

[0112] The farther away the obstacle, the wider the parking area, and the better the visibility, the higher the safety level and the higher the safety index score.

[0113] (2) Compliance indicators, used to measure whether candidate parking spots comply with parking regulations; During implementation, whether candidate parking spots comply with parking regulations can be determined by comparing high-precision map data or visual semantic recognition results to see if the candidate parking spots fall into preset no-parking zones. Examples include intersections, bus stops, tunnels, bridge sections, and sharp bends.

[0114] If a no-parking zone is entered, the score for compliance indicators will be significantly reduced or set to zero, depending on the penalty mechanism corresponding to each no-parking zone.

[0115] (3) Reachability index, used to measure the operational complexity of reaching candidate docking points; During implementation, the operational complexity of candidate stops can be calculated comprehensively based on the number of lane changes required to reach the candidate stop, the physical length of the reference path, and the estimated travel time.

[0116] The fewer lane changes, the shorter the path, and the shorter the estimated time, the lower the operational complexity and the higher the achievable score.

[0117] (4) Interference index, used to measure the degree of impact of parking at candidate stops on traffic; When implementing the policy, the impact of parking at candidate stops on traffic can be assessed based on whether the candidate stop occupies the main lane, is close to pedestrian-intensive areas, or affects the continuity of main traffic flow.

[0118] The less the disruption to traffic flow and the less the interference to pedestrians, the less the impact of parking at the stop on traffic, and the higher the score of the interference index.

[0119] (5) Recovery metrics, used to measure the probability of network recovery at candidate docking points and / or the probability of successful rescue.

[0120] In practice, the probability of restoring the network at a candidate stop can be measured based on the surrounding openness of the candidate stop (assessing signal obstruction) and the base station coverage density. The probability of successful rescue at a candidate stop can be measured based on assessing the physical accessibility of the rescue vehicle to that location.

[0121] During implementation, various characteristic parameters that affect network recovery and / or rescue success can be multiplied by their corresponding preset weight coefficients and then linearly accumulated to obtain the recovery index score of the final candidate docking point.

[0122] Alternatively, the raw information affecting network recovery and / or rescue success can be converted into input vectors, fed into a pre-trained neural network model for nonlinear mapping and feature extraction, and the neural network model can output recovery index scores for candidate docking points.

[0123] Alternatively, various raw information affecting network recovery and / or rescue success can be input into a pre-built decision tree model. The model will then traverse the path from top to bottom according to the feature judgment rules of the tree nodes until it reaches the leaf node. The recovery index score of the candidate docking point will be determined by the value corresponding to the leaf node.

[0124] In this embodiment, the scores of each of the above indicators can be multiplied by their corresponding preset weight coefficients and then linearly accumulated to obtain the parking quality score of the final candidate stop. The weight coefficients can be fixed or dynamically adjusted. For example, in an emergency situation of complete network outage, the weights of safety and recovery indicators can be increased, while the weights of interference indicators can be decreased, prioritizing vehicle and passenger safety and network recovery.

[0125] In this embodiment, a parking quality score is generated by comprehensively considering key indicators such as safety, compliance, accessibility, traffic interference, and recovery probability. This allows for the evaluation of each candidate parking location from different dimensions. Safety indicators ensure the safety of the vehicle and its surrounding environment when parking; compliance indicators ensure that parking behavior complies with relevant regulations and avoids the risk of violations; accessibility indicators help the vehicle select parking locations with lower operational complexity and easier accessibility; interference indicators reduce the impact of parking on normal traffic; and recovery indicators increase the likelihood of the vehicle reconnecting to the network or obtaining assistance after parking. The parking quality score generated by combining these indicators enables autonomous vehicles to select the most suitable parking location, thereby improving the safety, compliance, and overall efficiency of emergency parking.

[0126] Furthermore, in this embodiment of the disclosure, if the network interruption is of the predictable interruption type and the emergency parking plan cannot be read from the target storage space, the emergency parking plan can be obtained by adopting the handling measures under sudden interruption.

[0127] To prevent autonomous vehicles from failing to promptly detect network recovery after a network outage, thus avoiding prolonged periods of degraded operation, and to decouple the network self-recovery thread from the autonomous driving control logic, network scanning, carrier selection, and uplink / downlink reconstruction can be continuously executed in a background thread of the autonomous vehicle to monitor network recovery in real time. Upon detecting network recovery, the background thread can update the network status logically only, while the specific driving control logic can be determined based on the current behavior pattern. If the network is detected to be back to normal before implementing the emergency parking plan, continue to execute the target autonomous driving task.

