A connected autonomous driving method and cloud control platform based on vehicle-road-cloud integration

Abstract

The present invention provides a connected autonomous driving method based on vehicle-road-cloud integration. This method deploys the algorithmic functions of connected autonomous driving on a cloud platform, leveraging the platform's computing power and resource advantages to compensate for the insufficient resources of a single vehicle's computing unit. It can also handle the concurrent scenarios of multiple connected vehicles in parallel. By aligning the connected autonomous driving functional architecture deployed on the cloud platform with that of a single vehicle, it can simultaneously provide cloud-controlled perception, decision-making, and planning results, providing comprehensive data support for autonomous driving vehicles. This method can achieve multi-vehicle collaboration within the same scenario, with the cloud platform simultaneously sending decision-making and planning recommendations to all relevant vehicles. This can improve the performance and safety margins of single-vehicle autonomous driving, reduce the vehicle's perception, decision-making, and computational costs, and resolve the common problem of conflicting individual decisions among multiple vehicles in autonomous driving.

Description

Technical Field

[0001] The present invention relates to the field of autonomous driving technology, and more specifically, to a connected autonomous driving method and cloud control platform based on vehicle-road-cloud integration. Background Art

[0002] Vehicles are becoming increasingly intelligent, and autonomous driving is becoming part of our lives. Research in the field of autonomous driving is primarily focused on vehicle-road collaboration. Road information acquired through roadside sensing is processed in the cloud, or sent directly to the vehicle via roadside communication equipment (mostly RSUs) without cloud processing. The vehicle's controller then controls the vehicle's driving decisions.

[0003] Because the decision-making and calculations of existing vehicle-road-cloud collaborative solutions are mostly implemented by the vehicle as the ultimate implementer, all information is aggregated and calculated on the vehicle, which increases the vehicle's computing cost. In addition, existing vehicle-road-cloud collaborative solutions focus on the vehicle's own computing power and the interaction between individual vehicles, lacking the ability to coordinate multi-vehicle perception, decision-making and planning, and are unable to systematically solve the problem of multi-vehicle conflicts in traffic. At the same time, this focus leads to a low level of cloud platform participation, and the level of perception fusion and decision-making planning for global driving is not high, resulting in low global awareness and efficiency of autonomous driving. In addition, because the operating systems and architectures of the cloud and vehicle sides of existing vehicle-road-cloud collaborative solutions are usually two different systems, heterogeneous data restricts the convenience of implementing connected autonomous driving for connected vehicles. Summary of the Invention

[0004] The present invention provides a connected autonomous driving method and system based on vehicle-road-cloud integration to overcome at least one technical problem existing in the prior art.

[0005] In a first aspect, the present invention provides a connected autonomous driving method based on vehicle-road-cloud integration, which is applied to a cloud control platform, wherein the cloud control platform communicates with roadside sensors. The connected autonomous driving method includes:

[0006] Determine the area to be planned and lane position information of each monitored road in the area to be planned in a pre-prepared high-precision map;

[0007] Obtain environmental status information returned by roadside sensors on each monitored road and current driving information returned by vehicle sensors of vehicles traveling on the monitored roads;

[0008] Determine target information from environmental status information;

[0009] The environmental status information includes target information, environmental information, and current traffic event information, and the targets include pedestrians and vehicles;

[0010] The target information is matched and fused sequentially to form a continuous recognition sequence;

[0011] Generate a global planning path for each vehicle based on current traffic event information, environmental information, and current driving information;

[0012] Make decisions based on current traffic event information, environmental information, and the global planned path to generate the vehicle's current driving decision;

[0013] Based on the global path information and the current driving decision, a local path plan for vehicle driving is generated and sent to the vehicle to realize the vehicle's networked autonomous driving.

[0014] Among them, the lane position information includes the specific position of each lane in the high-precision map, the target information includes the specific position and status of the target in the high-precision map, the roadside sensors include video and radar sensors, and the environmental status information is video data information and / or radar data information.

