Navigation path planning method, computer equipment, computer readable storage medium, computer program product and mobile platform

Abstract

The invention discloses a navigation path planning method, computer equipment, a computer readable storage medium, a computer program product and a mobile platform. The navigation path planning method comprises the following steps: acquiring a current position and a destination position of a mobile platform; determining a corresponding memory map according to the current position and the destination position of the mobile platform; searching a global topological path at least according to the lane-level topological map to determine a global topological path from the current position to the destination position; and searching a local geometric path at least according to the global topological path and the road traffic information collected by the mobile platform in real time so as to determine the local geometric path. According to the method, the road smoothness from the current position of the vehicle to the destination position is ensured through global topological path search, and finer local navigation planning is realized through local geometric path search based on vehicle online sensing information.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automatic driving, and in particular to a navigation path planning method, a computer device, a computer readable storage medium, a computer program product and a mobile platform. BACKGROUND

[0002] There are successful precedents for the application of high-precision map-based assisted driving schemes, and certain effects have been achieved. However, the mapping and maintenance of high-precision maps are costly, which restricts the popularization of intelligent assisted driving schemes based on high-precision maps. With the improvement of vehicle sensor performance, the increase of automatic driving chip computing power and the progress of related algorithms, vehicles themselves have the ability to map common routes, and the memory map established has rich details that are not inferior to high-precision maps.

[0003] Therefore, how to develop a navigation scheme based on the memory map established by the vehicle itself to replace the high-precision map-based navigation scheme has become a problem to be solved. SUMMARY

[0004] The embodiments of the present application provide a navigation path planning method, a computer device, a computer readable storage medium, a computer program product and a mobile platform, which are used to at least solve one of the above technical problems.

[0005] In a first aspect, the embodiments of the present application provide a navigation path planning method, comprising:

[0006] obtaining a current position and a destination position of a mobile platform;

[0007] determining a corresponding memory map according to the current position and the destination position of the mobile platform, the memory map comprising a lane-level topological map from the current position to the destination position;

[0008] performing global topological path search according to at least the lane-level topological map to determine a global topological path from the current position to the destination position;

[0009] performing local geometric path search according to at least the global topological path and road traffic information collected by the mobile platform in real time to determine a local geometric path.

[0010] In some embodiments, the memory map further comprises a vehicle owner historical driving trajectory.

[0011] The at least lane-level navigation path planning from the current position to the destination position according to the lane-level topological map comprises:

[0012] performing global topology path search according to the lane-level topology map and the driver history driving trajectory to determine a global topology path from the current position to the destination position.

[0013] In some embodiments, the memory map further comprises the driver history driving trajectory.

[0014] performing local geometry path search according to at least the global topology path and the road traffic information collected by the mobile platform in real time to determine a local geometry path, comprising:

[0015] performing local geometry path search according to the global topology path, the driver history driving trajectory and the road traffic information collected by the mobile platform in real time to determine a local geometry path.

[0016] In some embodiments, the road traffic information comprises at least one of an obstacle, a road boundary and a lane line, and the local geometry path comprises local lane-changing information and curvature information of the local path.

[0017] In some embodiments, the memory map further comprises a cost map, wherein the cost map comprises cost values corresponding to each lane segment.

[0018] In some embodiments, the navigation path planning method further comprises: pre-processing weight values of lane segments before and after a junction point in the cost map to reduce the probability of lane changing of the vehicle before and after the junction point; and / or, pre-processing weight values of lane segments corresponding to the driver history driving trajectory in the cost map to generate a navigation path in line with the driver habit; and / or,

[0019] pre-processing weight values of lane segments where a diverging point and a merging point are located in the cost map to control the situation of the vehicle passing through the diverging point or the merging point; and / or,

[0020] pre-processing weight values of corresponding most lateral lane segments in the cost map according to road information and navigation information to reduce the probability of the vehicle driving in the most lateral lane; and / or,

[0021] pre-processing at least weight values of impassable lane segments in the cost map to limit the vehicle passing through.

[0022] In some embodiments, the memory map further comprises curvature values of the lane segments, and the method further comprises:

[0023] determining whether the corresponding lane is a curved lane or a U-turn lane according to the curvature values;

[0024] for the curved lane or the U-turn lane, calculating a longitudinal distance from the curved lane or the U-turn lane and / or calculating a maximum speed limit of the vehicle according to the corresponding curvature value.

[0025] Secondly, embodiments of this application provide a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the navigation path planning method described in any embodiment of this application.

[0026] Thirdly, embodiments of this application provide a computer-readable storage medium having a computer program / instructions stored thereon, which, when executed by a processor, implements the steps of the navigation path planning method described in any embodiment of this application.

[0027] Fourthly, embodiments of this application provide a computer program product, including a computer program / instructions, which, when executed by a processor, implement the steps of the navigation path planning method described in any embodiment of this application.

[0028] Fifthly, embodiments of this application provide a mobile platform on which the computer device described in any embodiment of this application is installed.

[0029] This application implements a two-layer, multi-level navigation path search method that combines global topological path search and local geometric path search. Global topological path search ensures smooth road access from the vehicle's current location to its destination, while local geometric path search based on the vehicle's online perception information enables more refined local navigation planning. Attached Figure Description

[0030] To more clearly illustrate the technical solutions of the embodiments of this application, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0031] Figure 1 A flowchart of an embodiment of the navigation path planning method of this application;

[0032] Figure 2 A flowchart of another embodiment of the navigation path planning method of this application;

[0033] Figure 3 A flowchart of another embodiment of the navigation path planning method of this application;

[0034] Figure 4 This is a schematic diagram of an embodiment of preprocessing memory maps in this application;

[0035] Figure 5 This is a schematic diagram of the path planning in the topology layer in this application;

[0036] Figure 6 This is a schematic diagram of the path planning at the geometric layer in this application;

[0037] Figure 7 This is a schematic diagram illustrating the determination of the longitudinal travel distance and the number of lateral lane changes by post-processing in this application;

[0038] Figure 8 This is a schematic diagram of the structure of an embodiment of the computer device of this application. Detailed Implementation

[0039] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other.

