L5 self-driving highway and L5 self-driving vehicle control method and device thereon

Abstract

This application belongs to the field of autonomous driving technology, and specifically relates to L5 autonomous driving control methods and devices based on L5 autonomous driving highways and on them. Addressing the shortcomings of existing technologies such as encoding / decoding and AI recognition failing to meet the accuracy requirements for autonomous driving safety, and the inefficiency of repetitive processing for each vehicle and each indicator, this application provides a scheme where the RSU at the entrance of each highway segment sends the corresponding lane attribute semantic table to all L5 autonomous vehicles entering the segment; all L5 autonomous vehicles read the corresponding lane number from the natural number memory set at all lane entrances of that segment; then, using the lane number as an index, they query the lane attribute semantic table to obtain lane attribute semantic information; and finally, based on this semantic information and other information, they form autonomous driving instructions. Its distinguishing feature is the technology of directly providing the entire lane's structured driving instructions to the vehicle at once through the road itself, changing the way vehicles acquire road information and representing a new form of vehicle-road cooperative method.

Description

Technical Field

[0001] This application relates to the field of autonomous driving technology, and in particular to a method and device for controlling L5 autonomous vehicles based on L5 autonomous driving highways and on them. Background Technology

[0002] The specific steps for lane-level precision positioning of vehicles in current autonomous driving systems are as follows:

[0003] 1. The Advanced Driver Assistance System (ADAS) is fused with a high-definition map (HD Map) to obtain a state vector data with latitude, longitude, altitude, attitude parameters, motion parameters, etc., with a precision of more than 8 decimal places.

[0004] 2. Convert latitude and longitude (WGS84) to the local coordinate system of the high-precision map (such as UTM (Universal Transverse Mercator)) or a Cartesian coordinate system to match the map data format. The high-precision map contains the following layers of data: Geometric layer: lane centerlines, boundaries, and widths (centimeter-level accuracy); Semantic layer: lane attributes (straight, left turn, right turn), speed limits, and intersection information (traffic lights, stop lines); Topological layer: lane connections and intersection paths; Feature layer: roadside landmarks (road signs, poles), and reflective points (for LiDAR matching), etc.

[0005] 3. Query process:

[0006] (1) Location matching: Use latitude, longitude and altitude to locate the road segment where the vehicle is located in the high-precision map.

[0007] (2) Lane determination: Combine attitude (heading angle) and lane lines detected by lidar / camera to determine the lane where the vehicle is located;

[0008] (3) Semantic extraction: Extract lane attributes (such as "left turn, speed limit 50km / h") and intersection distance (such as "50 meters from STOP line") from the map database;

[0009] (4) Path planning: Based on the topology layer, calculate the feasible path from the current lane to the target lane (e.g., crossing 1 lane).

[0010] However, current autonomous driving systems still face many challenges in lane-level positioning accuracy. For example, existing systems use too many expensive, complex, and computationally intensive devices for lane-level positioning; and the production and updating of high-precision maps are extremely costly. [1]Furthermore, high-precision positioning, high-precision map fusion, and roadside sign recognition all involve inefficient and wasteful repetitive processing of each vehicle and each sign. Since "repetitive processing of each vehicle and each sign" is a custom technical term included in the instruction manual summary, it is necessary to further explain and define it, as follows:

[0011] Take a "50 km / h speed limit" roadside sign as an example: This sign is mostly a sign standing by the roadside with a red circle at the top containing an image of the number "50". After the autonomous driving system captures this image, it performs real-time AI analysis to obtain the "50 km / h speed limit" autonomous driving instruction, thereby controlling the vehicle's speed. It is important to note that this process is repeated in real-time for all vehicles passing by a given sign; and for a given vehicle, it is repeated in real-time for all the same signs along its route to obtain the corresponding autonomous driving instruction. To make this repetitive processing more memorable, it can be named "for vehicles, all places; for signs, all vehicles," or more concisely, "vehicle-by-vehicle, sign-by-sign repetitive processing."

[0012] Currently, some existing technologies have proposed multi-task deep learning systems with an accuracy exceeding 98% in image recognition, object tracking, and classification tasks. [2][3] That is, an error rate of 10 2 However, with the increase in car ownership, the recognition error rate of autonomous driving systems at traffic lights at intersections must reach 10%. 8 The accuracy of existing autonomous driving systems' barcode decoding and AI recognition is insufficient to meet autonomous driving safety requirements. [4] .

[0013] At present, research on mixed-drive highways and fully autonomous driving highways can be roughly divided into two categories: solving the technical problems within mixed-drive highways under the conditions of ordinary autonomous driving (referred to as "ordinary autonomous driving" to L4 and below, to distinguish it from L5 autonomous driving) and mixed traffic of people and vehicles, and solving the phased transition from the current mixed-drive highways to L5 autonomous driving highways in terms of time.

[0014] Among these, regarding the technical challenges of mixed-traffic highways under conditions where ordinary self-driving vehicles and human-driven vehicles coexist: existing technologies can achieve safe and efficient mode-switching processes by setting up mode-switching zones and utilizing spatial physical isolation and precise positioning technologies, effectively solving a series of problems in the transition from autonomous driving to manual driving transportation; furthermore, existing technologies can also make full use of existing road resources, based on the premise of establishing separate autonomous driving lanes, allowing ordinary vehicles to use dedicated autonomous driving lanes when the utilization rate of autonomous driving lanes is low, realizing time-sharing of dedicated autonomous driving lanes on highways; in addition, some existing technologies have proposed a method for traffic control at intersections under conditions where ordinary self-driving vehicles and human-driven vehicles coexist under the above-mentioned problems.

[0015] Secondly, regarding the phased transition from current mixed-use highways to L5 autonomous driving highways, existing technologies can construct a highway design and traffic control system to support the coexistence of dedicated autonomous driving lanes and pedestrian / vehicle lanes on multi-lane highways. This transition can be achieved through mode switching subsystems, lane management subsystems, and merging / diversion subsystems. Existing technologies can also cover the systematic compatibility and migration steps of each stage—from mixed to partially autonomous driving to fully autonomous driving—and encompass the phased deployment and evolution framework of the entire regional road network, rather than being limited to a single highway or mining scenario. Some existing technical solutions can also be based on vehicle-to-everything (V2X) autonomous driving control methods, within an autonomous driving network migration framework under the V2X system. Starting with mixed traffic, L5 dedicated control can be deployed on certain road sections, gradually expanding to full network L5 full automation. Furthermore, existing technologies can also be based on a 5G-V2X vehicle-to-everything network migration framework, starting with mixed traffic, deploying some L5 dedicated road sections, and ultimately achieving full L5 network autonomy.