[0015] Optionally, the sequentially performing target matching and fusion on the target information to form a continuous recognition sequence includes:

[0016] According to the order of each roadside sensor in the monitored road, the target information is matched;

[0017] Forming the matched target information into a sequence;

[0018] The target information of the sequences formed by complete overlap and partial overlap is deduplicated to form an identification sequence.

[0019] The current driving information includes the driving destination and the driving trend.

[0020] Optionally, generating a global planned path for each vehicle based on current traffic event information, environmental information, and current driving information includes:

[0021] For each vehicle, determine the target lanes from the target location to the destination;

[0022] Determine whether a traffic event occurs on each target lane based on current traffic event information;

[0023] If a traffic incident occurs in a certain target lane, determine the location of the traffic incident in the high-precision map;

[0024] Based on environmental information and current driving information, predict the next state of the vehicle and whether a traffic incident may occur in a target lane in the next state;

[0025] When a traffic event is determined based on the current traffic event information and the predicted traffic event occurs in a target lane in the next state, a global planning path for the vehicle is generated.

[0026] Optionally, the environmental information includes traffic command information and traffic display information on the road, and predicting the next state of the vehicle and whether a traffic incident may occur in a target lane in the next state based on the environmental information and current driving information includes:

[0027] Predict the vehicle's next state based on traffic control information, traffic display information, and current driving information;

[0028] Whether a traffic incident may occur in the next state of each vehicle is determined based on the driving information of the next state of each vehicle.

[0029] Optionally, making a decision based on current traffic event information, environmental information, and the global planned path to generate a current driving decision of the vehicle includes:

[0030] The current traffic event information, environmental information and the global planning path are used as inputs of a preset decision model so that the decision model can perceive the environmental information and the current traffic event information, and make a current driving decision indicating the next state driving direction and driving route at the current position based on the global planning path.

[0031] Optionally, after generating the current driving decision of the vehicle, the connected autonomous driving method further includes:

[0032] The current driving decision is sent to the vehicle's control unit so that the vehicle can execute according to the current driving decision.

[0033] In a second aspect, the present invention provides a cloud control platform that executes the networked autonomous driving method based on vehicle-road-cloud integration described in the first aspect.

[0034] The innovative features of the embodiments of the present invention include:

[0035] 1. The present invention provides a connected autonomous driving method based on vehicle-road-cloud integration. It deploys the algorithm functions of the connected autonomous driving on a cloud platform, leveraging the platform's computing power and resource advantages to compensate for the insufficient resources of a single vehicle's computing unit. It can also handle the concurrent scenarios of multiple connected vehicles in parallel, which is also one of the innovations of the embodiments of the present invention.

[0036] 2. The present invention provides a connected autonomous driving method based on vehicle-road-cloud integration. By making the connected autonomous driving functional architecture deployed on the cloud platform the same as that of a single vehicle, it can simultaneously provide cloud-controlled perception, decision-making, and planning results, providing comprehensive data function support for autonomous driving vehicles. This is also one of the innovations of the embodiments of the present invention.

[0037] 3. The present invention provides a connected autonomous driving method based on vehicle-road-cloud integration, which can achieve multi-vehicle collaboration in the same scenario. The cloud platform simultaneously sends decision-making and planning suggestions to all relevant vehicles, and different vehicles simultaneously execute connected functional responses, which can solve the global objective optimization problem of individual autonomous driving vehicles. This is also one of the innovative features of the embodiments of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0039] Figure 1 A schematic diagram of a process flow of a connected autonomous driving method based on vehicle-road-cloud integration provided by an embodiment of the present invention;

[0040] Figure 2 A schematic diagram of a cloud platform provided in an embodiment of the present invention;

[0041] Figure 3 A flowchart of a connected autonomous driving method based on vehicle-road-cloud integration is provided for a specific embodiment of the present invention;

[0042] Figure 4 This is a schematic diagram of the interaction between multiple vehicles and a cloud control platform according to an embodiment of the present invention. DETAILED DESCRIPTION

[0043] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without creative work are within the scope of protection of the present invention.

[0044] It should be noted that the terms "including," "having," and any variations thereof in the embodiments of the present invention and the accompanying drawings are intended to cover non-exclusive inclusions. For example, a process, method, system, product, or apparatus comprising a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to the process, method, product, or apparatus.