[0040] It should also be noted that, in this document, the terms "comprising" or "including" include not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0041] In the process of developing this application, the inventors discovered that intelligent assisted driving solutions based on high-precision maps heavily rely on accurate and frequently updated high-precision maps. However, at present, the high-precision maps provided by various high-precision map service providers are difficult to meet user needs in terms of both update frequency and cost, as detailed below:

[0042] 1. High-precision maps have low coverage, only covering the main roads of key cities;

[0043] 2. High-precision maps have a low update frequency, and high-precision maps that cannot be updated frequently cannot meet the needs of intelligent assisted driving under complex road conditions.

[0044] 3. High-precision maps have high maintenance costs, and these high costs ultimately have to be borne by users, which limits the widespread adoption of high-precision maps;

[0045] 4. High-precision maps rely on satellite positioning, but there are problems such as poor satellite positioning signals in densely built-up urban areas.

[0046] The above problems can lead to a degradation of vehicle navigation capabilities, and in severe cases, even safety issues.

[0047] Furthermore, during their attempts to develop a rule-based memory map navigation scheme, the inventors discovered that while it could meet some of the needs of intelligent driving, the complex and varied road conditions and structures meant that the rule-based memory map navigation algorithm struggled to cover all road situations, leading to the following problems:

[0048] 1. Rule-based navigation schemes are difficult to reflect user preferences;

[0049] 2. Rule-based navigation schemes may miss route points;

[0050] 3. Rule-based navigation schemes are prone to incorrect lane changes;

[0051] 4. Rule-based navigation schemes have poor scalability and are difficult to cope with complex and ever-changing road conditions and structures.

[0052] Therefore, this application provides a navigation path planning method to solve at least one technical problem in the related art. Exemplarily, this navigation path planning method can be implemented as a driver assistance navigation module with long-distance planning capabilities (this driver assistance navigation module can be a computer program product, a combination of a computer program product and a computer processor, or a computer processor, etc.), providing navigation information and beyond-line-of-sight speed limit information to downstream driver assistance modules, and applied to driver assistance systems that do not rely on high-precision maps. This solution relies solely on offline memory maps and online perception information to achieve functions such as intelligent lane changing and beyond-line-of-sight speed limits during intelligent driving. For example, after the user enters the memory navigation function and successfully locates the vehicle, the driver assistance navigation module with long-distance planning capabilities described in this application will automatically enable and output navigation information and beyond-line-of-sight speed limit information.

[0053] like Figure 1 As shown, embodiments of this application provide a navigation route planning method. This navigation route planning method can be applied to mobile platforms, including but not limited to vehicles; the following embodiments use a vehicle as an example. The navigation route planning method includes:

[0054] S10. Obtain the current location and destination location of the mobile platform.

[0055] For example, the destination location is obtained based on the user's selection of a destination, and the current location of the mobile platform is obtained based on the vehicle's positioning information. For instance, if a user selects the memory navigation function on the vehicle's infotainment platform and enters a destination, the infotainment platform automatically determines the destination location based on the user's input and determines the vehicle's current location based on the positioning information obtained from the vehicle's positioning sensors.

[0056] S20. Determine a corresponding memory map based on the current location and destination location of the mobile platform. The memory map includes a lane-level topology map from the current location to the destination location.

[0057] For example, the memory map is constructed based on the user's historical driving data. For instance, after the user starts the vehicle and enters the destination, the user is prompted whether to select to build a memory map. If the user selects yes, the in-vehicle platform will collect road information, driving trajectory information, etc. in real time during the user's journey to the destination in order to construct a memory map from the vehicle's current location to the destination.

[0058] In some embodiments, an optimized memory map can be formed by fusing memory maps constructed multiple times for the same current location and destination. The current location can be, for example, the user's residence, and the destination can be the user's company location. The memory map includes a lane-level topology map from the current location to the destination location for lane-level navigation. The lane-level topology map includes multiple lane segments and the topological connections between these lane segments.

[0059] In some embodiments, multiple historical driving trajectories are recorded during the construction of memory maps between the same current location and destination multiple times. These multiple historical driving trajectories can be recommended to the user for selection during navigation route planning. Furthermore, before recommending historical driving trajectories to the user, based on the current navigation conditions (e.g., congestion levels of each historical driving trajectory), historical driving trajectories with better accessibility (e.g., relatively smooth traffic) are recommended to the user.

[0060] S30. Perform a global topology path search based at least on the lane-level topology map to determine a global topology path from the current location to the destination location.

[0061] For example, step S30 can be performed by a topology layer searcher. The topology layer searcher performs a global topology path search (as shown in the figure) on the lane-level topology map. A connected global topology path is obtained between the lane segment where the starting point (corresponding to the current position before the vehicle departs) and the lane segment where the ending point (corresponding to the destination of the vehicle) is located. This global topology path is a serialized combination of a series of lane segments.

[0062] In some embodiments, a global topological path search is performed from a lane-level topological map based on a user-selected historical driving trajectory. For example, a global topological path is formed by searching multiple lane segments covered by the historical driving trajectory in the lane-level topological map.

[0063] S40. At least based on the global topology path and the road traffic information collected in real time by the mobile platform, a local geometric path search is performed to determine the local geometric path.

[0064] Since the global topology path output by the global topology path search is a serialized combination of lane segments, including the starting point, ending point, whether a lane change or continuous lane change is needed in the vehicle's memory map, the mileage that can be continued if no lane change is needed, and the lane segments suggested to be traversed throughout the entire navigation process, the granularity is relatively coarse. Therefore, a more refined local geometric path search is required.

[0065] For example, step S40 can be performed by a geometry layer searcher. This geometry layer searcher combines the results of the vehicle's online perception with the global topology path to perform a geometry path search to determine a local geometry path. A local geometry path can be viewed as a local trajectory that the navigation system hopes the vehicle will travel.

[0066] For example, geometric path search uses a search tree to explore a local space. During the growth of the search tree, the geometric layer searcher combines perceived road traffic information such as obstacles, road boundaries, and lane lines to calculate a cost for each node in the search tree. The search tree generated during the search process can be viewed as an abstract representation of the expanded region. This is more granular and has higher resolution compared to lane segment combinations obtained through topological search.

[0067] This application implements a two-layer, multi-level navigation path search method that combines global topological path search and local geometric path search. Global topological path search ensures smooth road access from the vehicle's current location to its destination, while local geometric path search based on the vehicle's online perception information enables more refined local navigation planning.