[0128] In this embodiment of the disclosure, when the network is detected to have returned to normal asynchronously, the target autonomous driving task continues to be executed, which can avoid unnecessary stopping operations, enable the vehicle to continue to execute the target autonomous driving task, ensure the continuity and efficiency of the journey, and improve the flexibility and practicality of the vehicle's autonomous driving.

[0129] Once the autonomous vehicle reaches the designated stop and the network is detected to have returned to normal, the fault event is uploaded.

[0130] During implementation, the network status can be asynchronously checked. Once the network is confirmed to be back to normal, the autonomous vehicle can first upload information related to the fault event, such as the time, location, type, vehicle status, and perception and motion data before and after the fault, to a designated server or management platform. Subsequently, based on the current vehicle status and operational strategy, it autonomously decides whether to continue driving, change destinations, remain parked, or initiate manual takeover or remote handling procedures.

[0131] In this embodiment, upon detecting a return to normal network conditions, uploading the fault event ensures that the autonomous vehicle retains crucial fault scene information completely in the cloud immediately after network restoration, providing a reliable basis for subsequent fault analysis and system optimization. After uploading the fault information, the vehicle autonomously decides on subsequent actions based on its current state and operational strategy, enabling flexible responses. This ensures driving safety after a fault and maximizes resource allocation based on actual conditions, improving operational efficiency and ensuring the reliability and stability of autonomous driving.

[0132] In this embodiment of the disclosure, emergency parking planning to control the autonomous vehicle to drive to the parking point can be implemented based on the following steps: Step D1: Based on the reference path in the emergency parking plan, the autonomous vehicle generates operation instructions to reach the parking point using the sensor perception data and vehicle status of the autonomous vehicle. During implementation, a series of operational instructions can be calculated and generated by integrating the reference path, sensor perception data, and vehicle status. These operational instructions may include specific actions such as acceleration, deceleration, and steering of the autonomous vehicle to ensure that the vehicle can safely and accurately reach the stop point along the planned reference path.

[0133] Step D2: Based on the operation command, control the autonomous vehicle to park at the stop point.

[0134] In this embodiment of the disclosure, by integrating the reference path, sensor perception data and vehicle status, executable operation instructions are generated. This enables the autonomous vehicle to not only drive along the planned reference route when performing emergency parking, but also dynamically avoid surrounding obstacles and adapt to road conditions. This allows for safe and effective control of the autonomous vehicle to reach the parking point, thereby effectively improving the driving safety of the autonomous vehicle in the event of a network interruption.

[0135] In summary, the overall architecture diagram of the control method for autonomous vehicles provided in this disclosure embodiment is as follows: Figure 5As shown, the system includes a perception and input layer 51, a core processing layer 52, and an execution and post-processing layer 53. The perception and input layer 51 includes multi-source sensors 511, such as cameras, LiDAR, and millimeter-wave radar. The perception and input layer also includes a network unit 512 for network communication, including at least one communication module such as 4G / 5G / V2X. The core processing layer 52 includes a network status monitoring module 501, a network trend prediction module 502, an environment modeling module 503, a safe docking location decision module 504, a trajectory planning and control module 505, and a network self-recovery module 506.

[0136] The network status monitoring module 501 is primarily responsible for collecting network quality metrics, such as RSRP, SINR, latency, uplink and downlink speeds, packet loss rate, heartbeat timeout count, and V2X link status. Based on these metrics, it outputs enumerated values ​​for network quality levels, including: NET_NORMAL (normal network quality), NET_WARNING (degraded network quality but no disconnection), NET_LOST_PREDIC (predictable network outage), and NET_LOST_SUDDE (sudden outage).

[0137] The network trend prediction module 502 is responsible for predicting whether a network outage is likely to occur in the short term based on the aforementioned network indicator change trends. In practice, a trained neural network can be used to predict the duration of the network outage. Safe parking planning can be completed and stored within the predicted duration.

[0138] The environment modeling module 503 can perform multi-source perception fusion through vehicle-mounted sensors such as cameras, lidar, and millimeter-wave radar, and output an environment model containing environmental data such as drivable areas, lane structures, road boundaries, and obstacle distribution.