[0016] Furthermore, existing highways, where human-driven vehicles and Level 4 and below autonomous vehicles coexist, cannot achieve Level 5 fully autonomous driving. Why? To date, in human history, the only successful practical application of fully autonomous vehicles is in urban rail transit. For example, achieving GOA4 (Grade of Automation 4): UTO (Unattended Train Operation) level in the relevant technical guidelines, i.e., automated operation without human intervention. This is because in urban rail transit, vehicles are constrained to travel on rigid 1-dimensional tracks. More importantly, all vehicles on the tracks are autonomous, without human drivers. A crucial characteristic is that vehicles are uniformly controlled by a central control system within specific road sections. Imagine if some vehicles on the tracks were autonomous and others were human-driven; the accident rate would increase exponentially. In contrast to rail transit, vehicles on highways must be allowed to change lanes, meaning they travel on a 2-dimensional plane. The accident rate is an order of magnitude higher than in 1-dimensional rail transit. If autonomous and human-driven vehicles were allowed to coexist on a 2-dimensional highway, the accident rate would inevitably increase exponentially. If allowing some vehicles on the urban rail transit system to be self-driving and others to be human-driven is an unrealistic idea, then allowing some vehicles to be self-driving and others to be human-driven on a 2D plane road is even more unrealistic. The conclusion is: (1) On roads where self-driving and human-driven vehicles coexist, because human drivers are uncontrollable, Level 5 self-driving vehicles are a quagmire, and their practical application is far from being realized. (2) However, abandoning existing roads and rebuilding entirely new roads for all-self-driving vehicles (AV-only) requires a huge amount of land and funds, which is difficult to achieve. The realistic and feasible method is to transform existing mixed-use roads into pure Level 5 self-driving roads, and transform human-driven vehicles and ordinary self-driving vehicles into Level 5 self-driving vehicles.

[0017] In summary, the existing barcode decoding and artificial intelligence recognition of autonomous driving systems are insufficient to meet the accuracy requirements for autonomous driving safety, and the repeated processing of each vehicle and each barcode is inefficient. In addition, existing roads where human-driven vehicles and ordinary autonomous vehicles are used together cannot achieve L5 full autonomous driving, which urgently needs to be addressed. Summary of the Invention

[0018] This application provides a method and device for controlling L5 autonomous vehicles based on L5 autonomous driving highways and on them, in order to solve the problems that existing autonomous driving systems cannot meet the accuracy requirements of autonomous driving safety due to the difficulty of barcode decoding and artificial intelligence recognition, the low efficiency of repeated processing of each vehicle and each barcode, and the inability to achieve L5 full autonomous driving on existing highways where human-driven vehicles and ordinary autonomous vehicles are mixed.

[0019] This application provides a method for controlling L5 autonomous vehicles based on L5 autonomous driving highways, comprising the following steps: First active roadside units (RSU1) installed at the entrance of each segment of the target L5 autonomous driving highway communicate bidirectionally with the traffic control center of the corresponding segment and the L5 autonomous vehicles at the entrance of the corresponding segment, respectively, to send corresponding lane attribute semantic tables to all L5 autonomous vehicles entering each segment entrance; controlling all L5 autonomous vehicles to read the lane number (ID) corresponding to each lane from the natural number memory of all lane entrances of the corresponding segment, and querying the lane attribute semantic table using the lane ID as an index to obtain the corresponding lane attribute semantic information; second active roadside units (RSU2) installed at the entrance of each segment of the target L5 autonomous driving highway send corresponding lane attribute semantic tables to all L5 autonomous vehicles entering the corresponding segment. 5. The autonomous vehicle sends a dynamic autonomous driving instruction data summary table. Based on the dynamic autonomous driving instruction data summary table and the lane ID of the L5 autonomous vehicle, all autonomous driving dynamic information required for the road segment in the target L5 autonomous driving highway is obtained. The traffic control center of the road segment controls all L5 autonomous vehicles in the same lane to travel at the same speed and distance throughout the entire lane. When an L5 autonomous vehicle that needs to cross a lane reaches a preset distance before any intersection, a preset natural number memory is read. If the read data is the preset data, it means that the distance between the vehicle and the intersection is too short to complete the lane crossing action, so the vehicle stops crossing the lane.

[0020] Optionally, in one embodiment of this application, the first active roadside unit uses the C-V2X protocol or other protocols that meet the usage requirements to communicate bidirectionally with the traffic control center of each road segment and the L5 autonomous vehicles at the entrance of each road segment, so as to send the corresponding lane attribute semantic table to all L5 autonomous vehicles entering the entrance of each road segment. The lane attribute semantic table is a one-dimensional table (that is, unlike a two-dimensional table, it has two addresses pointing to one content in the vertical and horizontal directions). The "lane attribute semantic" table belonging to the road segment can be made according to the actual situation of the road segment, and maintained, modified and updated when the actual situation of the road segment changes.

[0021] Optionally, in one embodiment of this application, a passive natural number memory is installed at the entrance of all traffic lanes on any road segment. The memory stores the lane ID of that lane. The on-board unit (OBU) autonomous driving system of the approaching vehicle uses IT hardware readout technology to emit physical waves (light waves, electromagnetic waves, ultrasonic waves, etc.) and then receives the reflected waves from the memory to read the natural number. Depending on the different materials used to construct the memory, there may be a number of different types of physical characteristics of the reflected waves used for differentiation (assuming the number of such physical characteristics is K). The memory uses a K-ary system to store the lane ID natural number. For example, if the emitted wave is a millimeter wave, the memory uses materials that can reflect millimeter waves, such as metal strips (which reflect light), to represent "1", and materials that cannot reflect millimeter waves, such as wooden strips (which do not reflect light), to represent "0". In this case, K=2, and the memory uses a binary system. If the size, material, shape, surface finish, or concavity of the reflector causes differences in the frequency, waveform, peak value, width, or waveform integral of the reflected wave, resulting in three different reflected waves, then K=3, and the memory uses ternary representation. It is important to note that the binary and ternary representations here represent natural numbers, which are entirely different mathematical and informational concepts and elements from the encoding of barcodes, which uses different proportions of black and white bar widths (analog signals). Please note: Since the lane ID is used as a lookup index or address in the "lane attribute semantic table," if any (or several) of the listed lane attributes change, such as a change in the "speed limit," a new passive natural number memory must be set up at the location where the change occurs as the starting point. This memory stores the new lane ID as the starting point, even if the lanes before and after the new starting point are geographically the same lane. When a vehicle enters the next connected lane after the end of one lane, the vehicle's OBU must retrieve all lane attribute semantic content of the next lane from the one-dimensional table of lane attribute semantic content obtained from the active roadside unit RSU1 at the road segment entrance, with the new natural number lane ID as the lookup order (index or address), and then retrieve the new lane attribute semantic content of the next lane again with the new lane ID.