[0068] In some embodiments, the memory map further includes: the vehicle owner's historical driving trajectory; planning a lane-level navigation path from the current location to the destination location based at least on the lane-level topology map, including: performing a global topology path search based on the lane-level topology map and the vehicle owner's historical driving trajectory to determine a global topology path from the current location to the destination location.

[0069] For example, the vehicle owner's historical driving trajectory is the driving trajectory formed when the user drives the vehicle from the current location to the destination location during the construction of the memory map. This historical driving trajectory can be the trajectory obtained while driving the current vehicle, or it can be the historical driving trajectory obtained by the user driving other vehicles and then importing it into the current vehicle. This driving trajectory reflects the user's driving habits. By analyzing the lane segments covered by the vehicle owner's historical driving trajectory, the global topology path closest to the vehicle owner's historical driving trajectory can be determined based on the lane-level topology map.

[0070] For example, by analyzing the lane segments covered by the vehicle owner's historical driving trajectory, when planning a lane-level navigation path from the current location to the destination location based on the lane-level topology map, the lane segments covered by the vehicle owner's historical driving trajectory are prioritized for planning into the lane-level navigation path.

[0071] In some embodiments, the memory map further includes: the vehicle owner's historical driving trajectory; and performs local geometric path search based at least on the global topological path and the road traffic information collected in real time by the mobile platform to determine the local geometric path, including:

[0072] Local geometric paths are determined by searching the global topology path, the vehicle owner's historical driving trajectory, and real-time road traffic information collected by the mobile platform. The road traffic information includes at least one of obstacles, road boundaries, and lane lines. The local geometric path includes local lane change information and local path curvature information. The local lane change information and local path curvature information are used by downstream modules (such as…). Figure 3 (As shown in the diagram) Lateral and longitudinal beyond-line-of-sight decisions are made. For example, lateral decisions are made based on local lane-change information to control whether the vehicle should change lanes; longitudinal beyond-line-of-sight decisions are made based on the curvature information of the local path (e.g., determining whether the vehicle should decelerate in advance and the speed to be controlled based on curvature information). It should be noted that the above utilization of local lane-change information and local path curvature information is merely an example, and this application does not limit it.

[0073] For example, if real-time road traffic information collected by the mobile platform indicates a lane segment that is impassable due to an obstacle, but this lane segment is located on the driver's historical driving trajectory, then the real-time road traffic information should be used as the standard, and the lane segment should be excluded from the local geometric path. If real-time road traffic information collected by the mobile platform indicates a passable lane segment, and this lane segment is located on the driver's historical driving trajectory, then this lane segment should be prioritized for planning within the local geometric path. If real-time road traffic information collected by the mobile platform indicates that the right side is a road boundary, but the driver's historical driving trajectory crosses the road boundary to the right, then the real-time road traffic information should be used as the standard, and the corresponding lane segment should be excluded from the local geometric path. If the lane line collected by the mobile platform in real-time is a solid line, but the driver's historical driving trajectory involves a lane change at that solid line, then the real-time road traffic information should be used as the standard, and the corresponding lane segment after the lane change should be excluded from the local geometric path.

[0074] In this embodiment, local geometric path search is performed based on global topology path, vehicle owner's historical driving trajectory and road traffic information collected in real time by mobile platform. This not only makes the obtained local geometric path conform to the actual road drivability, but also ensures that the local geometric path is as close as possible to the vehicle owner's historical driving trajectory.

[0075] In some embodiments, the memory map further includes a cost map, which includes the cost value corresponding to each lane segment. Exemplarily, this application maintains a cost map at the topology level. This cost map assigns each lane segment in the memory map a cost value representing the passage through that lane segment, and a cost value representing a lane change within that lane segment. A higher cost value indicates a lower likelihood of subsequently finding a path through that lane segment or a lower likelihood of a path causing navigation-induced lane changes within that lane segment. During path search, the topology layer searcher considers the cost values ​​in the cost map and searches for the path with the lowest cumulative cost value.

[0076] For example, the memory map includes a first lane segment, a second lane segment, and a third lane segment. The first and second lane segments are located in the same lane (e.g., the second lane on the left side of the road), and the third lane segment is located in another lane (e.g., the first lane on the left side of the road). The second and third lane segments are subsequent lane segments to the first lane segment (vehicles can travel from the first lane segment to the second or third lane segment. Traveling from the first lane segment to the third lane segment constitutes a lane change). Optionally, the lane change cost is set to a value of 1-10.

[0077] For example, when a vehicle is not desired to enter the second lane from the first lane segment, it means that the vehicle does not want to pass through the second lane segment. In this case, a larger passing value (e.g., 8) can be assigned to the second lane segment, and a smaller passing value (e.g., 3) can be assigned to the third lane segment. In this way, during the path search process, the topology layer searcher will prioritize searching for paths that pass through the third lane segment.

[0078] For example, if the recorded driving history of the vehicle owner shows the vehicle passing through both the first and third lane segments, then to better align with user habits, it is not recommended that the vehicle enter the second lane segment from the first lane segment; if vehicle perception information indicates that the second lane segment is unavailable (e.g., under maintenance or frequently parked), then it is not recommended that the vehicle enter the second lane segment from the first lane segment. These are merely examples; other situations are possible, and this application does not limit the scope of these cases.

[0079] When you do not want a vehicle to change lanes from the first lane segment to the third lane segment (i.e., from the second lane on the left to the first lane on the left), it means you do not want to change lanes in the first lane segment. In this case, you can assign a larger lane change cost to the first lane segment (e.g., 8). In this way, during the path search process, the topology layer searcher will prioritize searching for paths that do not involve lane changes in the first lane segment.

[0080] For example, if the first, second, and third lanes are lanes before and after an intersection, with the first lane being the lane entering the intersection and the second and third lanes being the lanes exiting the intersection, then for driving safety, it is not desirable for a vehicle to change lanes from the first lane to the third lane. Alternatively, if the third lane is the lane where a divergence point is located, and the vehicle needs to continue straight, then for faster passage, it is not desirable for the vehicle to change lanes from the first lane to the third lane. These are merely examples; other situations may exist, and this application does not limit these possibilities.

[0081] In some embodiments, the navigation path planning method further includes: preprocessing the weight values ​​of lane segments before and after the corresponding intersection point in the cost map to reduce the probability of vehicles changing lanes before and after the intersection point.