[0139] The safe parking location decision module 504 is responsible for pre-planning safe parking points based on map and environmental modeling results during the early warning phase, and for searching for safe parking points in real time based on onboard sensors in the event of a sudden network outage.

[0140] The trajectory planning and control module 505 includes a normal driving planning submodule and an emergency stop planning submodule. The normal driving planning submodule can output a trajectory to maintain the current autonomous driving task; the emergency stop planning submodule can output a trajectory for safe parking.

[0141] The network self-recovery module 506 is responsible for performing actions such as network reconnection and re-registration. It only notifies the system after recovery without altering the autonomous driving control logic.

[0142] like Figure 5The execution and post-processing layer 53 shown includes a vehicle actuator 531 and an event recording and reporting module 507. The vehicle actuator 531 is responsible for executing tasks based on the decisions of the core processing layer, such as continuing to perform corresponding steering / braking / driving operations based on the autonomous driving task, or performing corresponding control operations based on the emergency parking planned path.

[0143] The event logging and reporting module 507 is responsible for recording information such as the time of network anomaly, vehicle status, parking location and trajectory locally, and reporting the local record to the cloud after the network is restored.

[0144] Based on the above schematic diagram of the overall architecture of the autonomous vehicle control method, the key contents of the autonomous vehicle control method provided in this disclosure are as follows: (1) Continuously monitor the network quality of autonomous vehicles, and classify and judge the network status based on the monitoring results.

[0145] In implementation, the aforementioned network metrics can be normalized in multiple dimensions and combined with preset weights to calculate a comprehensive network quality score. The current network connection status is then determined based on this score. For example, the determination logic is as follows: When the network is normally connected, if the network quality score is greater than the normal network quality threshold, the current network status is determined to be normal; if the network quality score is greater than the network quality warning threshold but less than the normal network quality threshold, the current network status is determined to be degraded but not disconnected; if the network quality score is less than the network quality warning threshold, the current network status is determined to be a warningable network outage. When the network is disconnected, if the network status before the disconnection was normal, the current network status is determined to be a sudden outage.

[0146] (2) Read the current network status, determine whether the network status has changed, and match and switch to the corresponding driving behavior mode according to the current network status.

[0147] In this embodiment of the disclosure, the provided driving behavior modes may include: DRIVE_NORMAL (normal autonomous driving), DRIVE_EMERGENCY_PLAN (emergency driving planning), DRIVE_PULL_OVER (execute parallel parking), and DRIVE_PARKED_STANDBY (safely stopped, waiting).

[0148] (3) When the network status is normal, set the driving behavior mode of the autonomous vehicle to autonomous driving. For example, normal following, cruise, lane change planning, etc.

[0149] (4) When the network status is network quality degradation but no disconnection, set the behavior mode to emergency driving planning.

[0150] During implementation, the driving trajectory can be set as the basic trajectory, with only safety restrictions imposed, and safe stopping points and offline routes to those safe stopping points pre-planned simultaneously. The specific implementation process is as follows: Figure 6 As shown, it includes: S601, start network monitoring.

[0151] S602 collects key network metrics, including RSRP, SINR, latency, and packet loss rate.

[0152] S603, determine if the network quality score is below the warning threshold. If it is not below the warning threshold, maintain normal autonomous driving mode and return to step S602 to continue collecting network metrics. If it is below the warning threshold, continue to execute S604.

[0153] S604 changes the network status to "network quality degraded but not disconnected" and activates the early warning plan.

[0154] S605, adjust driving strategy.

[0155] For example, adjusting driving strategies may include actions such as reducing the maximum speed limit and increasing the safe following distance.

[0156] S606, Initiate environment modeling scan.

[0157] For example, scanning the surrounding environment to obtain information such as lanes, shoulders, service areas, and parking lots.

[0158] S607, Generate a set of candidate docking points.

[0159] In cases where network quality degrades but the connection remains intact, a set of candidate docking points is generated based on the results of environmental modeling and the high-precision map provided by the network.

[0160] S608 provides a comprehensive multi-dimensional score for candidate stopping points.

[0161] During implementation, candidate docking points can be scored from dimensions such as safety, compliance, accessibility, interference, and recovery.

[0162] S609, select a quality stop.

[0163] During implementation, the optimal safe stopping point can be selected and determined based on the scoring results.

[0164] S610, plan offline reference path.