[0045] like Figure 1As shown, the present invention provides a connected autonomous driving method based on vehicle-road-cloud integration, which is applied to a cloud control platform. The cloud control platform communicates with roadside sensors. The connected autonomous driving method of the present invention includes:

[0046] S1: Determine the area to be planned and the lane position information of each monitored road in the area to be planned in a pre-prepared high-precision map;

[0047] Among them, the lane position information includes the specific position of each lane in the high-precision map.

[0048] It is worth noting that the high-precision map of the present invention can be an existing map that can accurately display specific buildings, specific roads, and lane lines. Of course, the map can also present the three-dimensional form of buildings, roads, and traffic equipment.

[0049] In this high-precision map, the present invention can select multiple areas to be planned to perform the autonomous driving path planning and decision-making process of the present invention. In this high-precision map, the position information of each lane includes the specific position of each lane in the high-precision map, and the position is a unique coordinate in the high-precision map.

[0050] S2: Obtaining environmental status information returned by roadside sensors of each monitored road and current driving information returned by vehicle sensors of vehicles traveling on the monitored road;

[0051] Roadside sensors include video and radar sensors, and environmental status information is video data and / or radar data. Environmental status information includes target information, environmental information, and current traffic event information.

[0052] It's worth noting that roadside sensors are typically installed on both sides of the road or at intersections. They have a specific monitoring range. Each roadside sensor detects objects within its monitoring range, thereby obtaining environmental status information. This includes vehicles and pedestrians in each lane, as well as traffic light status at intersections. This allows the environmental status information transmitted from the roadside sensors to the cloud control platform to include global information.

[0053] S3: Determine target information from environmental state information;

[0054] Targets include pedestrians and vehicles; target information includes the specific location and status of the target in the high-precision map, and current driving information includes the driving destination and driving trend.

[0055] It is worth noting that the present invention requires global and local path planning for vehicles, thus requiring the identification of vehicles and pedestrians. In global planning, vehicles influence each other while driving, and the status of pedestrians may also cause changes in the vehicle's driving state and driving trend. In global information, targets are highly coupled, and the mutual interference between targets must be considered during the global planning process. Therefore, it is necessary to identify the status and location of vehicles and pedestrians.

[0056] S4: Target information is matched and fused sequentially to form a continuous recognition sequence;

[0057] It is worth noting that since the monitoring ranges of each roadside sensor can overlap, the data transmitted back to the cloud control platform may also partially overlap. Therefore, the present invention requires target matching and fusion of the identified target information. The fusion process is to sequentially organize multiple frames of data according to the target's motion state. The geographical location of the roadside sensor in the high-precision map is unique. When the target is a vehicle, its target information can be organized in sequence according to the geographical location of the roadside sensor. Of course, when the target is a pedestrian, a continuous recognition sequence can be formed according to the data generation time, that is, the generation time of the target information.

[0058] S5: Generate a global planning path for each vehicle based on current traffic event information, environmental information, and current driving information;

[0059] It's worth noting that vehicles typically have a destination in mind. In global planning, multiple paths exist from the current location to the destination, each of which is influenced by other targets and the surrounding environment. Furthermore, these paths may include intersections, each with its own traffic lights. Vehicles can generate traffic events based on these lights. Traffic events between vehicles can also affect global path planning, and these events must be considered during the planning process.

[0060] S6: Make decisions based on current traffic event information, environmental information, and the global planned path to generate the vehicle's current driving decision;

[0061] It's worth noting that the current driving decision is based on current traffic event and environmental information, but is unified with and follows the global planned path. It's a driving decision optimized based on current traffic event and environmental information in the current state.

[0062] S7: Generate local path planning for vehicle driving based on global path information and current driving decisions, and send the local path planning to the vehicle to achieve connected autonomous driving of the vehicle.