[0082] For example, preprocessing the weight values ​​of lane segments before and after an intersection point includes assigning a higher lane-change cost value to the lane segments before and after the intersection point to reduce the probability of vehicles changing lanes in these lane segments and avoid safety hazards caused by lane changes. For instance, lane segments located before and after the intersection point in the cost map are assigned a higher lane-change cost value as a weight. Optionally, lane segments located before and after the intersection point in the cost map are also assigned a lower passage cost value as a weight.

[0083] In some embodiments, the navigation route planning method further includes: preprocessing the weight values ​​of lane segments corresponding to the driver's historical driving trajectory in the cost map in order to generate a navigation route that conforms to the driver's habits.

[0084] For example, lane segments covered by the driver's historical driving trajectory in the cost map are assigned a smaller through-cost value as a weight value, so that when searching for a route, the navigation route that is closer to the driver's driving habits can be obtained based on the through-cost value.

[0085] In some embodiments, the navigation path planning method further includes: preprocessing the weight values ​​of the lane segments where the divergence points and merging points are located in the cost map, so as to control the situation of vehicles passing through the divergence points or merging points.

[0086] The terms "divergence point" and "merging point" refer to points including, but not limited to, "ramp divergence" and "ramp merging." In addition to these two, it also includes other divergence points where the number of lanes increases and merging points where the number of lanes decreases. These divergence and merging points typically occur before intersections. Furthermore, it also includes situations where temporary road closures due to road construction result in a decrease in the number of lanes or an increase in the number of lanes.

[0087] During navigation route planning, the preprocessing of lane segments before and after divergence and merging points varies depending on the road type. This is because on highways, vehicle speeds are typically high, requiring strong intervention in route planning to ensure safety when vehicles pass through divergence and merging points. This intervention can be achieved, for example, by preprocessing the weight values ​​of the corresponding lane segments. On ordinary roads, vehicle speeds are typically low, making it safer for vehicles to pass through divergence and merging points, thus eliminating the need for strong intervention in route planning.

[0088] For example, preprocessing the weight values ​​of the lane segments where divergence points and merging points are located in the cost map to control vehicle passage through the divergence points or merging points includes:

[0089] If the divergence or merging point is on a highway, the lane segment containing that divergence or merging point is assigned a higher throughput value in the cost map. This guides vehicles to change lanes into their target lane after the divergence point or to change lanes in advance before the merging point. For example, for a divergence point on a highway, the rightmost lane of the highway consists of three consecutive lane segments: the first lane segment, the second lane segment, and the third lane segment. The second lane segment is where the divergence point is located. In actual driving, there are usually slow-moving vehicles about to enter the ramp on the lane segment where the divergence point is located. Therefore, if the target lane of the vehicle is the rightmost lane, it is not desirable for the vehicle to enter the rightmost lane before the divergence point (because it will be suppressed by slow-moving vehicles about to enter the ramp, and there may even be a risk of rear-end collision). Therefore, in this embodiment, the second lane segment where the divergence point is located is assigned a higher throughput value.

[0090] For example, at a merging point on a highway, the rightmost lane comprises three consecutive lane segments: the first lane segment, the second lane segment, and the third lane segment. The second lane segment is where the merging point is located. In actual driving, vehicles accelerating from ramps to merge often pass through this lane segment. Therefore, if the vehicle's target lane is the rightmost lane, it is undesirable for the vehicle to enter the rightmost lane after the merging point (as this could lead to a collision with vehicles accelerating from ramps). Therefore, in this embodiment, the second lane segment where the merging point is located is assigned a higher throughput value.

[0091] If the divergence or merging point is a divergence or merging point on a regular road, then the lane segment where the divergence or merging point is located is assigned a smaller passage cost (e.g., 0) in the cost map to reduce the impact on vehicles passing through the divergence or merging point.

[0092] In some embodiments, the navigation path planning method further includes: preprocessing the weight values ​​of the corresponding lateral lane segments in the cost map based on road information and navigation information, so as to reduce the probability of a vehicle traveling in the lateral lane.

[0093] For example, road information includes road width information and road diversion information, and navigation information includes situations where the vehicle needs to travel along the current road (e.g., needing to travel a long distance or needing to travel along the main road).

[0094] In some embodiments, the weight values ​​of the corresponding lateral lane segments in the cost map are preprocessed based on road information and navigation information to reduce the probability of vehicles traveling in the lateral lanes, including:

[0095] When the road width is determined to be greater than the width threshold (e.g., the width threshold is 25m) based on the road width information, and the distance to be traveled on the current road exceeds the distance threshold (e.g., the distance threshold is 3km), the rightmost lane segment near the non-intersection is assigned a higher passage cost in the cost map; when the road divergence information is determined to be a left divergence, and the vehicle needs to travel along the main road, the lane segment before the divergence point is assigned a higher passage cost.

[0096] In this embodiment, when the road width exceeds the road width threshold, the current road is generally considered to be an urban arterial road. The rightmost lane typically has more merging and diverging points, and vehicles in this lane usually travel at lower speeds and have poor maneuverability. Therefore, a higher maneuverability value is assigned to the rightmost lane segment, as vehicles are discouraged from using it. Furthermore, it should be noted that the reason this embodiment only assigns a higher maneuverability value to the rightmost lane segment not near intersections is because the rightmost lane near intersections usually has better maneuverability compared to other lanes near intersections (because the rightmost lane at an intersection is usually a right-turn + straight-ahead lane, and right turns are usually not restricted by red lights, thus vehicles in the rightmost lane near intersections have better maneuverability).

[0097] In addition, although the leftmost lane is usually the fast lane, when there is a left-hand divergence on the road, the lane segment where the divergence point is located usually has a lot of slow-moving vehicles, and the vehicle passability is low. Therefore, the leftmost lane gives the lane segment before the divergence point a higher passability value.

[0098] In some embodiments, at least the weight values ​​of impassable lane segments in the cost map are preprocessed to restrict vehicle passage. Impassable lanes include, but are not limited to, reverse lane segments, lane segments with poor mapping quality, and lane segments where temporarily parked vehicles are present for extended periods.