[0165] S611, save the pre-planned scheme, enter standby mode, and wait for changes in network status.

[0166] (5) When the network status is predictably down, set the driving behavior mode to execute pull-over.

[0167] During implementation, the pre-planned scheme can be prioritized. The specific implementation process is as follows: Figure 7 As shown, it includes: S701, Network Status Change Monitoring.

[0168] During implementation, the network status is continuously monitored to determine whether the network is continuously deteriorating or recovering. If the network is continuously deteriorating, proceed to step S702; if the network recovers, proceed to step 710.

[0169] S702 prioritizes executing pre-planned solutions when the network status indicates a foreseeable network outage.

[0170] S703, activate the pre-planned scheme and drive along the reference path in the pre-planned scheme.

[0171] During implementation, obstacle avoidance and speed fine-tuning can be performed in real time by using a reference path as a guide.

[0172] S704, perform safe docking operation.

[0173] For example, perform safe parking maneuvers such as pulling over or slowing down, while simultaneously activating the vehicle's hazard lights.

[0174] S705, the vehicle has safely stopped and is waiting for the network to be restored.

[0175] This means that after the vehicle has safely stopped, it enters a state of waiting for the network to be restored.

[0176] S706 continuously attempts to reconnect to the network and sends heartbeat checks.

[0177] During implementation, network reconnection attempts can be initiated periodically, and heartbeat detection packets can be sent to the network side.

[0178] S707 records network failure events and corresponding countermeasures.

[0179] For example, record the time of occurrence, duration, and response actions taken in response to the network failure.

[0180] S708, Determine if the network has been restored. If the network has been restored, proceed to step S709; if the network has not been restored, proceed to step S713.

[0181] S709: Report the complete event log and restore normal connection. Then proceed to step S712.

[0182] S710, network quality has returned to normal.

[0183] S711 changed the network status to normal network quality and resumed normal driving.

[0184] S712, end the pre-planning scheme and return to normal monitoring of network indicators.

[0185] S713, report the emergency and await human intervention.

[0186] During implementation, network failure emergencies can be reported to the backend or operations and maintenance personnel, and manual intervention instructions can be awaited.

[0187] S714, maintain a safe docking status and continuously monitor.

[0188] During implementation, the network status can be continuously monitored after the vehicle is safely parked until the network is restored or a manual intervention instruction is received.

[0189] (6) In the event of a sudden network interruption, set the behavior mode to execute pull-over parking.

[0190] During implementation, the system can first operate without interruption based on the basic trajectory, and then generate emergency stopping plans in real time. The specific implementation process is as follows: Figure 8 As shown, it includes: S801, a network link failure was detected.

[0191] S802, determine whether the previous network status was an alarm status, i.e., medium risk level or high risk level. If yes, continue to step S803; if no, continue to step S814.

[0192] S803 attempts to read the pre-planned scheme from the target storage space.

[0193] S804, determine if the pre-planned solution is valid. If valid, continue to step S805; if invalid, proceed to step S806.

[0194] Whether it is effective can also include whether the stopping points of the pre-planned scheme are within the set range ahead of the vehicle's driving path.

[0195] S805, execute the roadside driving trajectory in the pre-planned scheme, and after parking, continue to execute steps S811-S813.

[0196] S806, revert to real-time planning mode.

[0197] S807, re-search for safe docking points in real time.

[0198] S808 searches for safe docking points and assesses risks in real time.

[0199] S809 generates real-time edge-keeping trajectories, taking into account vehicle dynamic constraints.

[0200] S810 performs operations such as lane changing, deceleration, and pulling over.

[0201] S811, safely stop the vehicle and activate the warning lights.

[0202] S812 reports network outage events and records parking locations.

[0203] S813, continuous network reconnection has entered a waiting state.

[0204] S814 changes the network status to sudden network interruption.

[0205] S815 maintains the current driving strategy, imposing stricter speed limits and extremely high safe following distances.

[0206] S816 generates emergency parking plans based on candidate parking spots.

[0207] During implementation, the basic trajectory can be maintained initially, keeping the current vehicle speed and direction of travel constant but with stricter safety restrictions. For example, the minimum speed can be selected from road speed limits, set speeds, and emergency thresholds as the emergency speed limit, and the expected following distance can be significantly increased. Simultaneously, a parking plan that relies solely on local environment modeling can be started in parallel in a background thread. Once the planning is successful, the vehicle can immediately and smoothly switch to the driving trajectory corresponding to the parking plan. Steps S811-S813 can then be executed.