[0022] Optionally, in one embodiment of this application, the step of sending a dynamic self-driving instruction data summary table to all L5 self-driving vehicles in the corresponding road segment via a second active roadside unit (RSU2) located at the entrance of each segment of the target L5 self-driving highway, and obtaining all required self-driving dynamic information for the road segment in the target L5 self-driving highway based on the dynamic self-driving instruction data summary table and the lane ID of the L5 self-driving vehicle, includes: sending the dynamic self-driving instruction data summary table to all L5 self-driving vehicles in the corresponding road segment at a frequency that meets a preset high-frequency requirement via a second active roadside unit (RSU2) located at the entrance of each segment of the target L5 self-driving highway. The L5 autonomous vehicle obtains all the necessary dynamic information for driving on the target L5 autonomous driving highway segment based on the dynamic autonomous driving instruction data master table. The dynamic autonomous driving instruction data master table is a one-dimensional table. It includes a target index number (e.g., 0) pointing to the segment ID of the current road segment, index numbers above 0 pointing to all lane IDs of the current road segment, and index numbers above 0 pointing to all dynamic traffic signs and dynamic traffic control signals required for lane autonomous driving, including the dynamic traffic control signals for each lane's straight and left / right exits. After receiving this data master table from the vehicle's OBU receiver, the vehicle obtains all the necessary dynamic information for autonomous driving based on the obtained lane IDs. Specifically, the status of the traffic lights for straight-ahead and left / right exits in that lane (i.e., the traditionally understood "red, yellow, green" status) is obtained. This status information is not identified by AI methods, but rather the traffic light control system directly fills the dynamic autonomous driving instruction data table with code symbols corresponding to the "red, yellow, green" status. For example, 1, 2, and 3 represent red, yellow, and green status information, respectively. The corresponding dynamic autonomous driving instruction data table is received by the on-board unit (OBU) receiver of each L5 autonomous vehicle, and the dynamic autonomous driving instruction data table is queried based on the lane ID of each lane as an index to obtain all the autonomous driving dynamic information required for autonomous driving.

[0023] Optionally, in one embodiment of this application, the L5 autonomous vehicle, based on the global navigation satellite system, determines whether the lane in which the L5 autonomous vehicle is located meets the preset requirements for left or right turns or straight-ahead travel. If not, the L5 autonomous vehicle sends a lane change request to the road segment traffic control center. Upon receiving the lane change request, the road segment traffic control center sends yield control information to vehicles within the target range corresponding to the vehicle to be changed, controlling the vehicles within the target range to perform preset yield operations according to the yield control information, generating corresponding lane change positions, and controlling the vehicle to be changed to the lane change position. Furthermore, if the L5 autonomous vehicle... When a vehicle requests to cross a lane, it reads a preset natural number from a memory location at a predetermined distance before any intersection. If the read data matches the preset data, it indicates that the preset distance is insufficient to meet the required travel distance for the vehicle requesting to cross the lane. The vehicle must then stop its lane-crossing action. Based on the road segment traffic control center and the global navigation satellite system, and in conjunction with the lane attribute semantic information, the multiple autonomous driving dynamic information, and the preset mirror conveyor belt strategy, the road segment traffic control center controls all L5 autonomous vehicles in the same lane to travel at the same speed and distance to pass through the corresponding intersection or stop and wait before the intersection.

[0024] Optionally, in one embodiment of this application, the method further includes: setting a trip counter that meets a preset accuracy requirement in all L5 autonomous vehicles, and setting the trip counter to zero at each lane entrance and starting to count to obtain the corresponding trip count; based on the lane attribute semantic information, obtaining distance data from the lane entrance to the next lane entrance or intersection in each lane, and subtracting the trip count from the distance data to obtain distance data from the current position of the L5 autonomous vehicle to the next lane entrance or intersection in each lane; generating driving instructions corresponding to the L5 autonomous vehicle based on the distance data, and controlling the L5 autonomous vehicle to perform corresponding driving operations according to the driving instructions; setting multiple fixed distance markers on the target L5 autonomous driving highway, and automatically correcting the trip counter according to the multiple fixed distance markers to ensure that the trip counter meets the preset accuracy requirement.

[0025] Optionally, a Roadway Segment Traffic Control Center (RSTCC) is set up on the L5 autonomous driving highway. This center controls all L5 autonomous vehicles in each lane within the segment, transforming the 2D planar highway into a 1D-like soft lane. In this case, the parallel lanes can be likened to a row of parallel conveyor belts, each representing a lane; thus, the 2D planar highway is transformed into a 1D-like soft lane. The RTCC controls the L5 autonomous vehicles in each lane to maintain the same speed and distance, similar to a row of items placed on a conveyor belt, all moving at the same speed and maintaining a constant distance. The fixed lanes are likened to a moving conveyor belt, and the moving vehicles in the lanes to fixed items on the conveyor belt. This application refers to the analogy of "mirror" as the "mirror conveyor belt" method. Only when a vehicle needs to change lanes to complete its next left / straight / right turn as directed by GNSS (Global Navigation Satellite System) will it request permission from the traffic control center of that segment to allow it to move in. Once the vehicle has completed the lane change, it's like transferring items from one conveyor belt to another, and everything returns to its original operating pattern. Because the starting, stopping, speed, and distance between vehicles in the same lane are uniformly controlled by the central control system, vehicles in different lanes can pass through intersections in an orderly manner, thus replacing the function of traffic lights. Therefore, accidents such as violations of traffic light rules and driver erratic driving, common on such highways, can be eliminated. At this point, the accident rate on highways will be two orders of magnitude lower than the current rate, allowing for the widespread adoption of Level 5 autonomous driving.

[0026] Optionally, in one embodiment of this application, the target L5 autonomous driving highway is developed longitudinally from one segment of a mixed-use highway where human-driven vehicles and ordinary autonomous vehicles are used together. To illustrate this process in detail, this application defines two custom technical terms, which can also serve as a summary of the invention.

[0027] L5 self-driving highway definition:

[0028] 1. Eliminate all pedestrian crossings across highways and construct pedestrian overpasses or underpasses at all intersections; use railings to separate highways from pedestrian and non-motorized vehicle lanes, thus isolating pedestrians, non-motorized vehicles, and motorized vehicles in two separate worlds.

[0029] 2. All road segment entrances are equipped with active roadside units (RSU1), which use C-V2X or other communication protocols that meet the requirements to communicate bidirectionally with the road segment traffic control center and the vehicle OBUs at the road segment entrances. They transmit a one-dimensional lane attribute semantic table to the vehicles, with the natural number lane ID as the lookup order (index or address). Each natural number lane ID index points to the "lane attribute semantics" of that lane.

[0030] 3. At the entrance to all lanes on any road segment, a passive natural number memory is installed. The content stored in this memory is the lane ID for that lane. The OBU autonomous driving system of an approaching vehicle uses IT hardware readout technology to receive the reflected waves from the memory and read the natural number.