[0099] For example, a higher passage cost can be assigned to reverse lane segments, lane segments with poor mapping quality, and lane segments with long-term temporary parking vehicles in the cost map. This avoids planning these three types of lane segments into the global path when performing global topology path search, thereby restricting vehicles from passing through these three types of lane segments and ensuring driving safety.

[0100] Optionally, the lane segments before and after the intersection point, the lane segments of the driver's historical driving trajectory, the lane segments where the divergence / merging point is located, the most lateral lane segments, and the impassable lane segments in the aforementioned embodiments can be determined and recorded in combination with the perception information obtained by the vehicle during the construction of the memory map. The specific identification method can refer to the relevant prior art, and this application does not limit it.

[0101] like Figure 2The diagram shown is a flowchart of another embodiment of the navigation path planning method of this application. In this embodiment, the navigation path planning method further includes:

[0102] S50. Determine whether the corresponding lane is a curve or a U-turn lane based on the curvature value.

[0103] For example, the memory map stores the curvature values ​​of each lane segment. During navigation route planning, the curvature values ​​of the lane segments are used to determine the curvature of the lane segments, thereby determining whether the lane segment is a curve or a U-turn lane. The curvature values ​​of each lane segment are calculated during the construction of the memory map. For example, they can be calculated based on lane lines photographed during the mapping process; the specific algorithm is not limited in this application.

[0104] For example, the type of a lane segment (curve or U-turn lane) is determined by judging the magnitude of its curvature value. For instance, the curvature range corresponding to a curve and the curvature range corresponding to a U-turn lane can be predetermined, and the lane type can be determined by judging which curvature range the lane segment falls into. For example, when the curvature of a lane segment falls into the curvature range of a curve, the lane segment is determined to be a curve; when the curvature of a lane segment falls into the curvature range of a U-turn lane, the lane segment is determined to be a U-turn lane. The specific values ​​of the curvature range of the curve and the curvature range of the U-turn lane can be determined based on the analysis and statistical analysis of a large number of collected curvature values ​​of curves and U-turn lanes; the specific data is not limited in this application.

[0105] S60. For curves or U-turn lanes, calculate the current longitudinal distance from the curve or U-turn lane and / or calculate the maximum speed limit of the vehicle based on the corresponding curvature value.

[0106] For example, for lane segments identified as curves or U-turn lanes, the current longitudinal distance from the curve or U-turn lane can be further calculated. For instance, the vehicle's latest position is determined in real-time based on its location information, and the location information of the determined curve or U-turn lane (e.g., determined from a memory map) is used to calculate the vehicle's current longitudinal distance from the curve or U-turn lane. This longitudinal distance serves as a reminder to the user of the vehicle's current distance from the curve or U-turn lane, allowing the user to apply speed limits. The location information of the curve or U-turn lane can be the location information of the starting point of the lane segment corresponding to the curve or U-turn lane.

[0107] For example, for lane segments determined to be curves or U-turn lanes, the maximum speed limit for vehicles can be further determined based on the curvature value of the lane segment. For instance, when the curvature value corresponds to a highway ramp, the maximum speed limit for vehicles can be determined to be 60 km / h; when the curvature value corresponds to a right / left turn lane at an intersection, the maximum speed limit for vehicles can be determined to be 15 km / h; and when the curvature value corresponds to a U-turn lane, the maximum speed limit for vehicles can be determined to be 5 km / h.

[0108] For example, the memory map can pre-store the curvature values ​​or curvature ranges of each lane segment, the lane types corresponding to each curvature value or curvature range (e.g., highway ramps, left / right turn lanes at intersections, U-turn lanes), and the maximum speed limits corresponding to each lane type (or the maximum speed limits corresponding to each curvature value or curvature range), thereby reminding drivers to control their speed in real time, realizing beyond-line-of-sight speed limit reminders, and ensuring safe driving.

[0109] It should be noted that the above method for determining the maximum speed limit of a vehicle based on the curvature value of a lane segment is only an example, and other implementation methods can also be used, which are not limited in this application.

[0110] like Figure 3 The diagram shown is a flowchart of an embodiment of the navigation path planning method of this application, which includes the following steps:

[0111] Step 1. Input:

[0112] Input the vehicle's online perception information and pre-built memory maps.

[0113] Step 2. Preprocess the memory map:

[0114] The input memory map is preprocessed to identify the lane segments containing the start and end points, and to mark lane segments that are not intended for entry into or lane changes to be made. Lane segments that are not intended for entry into include, but are not limited to, lane segments at divergence points, lane segments at merging points, oncoming lane segments, lane segments with poor mapping quality, or lane segments where vehicles frequently stop temporarily. Lane segments that are not intended for lane changes include, but are not limited to, lane segments near intersections.

[0115] Step 3. Topology layer search:

[0116] A global topological path search is performed on the preprocessed memory map to obtain a coarse-grained global topological path;

[0117] Step 4. Geometry layer search:

[0118] Based on the global topological path and combined with the results of online sensing, a geometric layer search is performed within the road area to obtain a fine-grained local geometric path;

[0119] Step 5. Post-processing:

[0120] By combining memory maps, topological trajectories, geometric trajectories, and online perception, the system outputs lane change information needed for lateral decision-making and speed limit information needed for longitudinal beyond-line-of-sight decision-making.

[0121] For step 1:

[0122] A memory map is an abstract map with a much smaller coverage area than a high-precision map, obtained using onboard sensors (including but not limited to binocular cameras, monocular cameras, radar sensors, etc.) and mapping algorithms. Since car owners frequently use relatively fixed routes and drive on relevant road sections, the memory map serves as a substitute for a high-precision map of commonly used routes built by the vehicle itself. The memory map consists of interconnected lane segments that accurately reflect the road structure of the mapped area, and the combination of these lane segments forms an interconnected topological map. The mapping process of the memory map can record the location of the car owner's driving trajectory within the memory map, making it easier for the planner (e.g., a topology layer searcher) in this application to generate navigation paths that approximate the car owner's driving habits. For example, during mapping and subsequent use, the routes driven by the car owner each time are recorded. The planner described in this application will assign lower cost values ​​to frequently driven lane segments when constructing the cost map, so the paths generated by subsequent searches will be more likely to guide the vehicle into lanes that the car owner is accustomed to using.

[0123] Online perception can cover an area within 1-500m of the vehicle and provide real-time road traffic information (including but not limited to obstacles, road boundaries, and lane lines), which effectively compensates for the potential update lag problem of memory maps.