[0208] (7) Continuously perform network scanning, operator selection, uplink and downlink reconstruction and other actions in the background thread of the autonomous vehicle to monitor whether the network of the autonomous vehicle has been restored in real time.

[0209] During implementation, a process is implemented to monitor in real time whether the network for autonomous vehicles has been restored, such as... Figure 9 As shown, it includes: S901, initiates the network self-recovery thread.

[0210] S902 detects the current network status.

[0211] S903, Determine if the network is connected. If yes, proceed to step S904. If no, proceed to step S907.

[0212] S904, the network status has been updated to normal.

[0213] S905 sends a recovery notification to the main control system.

[0214] S906, restore network-related functions, return to execute S901.

[0215] S907, scan for available networks.

[0216] S908, attempting network registration / attachment.

[0217] S909, determine if registration was successful. If not, wait for a retry interval (t=2ⁿ seconds), then return to step S902 to re-check the network connection status. If yes, continue to step S910.

[0218] S910 establishes uplink and downlink data connections and activates network sessions.

[0219] S911, verify network connectivity (Ping test). Then return to step S902 to continuously monitor link status.

[0220] In summary, the control method for autonomous vehicles provided in this disclosure has the following beneficial effects: 1. Avoid unnecessary stopping due to network interruptions: The vehicle will not brake suddenly in the middle of the lane due to short-term network fluctuations or interruptions. Instead, it will find a safe stopping point through advance planning or real-time visual modeling while maintaining autonomous driving.

[0221] 2. Improve safety and driving continuity: Through network trend analysis, state classification and emergency behavior mode design, vehicles can maintain controllable and continuous driving when the network is unavailable, reducing the risk of secondary accidents.

[0222] 3. Enables fully offline safe parking capability: Even if the network is unavailable, the vehicle can still complete the safe parking or emergency stopping process completely offline based on local vision and environmental modeling results.

[0223] 4. Enhanced system architecture robustness: The network self-recovery thread is decoupled from the autonomous driving control logic to avoid blocking vehicle safety decisions and execution due to network module anomalies.

[0224] 5. It has good engineering feasibility and scalability: This method can be deployed on roads of different grades (urban roads, highways, elevated roads, tunnels, etc.) and can also be integrated with existing high-precision maps and cloud control platforms.

[0225] Based on the same technical concept, this disclosure also provides a control device 1000 for an autonomous vehicle, such as... Figure 10 As shown, it includes: The acquisition module 1001 is used to acquire an emergency parking plan when a network interruption is detected during the execution of the target autonomous driving task. The verification module 1002 is used to verify the validity of the parking points in the emergency parking plan based on the sensors of the autonomous vehicle. The first execution module 1003 is used to execute the emergency parking plan when the validity verification passes, so as to control the autonomous vehicle to drive to the parking point.

[0226] In some embodiments, the verification module is configured to: Within a set distance from the stopping point, environmental modeling is performed based on sensors to obtain environmental data; The validity check of the parking spot is deemed successful if the environmental data meets the following parking requirements: The stop was not occupied; The size of the docking point allows for the parking of autonomous vehicles; There are no obstructions on the reference path for the autonomous vehicle to reach the stop point; The process of an autonomous vehicle moving to a stopping point satisfies dynamic constraints.

[0227] In some embodiments, the parking requirement further includes: compliance with the location of the parking spot.

[0228] In some embodiments, the dynamic constraints include at least one of the following: The curvature of the reference path for autonomous vehicles is no greater than a curvature threshold. The deceleration of autonomous vehicles shall not exceed the deceleration threshold.

[0229] In some embodiments, an update module is also included, for: In the event of a network interruption, the vehicle speed and safe following distance in the autonomous driving task are updated based on preset control rules to obtain an updated autonomous driving task; wherein the updated vehicle speed is lower than the original vehicle speed and the updated safe following distance is greater than the original safe following distance. To control the autonomous vehicle to continue performing the updated autonomous driving task; If the validity of the stop point is verified, the updated autonomous driving task will be stopped.

[0230] In some embodiments, the acquisition module includes: The identification unit is used to identify the interruption type of the network interruption based on the network quality level; The acquisition unit is used to acquire the emergency parking plan based on the acquisition method corresponding to the interruption type of the network interruption.