[0031] 4. At the entrance of any road segment, an active roadside unit (RSU2) is installed. It broadcasts a dynamic autonomous driving instruction summary table of all dynamic traffic signs and traffic control signals of that road segment to all vehicles entering the segment at a high frequency, such as not less than 1 Hz.

[0032] 5. The dynamic traffic control signals at intersections on the L5 self-driving highway retain the original rules from before the highway upgrade, i.e., the passage time is arranged according to the importance of the intersecting roads. However, the traffic light status (i.e., the traditional "red, yellow, green" status) is not identified by AI methods. Instead, the traffic light control system directly fills the dynamic self-driving instruction data table with codes corresponding to the "red, yellow, green" statuses. For example, 1, 2, and 3 represent red, yellow, and green status information, respectively.

[0033] 6. At a certain distance before all intersections, install a memory storing a pre-set natural number, for example, 100, which is a natural number exceeding the total number of lanes on most road sections. After the vehicle's OBU reads the number 100, it stops attempting to cross lanes and, according to the instructions of the road section's traffic control center, proceeds through the intersection or stops and waits before the intersection. Because the vehicle failed to move to the target lane in time, after passing the intersection, it re-plans its route to its destination according to GNSS instructions.

[0034] 7. All highways should be equipped with lane markings for use by the Lane Keeping Subsystem (LKS). To ensure the lane markings remain effective in rainy or snowy weather, it is recommended to use metal plates as lane marking indicators. It is also recommended that the Lane Keeping Subsystem use millimeter-wave emission, which can be reflected back to vehicles by the metal plate. The metal plate can be housed in a small plastic box and embedded in the lane marking location, as millimeter waves can penetrate plastic and soil.

[0035] 8. On the L5 autonomous driving highway, set up a section traffic control center to coordinate and control all L5 autonomous vehicles on this section through protocols such as CV2X and use the "mirror conveyor belt" method.

[0036] 9. For two-way single-lane highways, vehicles are also coordinated and controlled by the road section traffic control center. Pedestrian overpasses or underpasses are built where necessary; the highway is separated from the roadside pedestrian and non-motorized vehicle lanes by railings, and the same treatment applies to two-way multi-lane highways.

[0037] 10. All ramps (all one-way), viaducts and elevated highways, tunnels, etc., shall be treated in the same way as two-way multi-lane highways.

[0038] 11. The process of L5 self-driving highways evolving from mixed driving to L5 highways is along the longitudinal direction of the highway, developing from one section of the highway to the next. In contrast, the process of existing technology highways evolving from mixed driving to partially self-driving to fully self-driving is along the transverse direction of the highway, gradually developing from the center of the highway outwards.

[0039] 12. All motor vehicles preparing to enter the L5 self-driving highway must be equipped with an OBU system that meets the L5 self-driving definition specified in this application.

[0040] L5 autonomous vehicle definition:

[0041] 1. All vehicles are equipped with an OBU, which can communicate bidirectionally with the traffic control center of the road segment and the RSU1 at the road segment entrance through protocols such as C-V2X. The vehicle receives a lane attribute semantic table from the RSU1 with the natural number lane ID as the lookup order (index or address). Each natural number lane ID index points to the "lane attribute semantics" of that lane.

[0042] 2. All vehicle OBUs can use IT's hardware readout technology to read the natural number memory set at the lane entrance. This natural number is the ID of the lane in which the vehicle is located.

[0043] 3. All vehicle OBUs can receive the dynamic autonomous driving instruction data summary table of all dynamic traffic signs and all dynamic traffic control signals of the road segment at high frequency from the active roadside unit RSU2 at the entrance of the road segment where the vehicle is located.

[0044] 4. All vehicle OBUs are capable of operating the lane keeping subsystem to keep the vehicle near the center line of the lane.

[0045] 5. All vehicle OBUs must be equipped with a travel counter with centimeter-level accuracy.

[0046] 6. All vehicles must be set to human-driven mode so that when a vehicle breaks down in a tunnel, indoor parking lot, or other location where it cannot be moved by helicopter, it can be moved in human-driven mode.

[0047] 7. Before obtaining legal approval for use and public recognition for deployment, L5 self-driving vehicles must be driven by a human driver holding a driver's license.

[0048] 8. All L5 fully automated vehicles entering highways must be equipped with an OBU system that meets the L5 autonomous vehicle definition specified in this application.

[0049] The effectiveness of L5 autonomous driving roads and technology must undergo appropriate technical evaluation. A testing area could be established by a relevant organization, such as a closed network of 10 horizontal and 10 vertical (or more) L5 ​​autonomous driving roads, where 100 (or more) L5 ​​autonomous vehicles could be placed and run continuously 24 hours a day. Several animals or robots could be released into the road to interfere with the operation of the test vehicles. A sufficiently long period (e.g., one year or longer) should be used to identify and resolve all problems. Finally, the area should be publicly exhibited for a certain period to obtain approval for legal use and public acceptance.

[0050] Because the OBUs of L5 autonomous vehicles on the L5 autonomous driving highway (RSU1, RSU2, and the highway itself) have been certified by authoritative institutions, although licensed human drivers can switch the aforementioned autonomous vehicles to human driving mode, most licensed human drivers prefer to keep the vehicles in autonomous mode, monitor their operation, and be ready to take over driving at any time. Given that the aforementioned L5 fully automated vehicles have obtained legal administrative approval for use and public acceptance, most drivers would conclude that it is safe to ride in L5 autonomous vehicles in autonomous driving mode.

[0051] Therefore, the embodiments of this application have the following beneficial effects:

[0052] The innovation of this application lies in abandoning the mainstream existing technical path of "graphics / images, AI recognition, and semantic understanding" and the technical method of repeated processing for each vehicle and each label. Instead, it adopts a simplified system of "natural number numbering, hardware reading, and table lookup command invocation." This simplified system is a simplified solution for the complete system architecture in the context of autonomous driving, especially in the current mainstream trend of "multimodal perception and AI reasoning," representing a counter-trend engineering optimization. This application changes the way vehicles acquire road information by providing the entire lane's structured driving instructions directly to the vehicle through the road itself, representing a new form of vehicle-road cooperative strategy.

[0053] Secondly, existing technologies, from mixed driving to partially autonomous driving and then to fully autonomous driving, develop gradually from the center of the highway outwards along the lateral direction. This application, however, develops along the longitudinal direction of the highway, from one segment to the next. The existing control rules and equipment for intersections and ramps on mixed driving highways can continue to be used. Because existing technologies develop gradually from the center of the highway outwards along the lateral direction, the ramps for autonomous vehicles to enter the center from the outer edge of the highway need to be continuously reconstructed as the transition from partially autonomous driving to fully autonomous driving begins. Simultaneously, the time and space allocation for pedestrians and vehicles crossing intersections must also change with the transition.