[0124] For step 2:

[0125] like Figure 4 The diagram shown is an embodiment of preprocessing a memory map in this application. The diagram includes lane segments and the vehicle's location as contained in the memory map. For lane lines that are dashed, lane-changing restrictions are relaxed; for lane lines that are solid, lane-changing is discouraged.

[0126] The memory map preprocessing steps in this application can be performed by a memory map preprocessor, which can be a physical processor chip or a virtual program module; this application does not limit this. The memory map preprocessor processes specific lane segments in the memory map as follows:

[0127] 1. Pre-treatment of lane sections before and after the intersection:

[0128] By preprocessing the lane segments before and after key route points such as intersections, the intention of vehicles to change lanes in that area is reduced, thus avoiding lane-changing behavior that crosses multiple lanes before and after intersections.

[0129] For example, preprocessing the lane segments before and after the intersection point includes:

[0130] 1.1) Determine if the intersection needs to be addressed: Ignore any consecutive small intersections.

[0131] 1.2). Determine if the intersection is a multi-to-short intersection: If it is a multi-to-short intersection (such as the Sanheyi intersection), then use curvature to determine whether the lanes of each section of the intersection are non-directly aligned or directly aligned. If there are directly aligned lanes, then try to guide vehicles into the directly aligned lanes.

[0132] 1.3) Iterate through the lane segments before the intersection from near to far, and assign a penalty for changing lanes: so that vehicles will be penalized if they change lanes before the intersection;

[0133] 1.4) Iterate over the lane segment after the intersection, using the same method as step 1.3), but consider a shorter distance, generally within 200 meters. The specific value is determined by the distance between the two intersections. If the distance between the two intersections is very close (e.g., 300 meters), this restriction is not applied. If the distance is very far (e.g., 2 kilometers), then lane changing behavior after passing through the intersection is restricted.

[0134] The memory map preprocessor described in this application assigns a higher lane-change cost to lane segments before intersections, thus suppressing vehicles' willingness to change lanes before intersections. If multiple lane changes occur before an intersection, multiple high lane-change costs will accumulate. Therefore, from the navigation's perspective, lane-change behavior that crosses multiple lanes before an intersection is too costly, and it will not tend to generate such paths.

[0135] 2. Pre-treatment of lane sections at merging and diverging points:

[0136] By preprocessing the lane segments where the divergence and merging points are located, vehicles can be guided to pass through or avoid passing through the divergence and merging points.

[0137] For example, divergence and merging points refer to points including, but not limited to, "ramp divergence" and "ramp merging." In addition to the above two, they also include divergence points where the number of lanes increases and merging points where the number of lanes decreases. These divergence and merging points usually appear before intersections. In addition, they may also include situations where the number of lanes increases or decreases due to temporary road closures caused by road construction.

[0138] For example, the preprocessing of the lane segment where the divergence and merging points are located includes the following steps:

[0139] 2.1) Identify the branching and merging points on the map;

[0140] 2.2) If the divergence or merging points on the highway tend to be assigned a higher "cost of passing through the location", vehicles are guided to change lanes into the target lane after the divergence point or to change lanes in advance before the merging point;

[0141] 2.3) If it is a divergence or merging point near an intersection on a regular road, then assign a value of 0 to "cost of passing through this location".

[0142] In the cost map of this application, if a vehicle is particularly unwilling to pass through a certain merging point, such as a section of a highway temporarily closed due to road construction, then the merging point before that section is assigned a particularly high "cost value". In addition, it is indicated that the driving distance of this section of the lane is relatively short. Then, the vehicle will have a very strong desire to change lanes before approaching such a merging point. This can avoid the danger caused by the vehicle changing lanes when it is close to the merging point or being unable to change lanes due to being suppressed by high-speed vehicles in the target lane.

[0143] 3. Pre-treatment of the most lateral lane segment:

[0144] By preprocessing the rightmost or leftmost lane in a multi-lane road, the intention of vehicles to travel long distances in the outermost lane is reduced, thereby increasing the intention of vehicles to travel in the fast lane.

[0145] For example, if the road is very wide and the vehicle needs to travel a long distance straight along the road, the rightmost lane near the intersection is assigned a higher "cost value" in the cost map; if there is a left shunting and the vehicle needs to travel straight along the main road, the lanes within a certain distance before the left shunting point are assigned a higher "cost value".

[0146] In this embodiment, the cost is achieved by assigning a higher "cost value" to the corresponding part in the cost map: when the cost value of the most lateral lane is higher, the cost value of the corresponding other lanes is relatively lower, so vehicles will tend to drive into areas with lower cost values ​​according to navigation.

[0147] 4. Pre-treatment of curves and U-turn lanes:

[0148] By preprocessing areas such as curves and U-turn lanes in the region, the system can prompt users to pay attention and limit their speed in advance when vehicles approach.

[0149] For example, specific preprocessing steps include:

[0150] 4.1) Determine whether the lane section is a high-curvature curve or a U-turn lane based on the local curvature;

[0151] 4.2) Calculate the maximum desired vehicle speed based on the local curvature;

[0152] 4.3) Send a marker position to the vehicle's infotainment system within a certain distance before the target lane, so that the vehicle's infotainment system can remind the user;

[0153] 4.4) The maximum speed of this lane is marked according to the curvature, and the longitudinal distance to this lane is given. The specific speed limit behavior is determined and executed by the downstream module.

[0154] In this embodiment, a marker is sent to the vehicle's infotainment system a certain distance before the target lane, and the infotainment system reminds the user (e.g., by controlling the microphone of the vehicle's infotainment system to remind the user to slow down); the maximum speed limit for this lane segment is marked according to the curvature, and the longitudinal distance to this lane segment is given. The specific speed limit behavior is determined by the downstream module (see...). Figure 3 ) Judgment and execution.

[0155] 5. Preprocessing of lanes in the mapping trajectory:

[0156] By preprocessing the lane segments frequently traversed by the user's driving trajectory during map creation, the topology route planning takes into account the user's driving habits and generates navigation routes that closely match the user's driving style.