[0231] In some embodiments, the acquiring unit is specifically used for: In the case of a predictable network interruption, the emergency parking plan is read from the target storage space.

[0232] In some embodiments, a pre-planning module is also included to generate a safe parking plan stored in the target storage space based on the following method: If the network quality level of the autonomous vehicle is at the first target level, at least one candidate docking point that meets the preset requirements is selected within a first distance range on the map; the first distance range is greater than the distance threshold. For each candidate parking spot, a corresponding safe parking plan is generated, resulting in multiple safe parking plans; Multiple safe parking plans are stored in the target storage space; among them, the emergency parking plan is the safe parking plan that is closest to the autonomous vehicle or the highest quality safe parking plan among the multiple safe parking plans.

[0233] In some embodiments, a task control module is also included: When the network quality level of the autonomous vehicle is the second target level, the vehicle speed and safe distance in the autonomous driving task are updated based on the preset control rules to obtain the updated autonomous driving task. Perform the updated autonomous driving task.

[0234] In some embodiments, the acquiring unit is specifically used for: In the event of a sudden network outage, the autonomous vehicle's sensors are used to identify the surrounding environment of the autonomous vehicle. Based on the surrounding environment of the autonomous vehicle, candidate parking points that meet preset requirements are selected within a second distance range; the second distance range is less than or equal to a distance threshold. Emergency parking plans are generated based on candidate parking spots.

[0235] In some embodiments, the preset requirements include: Candidate stops are stops that belong to the stop set, and the stop set includes at least one of the following: Designated parking spaces, hard shoulders, roadside parking areas, service areas, gas stations, emergency lanes, and locations with minimal obstruction; the location with minimal obstruction is the position within the third distance range of an autonomous vehicle that has the least impact on traffic.

[0236] In some embodiments, the preset requirement also includes: the parking quality score of the candidate stop meets the target condition; It also includes a quality assessment module, which generates a parking quality score for each candidate stop based on at least one of the following parking indicators: Safety indicators are used to measure the safety level of parking at candidate stops; Compliance metrics are used to measure whether candidate parking spots comply with parking regulations; Reachability metrics are used to measure the operational complexity of reaching a candidate stop. The interference index is used to measure the impact of parking at the candidate stops on traffic. Recovery metrics are used to measure the probability of restoring the network at candidate docking points and / or the probability of a successful rescue.

[0237] In some embodiments, the identification unit is specifically used for: In the event of a network outage, retrieve the currently stored network quality level; the network quality level is periodically evaluated and stored. The current stored risk level is used to indicate the type of network outage that is identified as a predictable outage in the event of a deterioration in network quality. The current risk level is used to indicate that, under normal network quality conditions, the interruption type of network outage is marked as a sudden interruption.

[0238] In some embodiments, a second execution module is further included, configured to: If the network is detected to be back to normal before implementing the emergency parking plan, the target autonomous driving task will continue.

[0239] In some embodiments, an upload module is also included, for: Once the autonomous vehicle arrives at the designated stop and the network is detected to be back to normal, it uploads the fault event.

[0240] In some embodiments, the first execution module includes: The instruction generation unit is used to generate operation instructions to reach the parking point based on the reference path in the emergency parking plan, the perception data of the autonomous vehicle's sensors, and the vehicle status. The execution unit is used to control the autonomous vehicle to park at the designated stop point based on operation commands.

[0241] The specific functions and examples of each module and submodule of the apparatus in this disclosure can be found in the relevant descriptions of the corresponding steps in the above method embodiments, and will not be repeated here.

[0242] The acquisition, storage, and application of user personal information involved in the technical solution disclosed herein comply with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0243] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.

[0244] Figure 11A schematic block diagram of an example electronic device 1100 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital assistants, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0245] like Figure 11 As shown, device 1100 includes a computing unit 1101, which can perform various appropriate actions and processes according to a computer program stored in read-only memory (ROM) 1102 or a computer program loaded into random access memory (RAM) 1103 from storage unit 1108. The RAM 1103 may also store various programs and data required for the operation of device 1100. The computing unit 1101, ROM 1102, and RAM 1103 are interconnected via bus 1104. Input / output (I / O) interface 1105 is also connected to bus 1104.