[0054] Furthermore, the roadside signage and environmental recognition proposed in this application have the following advantages compared to existing technologies:

[0055] (1) Improved accuracy: The current AI error rate is 10%. - ² [2] The barcode error rate is 10%. 6[5] The RSU error rate for industrial Ethernet transmission in this application is 10. -10[6] The read error rate of memory IT hardware is generally 10%. - ¹² [6] That is, reduce the error rate from 10 -6 Reduced to 10 -10 and 10 -12 .

[0056] (2) Efficiency optimization: The table lookup complexity of this application is O(1), which is better than O(n) encoding and decoding and O(n²) processing of AI (O(1) and O(N) are big O notation, used to describe the time complexity or space complexity of an algorithm, representing the performance of the algorithm in the worst case; specifically, O(1) represents constant time complexity. No matter how large the input data size N is, the execution time (or space requirement) of the algorithm is fixed and does not change with the increase of the data size. For example, direct addressing directly accesses the array index A[k] through the key value k, which only requires one operation, so the time complexity is O(1); O(N) represents linear time complexity, and the execution time (or space requirement) of the algorithm is proportional to the input data size N; the larger the data size, the linearly the execution time increases. For example, sequential search requires checking N elements in the array one by one, so the time complexity is O(N)).

[0057] (3) Cost reduction: from the existing technology of hundreds of millions of yuan (high-precision map, AI system) to tens of thousands of yuan (RSU), thousands of yuan (passive natural number memory) and low-priced OBU hardware and software.

[0058] (4) Paradigm innovation: Shift from the technical means of perception combined with AI to the technical means of reading combined with table lookup.

[0059] Additional aspects and advantages of this application will be set forth in the following description. Attached Figure Description

[0060] Figure 1 This is a flowchart illustrating a method for controlling L5 autonomous vehicles based on L5 autonomous driving highways and on them, according to an embodiment of this application. Detailed Implementation

[0061] The embodiments of this application are described in detail below. The embodiments described with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0062] The following describes, with reference to the accompanying drawings, an L5 autonomous vehicle control method and apparatus based on L5 autonomous driving highways and on them, according to embodiments of this application. Addressing the problems of the prior art mentioned in the background section, this application provides an L5 autonomous vehicle control method based on L5 autonomous driving highways and on them. In this method, a first active roadside unit located at the entrance of each segment of the target L5 autonomous driving highway establishes bidirectional communication with the traffic control center of the corresponding segment and the L5 autonomous vehicles at the entrance of the corresponding segment, respectively, to send the corresponding lane attribute semantic table to all L5 autonomous vehicles entering each segment entrance; controls all L5 autonomous vehicles to read the lane ID corresponding to each lane from the natural number memory of all lane entrances of the corresponding segment, and queries the lane attribute semantic table using the lane ID as an index to obtain the corresponding lane attribute semantic information; and controls the control of each L5 autonomous vehicle located at the entrance of the target L5 autonomous driving highway to... At the entrance of each road segment, the second active roadside unit sends a dynamic autonomous driving instruction data summary table to all L5 autonomous vehicles in the corresponding road segment. Based on the dynamic autonomous driving instruction data summary table and the lane ID of the L5 autonomous vehicle, it obtains the required autonomous driving dynamic information for each intersection and all lanes in the target L5 autonomous driving highway. Based on the global navigation satellite system, lane attribute semantic information, autonomous driving dynamic information, and a preset mirror conveyor belt strategy, the road segment traffic control center controls all L5 autonomous vehicles in the same lane to travel at the same speed and distance throughout the entire lane. When an L5 autonomous vehicle reaches a preset distance before any intersection, it reads a natural number from the memory. In response to the read data being the preset data, it stops requesting vehicles to cross lanes. This application does not require high-precision maps and AI; it only needs to query the lane attribute semantic table of the road segment based on the road segment ID and lane ID to obtain the required autonomous driving instructions, thereby improving the efficiency and reliability of vehicle control. This solves the problems of existing autonomous driving systems, such as the difficulty in meeting the accuracy requirements of autonomous driving safety through barcode decoding and artificial intelligence recognition, the low efficiency of repeated processing of each vehicle and each barcode, and the inability to achieve full L5 autonomous driving on existing roads where human-driven vehicles and L4 and below autonomous vehicles are mixed.

[0063] Specifically, Figure 1 A flowchart illustrating an L5 autonomous driving control method based on an L5 autonomous driving highway and L5 autonomous vehicles provided in this application embodiment.

[0064] like Figure 1 As shown, the method for controlling L5 autonomous vehicles based on L5 autonomous driving highways includes the following steps:

[0065] In step S101, the first active roadside unit set at the entrance of each segment of the target L5 autonomous driving highway communicates bidirectionally with the traffic control center of the corresponding segment and the L5 autonomous driving vehicle at the entrance of the corresponding segment, so as to send the corresponding lane attribute semantic table to all L5 autonomous driving vehicles entering each segment entrance.

[0066] In actual implementation, the embodiments of this application only set up the first active roadside unit RSU1 at the entrance of all road segments, and use the C-V2X protocol or other protocols that meet the usage requirements to communicate bidirectionally with the road segment traffic control center and the vehicle OBU at the road segment entrance, so as to transmit a one-dimensional table with natural number lane ID as the lookup order (index or address) to the vehicle (i.e., L5 autonomous vehicles equipped with OBU and other components, i.e., autonomous vehicles that do not require human supervision at all). That is, each natural number lane ID index points to all lane attribute semantics of that lane.

[0067] Optionally, in one embodiment of this application, the target L5 autonomous driving highway is developed longitudinally from one segment of the mixed driving highway, where human-driven vehicles and L4 and below autonomous vehicles are used together.

[0068] As described in the invention above, the effectiveness of L5 autonomous driving highways and L5 autonomous driving vehicles must undergo relevant technical assessments and obtain corresponding approvals and public recognition.

[0069] This application embodiment takes a regional approach, and all roads, including ramps, undergo modification projects that meet the requirements of the above-mentioned L5 self-driving highway definition, while retaining existing roadside signs and traffic control systems and ensuring normal mixed-driving traffic.

[0070] It should be noted that during the transition phase, when all vehicles are equipped with OBUs that meet the requirements of the L5 autonomous vehicle definition, but the L5 autonomous driving highway renovation project is not yet completed, all vehicles must be supervised by a human driver with a valid driver's license, and all intersections will be directed by the existing mixed driving traffic control system. Secondly, during this transition phase, the road segment traffic control center can identify whether all vehicles in the road segment are in human-driven or L5 autonomous driving mode. The road segment traffic control center can control autonomous driving convoys (such as 5 or more vehicles) to pass through intersections using a mirror conveyor belt method.

[0071] Furthermore, after completing all the modifications required to meet the definition of an L5 autonomous driving highway, and when the traffic control center identifies that all vehicles in the road segment are not driven by humans and are in L5 autonomous driving mode, the transition phase of that road segment is completed, and the modification work for the next adjacent road segment can begin. In addition, when all road segments in a region announce the completion of their transition phase, the transition phase of that region is complete.