[0157] For example, specific preprocessing steps include:

[0158] 5.1) During the mapping process, the route traveled by the vehicle will be recorded, which is a series of specific GPS point sequences;

[0159] 5.2) By comparing the GPS point sequence with the memory map, it is possible to determine which lanes the driver usually drives in;

[0160] 5.3) By assigning lower values ​​to lane segments that drivers are accustomed to driving through, lane-level navigation routes that closely match user habits can be obtained through search, thus achieving the above objectives.

[0161] It should be noted that the preprocessing results described in this application are all soft constraints. Therefore, even if there are errors or imperfections in the preprocessing rules, the performance of the navigation module of intelligent assisted driving may be reduced, but there will still be no erroneous navigation signals such as missing route points or wrong lane changes.

[0162] For example, during map construction, there might be instances of false or missed intersections, or lane confusion. However, any search algorithm essentially traverses the entire map in a certain order. Therefore, if the starting and ending points are connected, both can be traversed, and the subsequent backtracking can generate a navigation path. Even if a lane segment is incorrectly marked with a high cost in the cost map, it will only result in a higher overall cost for the searched path, not prevent the path from being generated altogether. Therefore, it can still guide vehicles to "not miss route points."

[0163] For steps 3-4:

[0164] The topology layer search and geometry layer search of this application can be executed by a planner, which further includes a topology layer searcher and a geometry layer searcher. The planner, topology layer searcher and geometry layer searcher can be implemented by hardware chips or as software program modules. This application does not limit the implementation in this regard.

[0165] For example, the planner's search process is divided into two layers, the first layer being the topology search performed by the topology layer. For example... Figure 5 The diagram illustrates the topology-level path planning in this application, where the path indicated by the purple arrow is the planned path. Topology search is performed on a memory map. A connected topology path is obtained between the starting and ending lane segments through searching. This topology path is a sequential combination of a series of lane segments. The memory map includes a lane-level topology map and a cost map. Nodes in the lane-level topology map represent lane segments, and edges represent the connections between lane segments. The cost map includes the through cost and / or lane change cost for each lane segment. Optionally, topology-level search simultaneously searches for a globally lower-cost topology path based on both the topology map and the cost map.

[0166] The topology search process uses a classic graph search algorithm. Therefore, as long as the connectivity of the memory map is guaranteed, the planner described in this patent can provide lateral lane change decision guidance information during vehicle travel, guiding the vehicle not to miss route points and avoiding erroneous lane changes.

[0167] Since the topology path output by the topology search is a serialized combination of lane segments, including the vehicle's location in the memory map, the final destination, whether a lane change or continuous lane change is needed, the remaining distance if no lane change is needed, and the lane segments suggested to be traversed throughout the navigation process, the granularity of the output of this module is relatively coarse. Therefore, a more refined geometric search is required.

[0168] like Figure 6The diagram illustrates the path planning at the geometric layer in this application, including the vehicle's position, topological path, geometric path, and lane segments in the memory map. Because the memory map and topological path are too coarse-grained, they can only be used for coarse-grained decision-making and navigation. Therefore, the inventors designed a second layer for the planner described in this application: the geometric layer searcher. The geometric layer searcher combines the results of the vehicle's online perception to obtain a real-time, finer-grained environmental representation. The geometric search combines information such as the topological path, the solid and dashed lines of the lanes, and the memory trajectory (the route traveled when the vehicle owner uses the vehicle to create a memory map; in implementation, this route is recorded using a broken line, hence the "memory trajectory") to generate a geometric path that conforms to the vehicle's kinematic constraints. The geometric path can be viewed as a local trajectory that navigation aims for the vehicle to travel. This trajectory conforms to certain kinematic constraints, including local lane-changing information and the curvature information of the local trajectory, which will affect lateral and longitudinal decisions.

[0169] Geometric search uses a search tree to explore a local space. During the growth of the search tree, it incorporates perceived information such as obstacles, road boundaries, and lane lines to calculate a cost for each node. The search tree generated during the search process can be viewed as an abstract representation of the expanded region. This offers finer granularity and higher resolution compared to lane segment combinations obtained through topological search.

[0170] Optionally, each node in the search tree corresponds to a real point in a two-dimensional plane and contains the position information of each point (e.g., two-dimensional coordinates x and y, and the θ angle, where the θ angle is the angle between the vehicle's direction of travel and the X-axis in the two-dimensional coordinate system).

[0171] Furthermore, the starting parent node in the search tree can be the vehicle's origin (e.g., the vehicle's current position). Then, using this starting parent node as a reference, subsequent child nodes are searched based on information obtained from perception, such as obstacles, road boundaries, and lane lines, and each subsequent child node is assigned a value. Further, these subsequent child nodes are prioritized according to their value, with lower values ​​indicating higher priority. The child node with the lowest value is then iterated to become the new parent node for the next round of searching, and so on, achieving local path search within the vehicle's perception range.

[0172] Optionally, subsequent child nodes can be searched based on information obtained from combined perception, such as obstacles, road boundaries, and lane lines, and each subsequent child node can be assigned a value, including:

[0173] Determine the edges connecting the parent node to the subsequent child node and whether the subsequent child node collides with an obstacle. If so, assign a first-generation value to the subsequent child node. And / or, determine the edges connecting the parent node to the subsequent child node and whether the subsequent child node collides with a road boundary. If so, assign a second-generation value to the subsequent child node. And / or, determine whether the edges connecting the parent node to the subsequent child node intersect with a lane line. If so, assign a third-generation value to the subsequent child node. Finally, determine the generation value corresponding to the subsequent child node based on the first-generation value and / or the second-generation value and / or the third-generation value.

[0174] For example, determining whether the edge connecting the parent node to the subsequent child node and whether the subsequent child node collides with an obstacle includes: determining whether the distance between the two-dimensional coordinates (x, y) of the subsequent child node and the two-dimensional coordinates of the obstacle is within a preset range; if so, the subsequent child node is considered to have collided with the obstacle; determining whether the edge connecting the parent node to the subsequent child node passes through the location of the obstacle (for example, it can be determined whether the distance from the two-dimensional coordinates of the obstacle to the edge connecting the parent node to the subsequent child node is within a preset range; if so, it indicates that the obstacle has been passed through), if so, it is determined that the edge connecting the parent node to the subsequent child node has collided with the obstacle. Optionally, the preset range can be determined based on the size of the obstacle and the size of the vehicle; this application does not limit this.