[0246] Multiple components in device 1100 are connected to I / O interface 1105, including: input unit 1106, such as keyboard, mouse, etc.; output unit 1107, such as various types of monitors, speakers, etc.; storage unit 1108, such as disk, optical disk, etc.; and communication unit 1109, such as network card, modem, wireless transceiver, etc. Communication unit 1109 allows device 1100 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.

[0247] The computing unit 1101 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 1101 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special-purpose artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 1101 performs the various methods and processes described above, such as control methods for autonomous vehicles. For example, in some embodiments, the control methods for autonomous vehicles can be implemented as computer software programs tangibly contained in a machine-readable medium, such as storage unit 1108. In some embodiments, part or all of the computer program can be loaded and / or installed on device 1100 via ROM 1102 and / or communication unit 1109. When the computer program is loaded into RAM 1103 and executed by the computing unit 1101, one or more steps of the control methods for autonomous vehicles described above can be performed. Alternatively, in other embodiments, the computing unit 1101 may be configured to perform control methods for an autonomous vehicle by any other suitable means (e.g., by means of firmware).

[0248] Various embodiments of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field-programmable gate arrays (FPGAs), application-specific integrated circuits (ASICs), application-specific standard products (ASSPs), systems-on-a-chip (SoCs), payload-programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.

[0249] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.

[0250] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.

[0251] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).

[0252] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as a data server), or computing systems that include middleware components (e.g., an application server), or computing systems that include frontend components (e.g., a user computer with a graphical user interface or web browser through which a user can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., a communication network). Examples of communication networks include local area networks (LANs), wide area networks (WANs), and the Internet.

[0253] Computer systems can include clients and servers. Clients and servers are generally located far apart and typically interact via communication networks. Client-server relationships are created by computer programs running on the respective computers and having a client-server relationship with each other. Servers can be cloud servers, servers in distributed systems, or servers incorporating blockchain technology.

[0254] Based on the aforementioned electronic devices, this disclosure also provides a vehicle that may include electronic devices, and may also include communication components, a display screen for realizing a human-machine interface, and an information collection device for collecting information about the surrounding environment, etc., wherein the communication components, the display screen, the information collection device and the electronic devices are communicatively connected.

[0255] According to embodiments of this disclosure, the electronic device can be integrated with the communication component, display screen, and information acquisition device, or it can be separately configured with the communication component, display screen, and information acquisition device.

[0256] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this disclosure can be achieved, and this is not limited herein.

[0257] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the principles of this disclosure should be included within the scope of protection of this disclosure.

Claims

1. A control method for an autonomous vehicle, comprising: If a network interruption is detected during the execution of the target autonomous driving task, an emergency parking plan is obtained. The validity of the parking points in the emergency parking plan is verified based on the sensors of the autonomous vehicle. If the validity verification passes, the emergency parking plan is executed to control the autonomous vehicle to drive to the parking point.

2. The method according to claim 1, wherein, The sensors based on the autonomous vehicle perform validity verification on the parking points in the emergency parking plan, including: Within a set range from the designated stop point, environmental modeling is performed based on the sensors to obtain environmental data; The validity check of the parking spot is deemed successful if the environmental data meets the following parking requirements: The stop was not occupied; The dimensions of the docking point allow the autonomous vehicle to be parked; There are no obstructions on the reference path for the autonomous vehicle to reach the stop point; The process of the autonomous vehicle moving to the stopping point satisfies dynamic constraints.

3. The method according to claim 2, wherein, The parking requirements also include that the location of the parking spot is compliant.

4. The method according to claim 2, wherein, The dynamic constraints include at least one of the following: The curvature of the reference path of the autonomous vehicle is not greater than a curvature threshold. The deceleration of the autonomous vehicle does not exceed the deceleration threshold.

5. The method according to claim 1, further comprising: In the event of a network interruption, the vehicle speed and safe following distance in the autonomous driving task are updated based on preset control rules to obtain an updated autonomous driving task; wherein the updated vehicle speed is lower than the original vehicle speed and the updated safe following distance is greater than the original safe following distance. Control the autonomous vehicle to continue performing the updated autonomous driving task; If the validity verification of the stop point passes, the updated autonomous driving task is stopped.

6. The method according to claim 1, wherein, In the event of a detected network outage, the process of obtaining an emergency parking plan includes: Identify the interruption type of the network outage based on the network quality level; The emergency parking plan is obtained based on the acquisition method corresponding to the interruption type of the network interruption.