[0063] The present invention provides a connected autonomous driving method based on vehicle-road-cloud integration. This method deploys the algorithmic functions of connected autonomous driving on a cloud platform, leveraging the platform's computing power and resource advantages to compensate for the insufficient resources of a single vehicle's computing unit. It can also handle the concurrent scenarios of multiple connected vehicles in parallel. By aligning the connected autonomous driving functional architecture deployed on the cloud platform with that of a single vehicle, it can simultaneously provide cloud-controlled perception, decision-making, and planning results, providing comprehensive data support for autonomous driving vehicles. This method can achieve multi-vehicle collaboration within the same scenario, with the cloud platform simultaneously sending decision-making and planning recommendations to all relevant vehicles. This can improve the performance and safety margins of single-vehicle autonomous driving, reduce the vehicle's perception, decision-making, and computational costs, and resolve the common problem of conflicting individual decisions among multiple vehicles in autonomous driving.

[0064] In an optional embodiment, as Figure 2 As shown, the target information is matched and fused in sequence to form a continuous recognition sequence including:

[0065] S21: matching the target information according to the order of each roadside sensor in the monitored road;

[0066] It's worth noting that each roadside sensor has a directional order on the path and a unique position on the high-precision map. This allows for target matching based on their order on the monitored road, accelerating speed.

[0067] S22: forming a sequence of the matched target information;

[0068] S23: Deduplication processing is performed on target information that is completely overlapped or partially overlapped to form a sequence to form an identification sequence.

[0069] It is worth noting that: since some target information of each roadside sensor may overlap during the monitoring process, deduplication processing is required to reduce the amount of target information in the subsequent process and improve accuracy.

[0070] In a specific embodiment, reference Figure 3 , based on the current traffic event information, environmental information and current driving information, the global planning path of each vehicle is generated, including:

[0071] S31: For each vehicle, determine each target lane from the target location to the destination;

[0072] S32: Determine whether a traffic event occurs in each target lane based on the current traffic event information;

[0073] S33: If a traffic incident occurs in a certain target lane, determine the location of the traffic incident in the high-precision map;

[0074] S34: Predicting the next state of the vehicle and whether a traffic incident may occur in a target lane in the next state based on the environmental information and the current driving information;

[0075] For example, if the traffic light at the intersection ahead turns red 5 seconds before it turns green, vehicles about to pass through the intersection may brake and pause to wait for the red light. Therefore, traffic events that could affect vehicle movement can be anticipated. This paves the way for subsequent global route planning.

[0076] S35: When a traffic event is determined based on the current traffic event information and the predicted result of whether a traffic event occurs in a target lane in the next state, a global planned path of the vehicle is generated.

[0077] In a specific embodiment, the environmental information includes traffic control information and traffic display information on the road. Based on the environmental information and the current driving information, predicting the next state of the vehicle and whether a traffic incident may occur in a target lane in the next state includes:

[0078] Step 1: Predict the vehicle's next state based on traffic control information, traffic display information, and current driving information;

[0079] Step 2: Determine whether a traffic incident may occur in the next state of each vehicle based on the driving information of the next state of each vehicle.

[0080] It can be understood that if a traffic accident occurs in a certain lane, it can be predicted that it is impossible for a vehicle to pass through the lane in the next state, thereby predicting the traffic event information in the next state.

[0081] In an optional embodiment, making a decision based on current traffic event information, environmental information, and the global planned path to generate a current driving decision of the vehicle includes:

[0082] The current traffic event information, environmental information and global planning path are used as inputs of the preset decision model, so that the decision model can perceive the environmental information and the current traffic event information, and according to the global planning path, make the current driving decision of the next state indicating the driving direction and the driving route at the current position, and send the current driving decision to the control unit of the vehicle so that the vehicle can execute according to the current driving decision.

[0083] like Figure 4 As shown, the present invention provides a cloud control platform that executes any of the above-mentioned networked autonomous driving methods based on vehicle-road-cloud integration.

[0084] exist Figure 4In the process, there are multiple interactions between vehicles and the cloud control platform. Vehicle sensors perceive vehicle information, and roadside sensors fuse perceived environmental information to further identify target information. In the global path planning process, dynamic time and situation recognition is performed based on environmental information and vehicle driving information to determine event information. Combined with previous target information, motion decision recommendations are made and sent to the vehicle. The cloud control platform performs global path planning based on environmental information and vehicle driving information, and then predicts the event information of the next state to perform local path planning. The local path planning and decision recommendations of the present invention are unified in the global path planning, and real-time decisions and execution of local planned paths can be made according to the vehicle environment.