[0072] After the transition phase is completed, privately owned L5 self-driving cars must park in the parking lot of the nearest self-driving platform after arriving at their destination or returning home. When in use, a privately owned L5 self-driving car can be called, and after the L5 self-driving car receives the call, it will drive to the designated location to pick up the passenger and drive to the destination.

[0073] In summary, in the embodiments of this application, the process of a highway progressing from mixed driving to partial L5 to full L5 is along the longitudinal direction of the highway, developing from one segment of the highway to the next, whereas the process of a non-existent highway progressing from mixed driving to partial autonomous driving to full autonomous driving is along the transverse direction of the highway, gradually developing from the center of the highway outwards.

[0074] As one possible approach, embodiments of this application can install an active roadside unit (RSU1) at the entrance of any road segment. Since the communication distance of DSRC (Dedicated Short Range Communications) is 300–1000 meters (depending on line of sight and environment), the RSU can use the C-V2X protocol of the cellular network (Uu interface) to communicate bidirectionally with the vehicle and transmit a table to the vehicle with lane IDs starting from 1 as the lookup order (index or address). Each lane ID address points to the lane attribute semantics of that lane. This data includes lane attribute semantics from currently widely used high-precision map databases, with additions and subtractions.

[0075] It should be noted that the lane attribute semantic table content in this application embodiment includes the following:

[0076] (1) The segment ID with address 0;

[0077] (2) All lane attributes of all lanes in this road section, including:

[0078] 1) Straight lane;

[0079] 2) Left turn lane;

[0080] 3) Right turn lane;

[0081] 4) Left turn / straight lane;

[0082] 5) Right turn / straight lane;

[0083] 6) Lane for both left and right turns and straight driving;

[0084] 7) This lane is a left-turn lane. To go straight, you need to cross p lanes to the right. To turn right, you need to cross q lanes to the right (p and q are natural numbers).

[0085] 8) This lane is a right-turn lane. To go straight, you need to cross p lanes to the left. To turn left, you need to cross q lanes to the left.

[0086] 9) This lane is for straight-ahead travel. To turn left, you need to cross p lanes to the left; to turn right, you need to cross q lanes to the right, and so on.

[0087] (3) Speed ​​limit for this lane;

[0088] (4) The distance from the start of this lane to the left / right exit and the straight exit / intersection / STOP sign;

[0089] (5) Traffic light IDs for straight-ahead and left / right exits in this lane;

[0090] (6) All data necessary for driving, such as the ID of the next lane connected to this lane.

[0091] (7) etc.

[0092] The semantic representation of lane attributes, which uses natural number lane IDs as the lookup order (index or address), is shown in Table 1:

[0093] Table 1

[0094]

[0095] It is understood that the embodiments of this application do not require high-precision maps or database structures; they only require constructing a one-dimensional table with lane IDs as the order (index or address) and the lane attribute semantic content they point to. It should be noted that this lane attribute semantic table belonging to this road segment can be created by highway administration departments, relevant industry associations, or other authoritative entities based on the actual situation of that road segment, and can be maintained, modified, and updated when changes occur.

[0096] Therefore, the embodiments of this application only require setting an active roadside unit RSU1 at the road segment entrance to transmit a lane attribute semantic table to the vehicle. When the road conditions change subsequently, only data modification is needed in the lane attribute semantic table, and a small number of engineering modifications are required to the natural number memory of a few lane entrances, reducing the engineering construction cost during the modification. Thus, the embodiments of this application explicitly introduce a driving instruction table (i.e., a lane attribute semantic table) with natural numbers as addresses, so that the vehicle driving behavior becomes a direct execution result of the deterministic structured information of the roadside.

[0097] In step S102, all L5 autonomous vehicles are controlled to read the lane ID corresponding to each lane in the natural number memory of all lane entrances of the corresponding road segment, and query the lane attribute semantic table by using the lane ID as the index to obtain the corresponding lane attribute semantic information.

[0098] Furthermore, in embodiments of this application, a passive natural number memory is set at the entrance of all lanes on any road segment. The content stored in the memory is the lane ID of the lane. Vehicles entering the corresponding lane can read the natural number lane ID based on the IT hardware readout technology of the OBU autonomous driving system, and then use the lane ID to query the lane attribute semantic table to obtain all lane attribute semantic information of the corresponding lane.

[0099] Therefore, in this embodiment of the application, when a vehicle enters the corresponding lane entrance, a lookup table can be performed in the lane attribute semantic table to obtain all lane attribute semantic information of the corresponding lane at once, thereby providing the autonomous driving system with more time for driving planning (such as changing lanes).

[0100] Optionally, in one embodiment of this application, all L5 autonomous vehicles read the lane ID corresponding to each lane from the natural number memory at all lane entrances of the corresponding road segment. This includes: transmitting physical waves (light waves, electromagnetic waves, ultrasonic waves, etc.) from the L5 autonomous vehicles to the natural number memory located at all lane entrances of each road segment, and reflecting the reflected waves corresponding to the physical waves back to the L5 autonomous vehicles after the natural number memory receives the physical waves, so that the L5 autonomous vehicles can obtain the lane ID of the digital natural number type based on the reflected waves, and obtaining the new lane ID of the next lane when the L5 autonomous vehicles drive to the next lane, so as to re-query the lane attribute semantic table based on the new lane ID of the next lane to obtain the lane attribute semantic information corresponding to the next lane.

[0101] This application embodiment uses physical devices (such as RSU, natural number memory, etc.) and on-board table lookup operations to enable the L5 autonomous driving highway to output structured instructions, rather than controlling the vehicle to calculate them itself, thereby achieving the following beneficial effects:

[0102] (1) Determinism: The instruction response time is constant and is not affected by the complexity of the decoding algorithm or the fluctuation of signal quality.

[0103] (2) Reliability: It avoids the risk of misinterpretation and AI understanding errors caused by defects in decoding algorithms or signal interference.

[0104] (3) Simplicity: The vehicle unit does not need to have powerful signal processing or decoding operations or AI understanding capabilities, which greatly reduces the cost of software and hardware and power consumption.

[0105] Therefore, the embodiments of this application can construct a technical system that can directly provide deterministic, highly reliable and simple vehicle driving instructions without relying on on-board real-time decoding and AI understanding, but only through road physical devices.