7. The method according to claim 6, wherein, The method for obtaining the emergency parking plan based on the interruption type corresponding to the network interruption includes: In the case of a predictable network interruption, the emergency parking plan is read from the target storage space.

8. The method of claim 7, further comprising generating a safe parking plan stored in the target storage space based on the following method: If the network quality level of the autonomous vehicle is at the first target level, at least one candidate stopping point that meets the preset requirements is selected within a first distance range on the map; the first distance range is greater than a distance threshold. For each candidate parking spot, a corresponding safe parking plan is generated, resulting in multiple safe parking plans; The multiple safe parking plans are stored in the target storage space; wherein, the emergency parking plan is the safe parking plan that is closest to the autonomous vehicle among the multiple safe parking plans, or the safe parking plan with the highest quality.

9. The method according to claim 7, further comprising: When the network quality level of the autonomous vehicle is the second target level, the vehicle speed and safe distance in the autonomous driving task are updated based on the preset control rules to obtain the updated autonomous driving task. Perform the updated autonomous driving task.

10. The method according to claim 6, wherein, The method for obtaining the emergency parking plan based on the interruption type corresponding to the network interruption includes: In the event of a sudden network interruption, the surrounding environment of the autonomous vehicle is identified based on the sensors of the autonomous vehicle. Based on the surrounding environment of the autonomous vehicle, candidate stopping points that meet preset requirements are selected within a second distance range; the second distance range is less than or equal to a distance threshold. The emergency parking plan is generated based on the candidate parking spots.

11. The method according to claim 8 or 10, wherein, The preset requirements include: The candidate stop point is a stop point that belongs to the stop point set, which includes at least one of the following: Designated parking spaces, hard shoulders, roadside parking areas, service areas, gas stations, emergency lanes, and locations with minimal obstruction; the location with minimal obstruction is the position within a third distance range of the autonomous vehicle that has the least impact on traffic.

12. The method according to claim 11, wherein the preset requirement further includes: The parking quality scores of the candidate stops meet the target conditions; The method further includes: for each candidate stop, generating a parking quality score for the candidate stop based on at least one of the following parking indicators: Safety indicators are used to measure the safety level of parking at the candidate stop points; Compliance metrics are used to measure whether the candidate parking spots comply with parking regulations; The reachability metric is used to measure the operational complexity of reaching the candidate docking points. The interference index is used to measure the impact of parking at the candidate stops on traffic. Recovery metrics are used to measure the probability of network recovery and / or successful rescue at the candidate docking point.

13. The method according to claim 6, wherein, The step of identifying the interruption type of the network outage based on the network quality level includes: In the event of a network outage, retrieve the currently stored network quality level; the network quality level is periodically evaluated and stored. The currently stored risk level is used to indicate the type of network outage as a predictable outage in the event of a deterioration in network quality. The currently stored risk level is used to determine the network outage type as a sudden outage when the network quality is normal.

14. The method according to claim 1, further comprising: If the network is detected to have returned to normal before the emergency parking plan is executed, the target autonomous driving task will continue to be executed.

15. The method according to claim 1, further comprising: After the autonomous vehicle arrives at the designated stop and detects that the network has returned to normal, it uploads the fault event.

16. The method according to claim 1, wherein, The execution of the emergency parking plan to control the autonomous vehicle to drive to the parking point includes: Based on the reference path in the emergency parking plan, the autonomous vehicle generates operation instructions to reach the parking point based on the sensor perception data and vehicle status. Based on the operation command, the autonomous vehicle is controlled to park at the stop point.

17. A control device for an autonomous vehicle, comprising: The acquisition module is used to acquire emergency parking plans when a network interruption is detected during the execution of the target autonomous driving task. The verification module is used to verify the validity of the parking points in the emergency parking plan based on the sensors of the autonomous vehicle. The first execution module is used to execute the emergency parking plan if the validity verification passes, so as to control the autonomous vehicle to drive to the parking point.

18. An electronic device comprising: At least one processor; as well as A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-16.

19. A non-transitory computer-readable storage medium storing computer instructions, wherein, The computer instructions are used to cause the computer to perform the method according to any one of claims 1-16.

20. A computer program product comprising a computer program that, when executed by a processor, implements the method according to any one of claims 1-16.

21. An autonomous vehicle, including the electronic equipment as claimed in claim 18.

Images