[0085] Those skilled in the art will appreciate that the accompanying drawings are merely schematic diagrams of an embodiment, and the modules or processes in the accompanying drawings are not necessarily required to implement the present invention.

[0086] Those skilled in the art will appreciate that the modules in the systems of the embodiments may be distributed throughout the systems of the embodiments as described in the embodiments, or may be modified accordingly and located in one or more systems different from the embodiments. The modules in the above embodiments may be combined into one module or further divided into multiple submodules.

[0087] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A connected autonomous driving method based on vehicle-road-cloud integration, applied to a cloud control platform that communicates with roadside sensors, characterized in that: include: Determine the area to be planned and lane position information of each monitored road in the area to be planned in a pre-prepared high-precision map; Obtain environmental status information returned by roadside sensors on each monitored road and current driving information returned by vehicle sensors of vehicles traveling on the monitored roads; Determining target information from environmental status information; wherein the environmental status information includes target information, environmental information, and current traffic event information, and the targets include pedestrians and vehicles; The target information is matched and fused sequentially to form a continuous recognition sequence; Generate a global planning path for each vehicle based on current traffic event information, environmental information, and current driving information; Make decisions based on current traffic event information, environmental information, and the global planned path to generate the vehicle's current driving decision; Generate local path planning for the vehicle based on global path information and current driving decisions, and send the local path planning to the vehicle to achieve connected autonomous driving. The networked autonomous driving functional architecture deployed by the cloud control platform is the same as that of a single vehicle; The current driving information includes a driving destination and a driving trend, and generating a global planning path for each vehicle based on the current traffic event information, environmental information, and current driving information includes: For each vehicle, determine the target lanes from the target location to the destination; Determine whether a traffic event occurs on each target lane based on current traffic event information; If a traffic incident occurs in a certain target lane, determine the location of the traffic incident in the high-precision map; Based on environmental information and current driving information, predict the next state of the vehicle and whether a traffic incident may occur in a target lane in the next state; When a traffic event is determined based on the current traffic event information and the predicted result of whether a traffic event occurs in a target lane in the next state, a global planning path for the vehicle is generated; The environmental information includes traffic control information and traffic display information on the road. The prediction of the next state of the vehicle and whether a traffic incident may occur in a target lane in the next state based on the environmental information and the current driving information includes: Predict the vehicle's next state based on traffic control information, traffic display information, and current driving information; Whether a traffic incident may occur in the next state of each vehicle is determined based on the driving information of the next state of each vehicle.

2. The connected autonomous driving method according to claim 1, characterized in that: The lane position information includes the specific position of each lane in the high-precision map, and the target information includes the specific position and status of the target in the high-precision map.

3. The connected autonomous driving method according to claim 1, wherein: The roadside sensors include video and radar sensors, and the environmental status information is video data information and / or radar data information.

4. The connected autonomous driving method according to claim 1, wherein: The target information is matched and integrated in sequence to form a continuous recognition sequence, which includes: According to the order of each roadside sensor in the monitored road, the target information is matched; Forming the matched target information into a sequence; The target information of the sequences formed by complete overlap and partial overlap is deduplicated to form an identification sequence.

5. The connected autonomous driving method according to claim 1, wherein: The decision-making based on the current traffic event information, environmental information and the global planned path to generate the current driving decision of the vehicle includes: The current traffic event information, environmental information and the global planning path are used as inputs of a preset decision model so that the decision model can perceive the environmental information and the current traffic event information, and make a current driving decision indicating the next state driving direction and driving route at the current position based on the global planning path.

6. The connected autonomous driving method according to claim 5, characterized in that: After generating the current driving decision of the vehicle, the connected autonomous driving method further includes: The current driving decision is sent to the vehicle's control unit so that the vehicle can execute according to the current driving decision.

7. A cloud control platform, characterized in that: Execute the connected autonomous driving method based on vehicle-road-cloud integration as described in any one of claims 1 to 6.

Images