[0106] Furthermore, in the embodiments of this application, a specific lookup strategy can be creatively applied to the field of directly providing vehicle driving instructions through only road physical devices, replacing the deeply ingrained decoding and AI understanding paradigm. A matching hardware and software co-architecture of road surface RSU, natural number memory, and onboard "lane attribute semantic table" is designed to achieve this. This abandons the mainstream existing technical path of "graphics / images, AI recognition, semantic understanding" and the technical method of repeated processing per vehicle and per label, instead adopting a simplified system of "natural number numbering, hardware reading, and lookup instruction invocation." Especially in the current mainstream trend of "multimodal perception and AI reasoning," the embodiments of this application represent a counter-trend engineering optimization, using a technical means to directly provide vehicles with structured driving instructions for the entire lane at once through the road ontology, thereby changing the way vehicles acquire road information and representing a new form of vehicle-road cooperative method.

[0107] Optionally, in one embodiment of this application, the method further includes: setting a trip counter that meets a preset centimeter-level accuracy requirement in all L5 autonomous vehicles to obtain distance data from the L5 autonomous vehicle to the next lane entrance or intersection in each lane; and setting multiple fixed distance markers (e.g., 500 meters) on the target L5 autonomous driving road to automatically correct the trip counter to ensure that the trip counter meets the preset accuracy requirement.

[0108] It should be noted that existing technologies include devices such as active or passive reflective barcode patterns on the road surface, and the corresponding lane information can only be decoded and obtained when the vehicle reaches the location of the device. However, the embodiment of this application only sets an active roadside unit (RSU1) at the entrance of the road segment to transmit a lane attribute semantic table to the vehicle. The index of this table is the ID of the lane in the road segment, and the index points to all the lane attributes of the lane (such as the distance of the vehicle from the left or right exit or straight exit, intersection, STOP sign, and the traffic light IDs for the straight and left / right exits of this lane). The embodiment of this application can obtain all the lane attribute semantic table information at once when the vehicle enters the lane entrance, thus gaining execution time for some driving instructions, such as vehicles that need to change lanes according to GNSS navigation requirements.

[0109] As one possible approach, all highways in this embodiment can also be equipped with a Lane Keeping Subsystem (LKS) to identify lane lines. To ensure the lane line identification markers function properly in rainy or snowy weather, this embodiment can use a metal sheet as the lane line identification marker. The Lane Keeping Subsystem can emit millimeter waves, which can be reflected back to the vehicle by the metal sheet. Since millimeter waves can penetrate plastic and soil, the metal sheet can be housed in a small plastic box and buried at the lane line identification location. Thus, this embodiment can operate the Lane Keeping Subsystem to control the vehicle to run near the lane center line.

[0110] In step S103, the second active roadside unit set at the entrance of each segment of the target L5 autonomous driving highway sends a dynamic autonomous driving instruction data summary table to all L5 autonomous driving vehicles in the corresponding segment. Based on the dynamic autonomous driving instruction data summary table and the lane ID of the L5 autonomous driving vehicle, the autonomous driving dynamic information required for each segment intersection and all lanes in the target L5 autonomous driving highway is obtained.

[0111] Subsequently, in embodiments of this application, an active roadside unit (RSU2) can be installed at the entrance of any road segment to publish a dynamic autonomous driving instruction data summary table of all dynamic traffic signs and all traffic control signals of the road segment to all vehicles entering the road segment at a high frequency (e.g., not less than 1Hz). This allows the vehicle to obtain all autonomous driving dynamic information required for autonomous driving on the L5 autonomous driving highway segment from the dynamic autonomous driving instruction data summary table.

[0112] It should be noted that in the embodiments of this application, the dynamic traffic signal at the intersection of the L5 autonomous driving highway retains the rules before the highway renovation, that is, the passage time is arranged according to the importance of the intersecting roads; the embodiments of this application do not obtain the traffic light status of the corresponding lanes for straight and left and right exits through AI recognition technology, but the traffic light control system directly fills the dynamic autonomous driving instruction data table with code symbols corresponding to the "red, yellow, and green" status, for example, 1, 2, and 3 represent red, yellow, and green status information respectively.

[0113] In step S104, based on the global navigation satellite system, lane attribute semantic information, autonomous driving dynamic information, and the preset mirror conveyor belt strategy, the road segment traffic control center controls all L5 autonomous vehicles in the same lane to travel at the same speed and distance throughout the entire lane. When an L5 autonomous vehicle travels to a preset distance before any intersection, it reads a preset natural number memory. In response to the read data being the preset data, the requirement for vehicles to cross lanes to cross lanes is stopped.

[0114] Subsequently, embodiments of this application can install a natural number memory storing a specially preset value at a certain distance before all intersections, for example, a natural number exceeding 100, which is greater than the number of lanes in most road segments. After the vehicle's OBU reads 100, it stops requesting vehicles to cross lanes. Then, the road segment traffic control center directs the corresponding vehicles to pass through the intersection or stop and wait before the intersection. Since the vehicle did not move to the target lane in time, after passing the intersection, GNSS can be used to replan the route to the destination.

[0115] It should be noted that the embodiments of this application can set up a road segment traffic control center on the L5 autonomous driving highway to coordinate and control all L5 autonomous vehicles on the road segment through the CV2X (Cellular Vehicle to Everything) protocol and mirror conveyor belt strategy. It is understood that most of the autonomous driving commands involved in the embodiments of this application are obtained by the OBU reading the lane ID from the natural number memory at each lane entrance within the road segment and then looking it up in the lane attribute semantic table. Therefore, the embodiments of this application only require a road segment traffic control center to control one road segment.

[0116] Because the starting, stopping, speed, and distance between all vehicles in the same lane are uniformly controlled by the central control system of the road segment, vehicles in different lanes can take turns passing through the intersection in an orderly manner, thus replacing the function of traffic lights, thereby reducing the probability of traffic accidents and improving the user's driving experience.

[0117] As an feasible approach, for two-way single-lane highways, vehicles can be coordinated and controlled by the roadside traffic control center. Pedestrian overpasses or underpasses can be constructed at necessary locations, and barriers can be used to separate the highway from pedestrian and non-motorized vehicle lanes. Simultaneously, all L5 autonomous vehicles must be set to human-driven mode so that if a vehicle breaks down in a tunnel, indoor parking lot, or other location where helicopter removal is not possible, it can be moved in human-driven mode. The same treatment applies to two-way multi-lane highways. Furthermore, all ramps (all one-way), viaducts and elevated highways, tunnels, etc., are treated in the same way as two-way multi-lane highways.

[0118] It is understood that the embodiments of this application have the technical effects of improved accuracy, optimized efficiency, reduced cost, and paradigm innovation compared with the prior art.

[0119] Secondly, this application embodiment also provides an L5 autonomous driving vehicle control device based on L5 autonomous driving highway and above. It should be noted that the foregoing explanation of the L5 autonomous driving vehicle control method embodiment based on L5 autonomous driving highway and above also applies to the L5 autonomous driving vehicle control device based on L5 autonomous driving highway and above in this embodiment, and will not be repeated here.

[0120] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0121] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

Claims

1. A L5 self-driving highway and a L5 self-driving vehicle control method thereon, characterized in that, Includes the following steps: By setting up a first active roadside unit at the entrance of each segment of the target L5 autonomous driving highway, two-way communication is established between the traffic control center of the corresponding segment and the L5 autonomous driving vehicle at the entrance of the corresponding segment, so as to send the corresponding lane attribute semantic table to all L5 autonomous driving vehicles entering each segment entrance. The system controls all L5 autonomous vehicles to read the lane ID corresponding to each lane from the natural number memory of all lane entrances of the corresponding road segment, and to query the lane attribute semantic table using the lane ID as an index to obtain the corresponding lane attribute semantic information. By sending a dynamic self-driving instruction data summary table to all L5 self-driving vehicles in the corresponding road segment through a second active roadside unit set at the entrance of each road segment of the target L5 self-driving highway, and obtaining the required self-driving dynamic information for each road segment intersection and all lanes in the target L5 self-driving highway based on the dynamic self-driving instruction data summary table and the lane ID of the L5 self-driving vehicle. Based on the global navigation satellite system, the lane attribute semantic information, the autonomous driving dynamic information, and the preset mirror conveyor belt strategy, the road segment traffic control center controls all L5 autonomous vehicles in the same lane to travel at the same speed and distance throughout the entire lane. When an L5 autonomous vehicle travels to a preset distance before any intersection, it reads a preset natural number from the memory. In response to the read data being the preset data, the system stops the cross-lane action of vehicles that are required to cross lanes.

2. The L5 self-driving highway and the L5 self-driving vehicle control method thereon according to claim 1, characterized in that, The first active roadside unit uses a communication protocol that meets preset usage requirements to communicate bidirectionally with the traffic control center of each road segment and the L5 autonomous vehicles at the entrance of each road segment, so as to send the corresponding lane attribute semantic table to all L5 autonomous vehicles entering the entrance of each road segment. The lane attribute semantic table is a one-dimensional table, the index of the lane attribute semantic table is the lane ID, and the table content pointed to by the index is the lane attribute semantic of the corresponding lane. 3.The L5 self-driving highway and L5 self-driving vehicle control method thereon based on claim 1, characterized in that, The L5 autonomous vehicles read the lane ID corresponding to each lane from the natural number memory of all lane entrances of the corresponding road segment, including: The L5 autonomous vehicle emits physical waves to natural number memory locations at the entrances of all lanes on each road segment. After receiving the physical waves, the natural number memory reflects the corresponding reflected waves back to the L5 autonomous vehicle. This allows the L5 autonomous vehicle to obtain a lane ID of type digital natural number based on the reflected waves. When the L5 autonomous vehicle moves to the next lane, it obtains a new lane ID for the next lane and re-queries the lane attribute semantic table based on the new lane ID to obtain the lane attribute semantic information for the corresponding next lane. 4.The L5 self-driving highway and L5 self-driving vehicle control method thereon based on claim 1, characterized in that, The process involves sending a dynamic self-driving instruction data summary table to all L5 autonomous vehicles in the corresponding road segment via a second active roadside unit located at the entrance of each segment of the target L5 autonomous driving highway. Based on the dynamic self-driving instruction data summary table and the lane ID of the L5 autonomous vehicle, the required autonomous driving dynamic information for each intersection and all lanes of the target L5 autonomous driving highway is obtained, including: By setting a second active roadside unit at the entrance of each segment of the target L5 self-driving highway to send a dynamic self-driving instruction data summary table to all L5 self-driving vehicles in the corresponding segment at a frequency that meets the preset high-frequency requirements, the L5 self-driving vehicles obtain all the self-driving dynamic information required for the segment of the target L5 self-driving highway based on the dynamic self-driving instruction data summary table. The dynamic self-driving instruction data summary table is a one-dimensional table. The table content pointed to by the target index number is the segment ID of the current segment. The other index numbers besides the target index number are all lane IDs of the current segment. The table content pointed to by all lane IDs includes all dynamic traffic signs and dynamic traffic control signals required for self-driving, including the dynamic traffic control signals of the straight exit and left and right exits of each lane.

5. The method for controlling L5 autonomous vehicles based on L5 autonomous driving highways and on them, as described in claim 1, is characterized in that... The method involves controlling all L5 autonomous vehicles in the same lane through the traffic control center of the road segment to travel at the same speed and distance throughout the entire lane. When an L5 autonomous vehicle reaches a preset distance before any intersection, a preset natural number is read from the memory. In response to the read data being the preset data, the method stops the lane-crossing action of vehicles requesting to cross lanes. This includes: Using the L5 autonomous vehicle based on the global navigation satellite system, it is determined whether the lane where the L5 autonomous vehicle is located meets the preset requirements for left or right turn or straight-through. If not, the L5 autonomous vehicle sends a lane change request to the traffic control center of the road segment. After receiving the lane change request, the traffic control center sends yield control information to vehicles within the target range corresponding to the vehicle to be changed, so as to control the vehicles within the target range to perform preset yield operations according to the yield control information, generate the corresponding lane change position, and control the vehicle to be changed to change lanes to the lane change position. At a predetermined distance before any intersection, the L5 autonomous vehicle reads a predetermined natural number from a memory, obtains the read data, and determines whether the read data is a predetermined natural number. If the read data is the predetermined natural number, the vehicle is stopped from crossing lanes. Based on the global navigation satellite system, and combined with lane attribute semantic information, multiple autonomous driving dynamic information, and a predetermined mirror conveyor belt strategy, the traffic control center controls all L5 autonomous vehicles in the same lane to travel at the same speed and distance to pass through the corresponding intersection or stop and wait before the intersection.

6. The method for controlling L5 autonomous vehicles based on L5 autonomous driving highways and on them, as described in claim 1, is characterized in that... Also includes: In all L5 autonomous vehicles, a travel counter that meets the preset accuracy requirements is set, and the travel counter is reset to zero at the entrance of each lane and then starts counting to obtain the corresponding travel count. Based on the lane attribute semantic information, the distance data from the lane entrance to the next lane entrance or intersection in each lane is obtained, and the distance data from the trip count to the current position of the L5 autonomous vehicle to the next lane entrance or intersection in each lane is obtained by subtracting the distance data from the distance data. Based on the distance data, the driving command corresponding to the L5 autonomous vehicle is generated, so as to control the L5 autonomous vehicle to perform the corresponding driving operation according to the driving command. Multiple fixed distance markers are set on the target L5 self-driving highway, and the trip counter is automatically calibrated according to the multiple fixed distance markers to ensure that the trip counter meets the preset accuracy requirements. 7.The L5 self-driving highway and L5 self-driving vehicle control method thereon of claim 1, wherein, The target L5 autonomous driving highway is a mixed driving highway used by both human-driven vehicles and L4 and below autonomous vehicles, developed longitudinally from one segment of the mixed driving highway to the next.

Images