[0175] For example, determining whether the edge connecting the parent node to the subsequent child node and whether the subsequent child node collides with the road boundary includes: determining whether the two-dimensional coordinates (x, y) of the subsequent child node are on the road boundary, and if so, considering that the subsequent child node collides with the road boundary; determining whether the edge connecting the parent node to the subsequent child node intersects with the road boundary, and if so, considering that the subsequent child node collides with the road boundary.

[0176] For example, determining whether the edge connecting the parent node to the subsequent child node intersects the lane line includes: determining whether the parent node and the subsequent child node are distributed on both sides of the lane line based on the two-dimensional coordinates of the parent node and the two-dimensional coordinates (x, y) of the subsequent child node; if so, it is considered that the edge connecting the parent node to the subsequent child node intersects the lane line.

[0177] For example, subsequent child nodes include a first child node, a second child node, and a third child node. The cost values ​​of the first child node, the second child node, and the third child node are calculated accordingly based on the above method. If the cost value of the first child node is the smallest, then the corresponding first child node is determined as the new parent node for subsequent searches.

[0178] For example, for the first child node, the first-generation value, second-generation value, and third-generation value of the first child node are calculated according to the method described in the foregoing embodiments. Optionally, a preset calculation method is used to determine the first child node generation value of the first child node based on the first-generation value, second-generation value, and third-generation value. For example, the first child node generation value of the first child node is obtained by weighted summation of the first-generation value, second-generation value, and third-generation value. It should be noted that the above is only an example, and this application does not limit the preset calculation method for calculating the cost value of the first child node. Similarly, the second child node generation value of the second child node and the third child node generation value of the third child node can be calculated.

[0179] For example, this application uses a bicycle model to model the vehicle. The bicycle model is a simplified representation of vehicle dynamics used to simulate and analyze the vehicle's handling behavior. This model simplifies the vehicle to two wheels to reduce computational complexity and capture its fundamental dynamic characteristics. During the search tree growth process to find child nodes, the vehicle's wheel-direction angle and acceleration are deterministically and uniformly sampled, and the propagation process yields the child nodes and the corresponding connections between parent and child nodes.

[0180] The longitudinal decision-making process uses the curvature of the trajectory obtained from the geometric search to determine whether to slow down in advance, and whether to limit the speed within a certain range based on whether this trajectory passes through a divergence / merging point. Information from divergence / merging points also influences lateral decision-making. For example, in highway scenarios, lateral decision-making tends to favor changing lanes after a divergence point to enter the desired lane, rather than driving along the divergence point. In urban scenarios, the opposite is true; the lateral decision-making module will choose to drive past the divergence point to enter the target lane, rather than changing lanes.

[0181] For step 5:

[0182] The core of autonomous driving remains decision-making. The decision-making module cannot directly utilize the output results in the three-dimensional coordinate system, such as the global topology path and local geometric path described in this application. Therefore, post-processing is required to generate navigation semantic information that can be directly used by the lateral and longitudinal decision-making modules. For example, in areas where continuous lane changes are necessary, the solution in this application sends information to the downstream lateral and longitudinal decision-making modules, indicating that multiple lanes need to be crossed consecutively within a short distance to avoid missing the route point. The downstream lateral and longitudinal decision-making modules then strengthen their willingness to change lanes laterally and begin to focus on limiting longitudinal speed to avoid missing the route point.

[0183] Post-processing could include, for example, generating the remaining longitudinal distance along each lane and the number of lane changes required laterally, to avoid missing routing points (e.g., branching points). The remaining longitudinal distance along each lane and the number of lane changes required laterally constitute navigation semantic information.

[0184] Optionally, the post-processing described above involves determining the driving task to be performed by the vehicle based on the global topology path and local geometric path obtained in the preceding embodiments. The driving task includes, but is not limited to, the longitudinal distance the vehicle needs to travel along each lane and the number of lane changes the vehicle needs to complete.

[0185] like Figure 7 The diagram illustrates how post-processing is used to determine the longitudinal travel distance and lateral lane change frequency of a vehicle in this application. This example corresponds to a two-to-three lane splitting road. Each rectangle represents a road segment, with the starting point located in the leftmost lane (corresponding to the upper side in the attached diagram). The yellow rectangles represent the topological path, and the furthest possible position a vehicle can reach along its lane is marked in red in the attached diagram. In this embodiment, the two-to-three lane splitting road consists of two main roads and one branching auxiliary road, and the main and auxiliary road lanes separated by sight distance are marked in the attached diagram. Vehicles need to make two lane changes while traveling along the topological path.

[0186] like Figure 7 As shown, the post-processing steps of this application calculate the driving tasks that the vehicle needs to perform from the starting point to the end point. These driving tasks include two lane changes and determining the furthest achievable position along each lane (which can also be converted into the longitudinal distance traversable along each lane). Furthermore, when approaching a high-curvature curve at a certain distance (e.g., 200m away), the post-processing steps calculate the curvature and distance of the upcoming curve based on the global topology path and local geometric path. The downstream module then uses this output to make longitudinal decisions and guide the vehicle safely through the target curve.

[0187] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of combined actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Secondly, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application. In the above embodiments, the descriptions of each embodiment have their own emphasis; for parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0188] In some embodiments, this application provides a computer device including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the navigation path planning method described in any embodiment of this application.

[0189] In some embodiments, this application provides a computer-readable storage medium storing a computer program / instructions thereon, which, when executed by a processor, implements the steps of the navigation path planning method described in any embodiment of this application.

[0190] In some embodiments, this application provides a computer program product including a computer program / instructions that, when executed by a processor, implement the steps of the navigation path planning method described in any embodiment of this application.

[0191] In some embodiments, this application provides a mobile platform on which the computer device described in any embodiment of this application is installed.

[0192] Figure 8 This is a schematic diagram of the hardware structure of a computer device for executing a navigation path planning method according to another embodiment of this application, as shown below. Figure 8 As shown, the device includes:

[0193] One or more processors 810 and memory 820, Figure 8 Take the 810 processor as an example.

[0194] The device for performing the navigation path planning method may also include an input device 830 and an output device 840.

[0195] The processor 810, memory 820, input device 830, and output device 840 can be connected via a bus or other means. Figure 8 Taking the example of a connection between China and Israel via a bus.

[0196] The memory 820, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules, such as the program instructions / modules corresponding to the navigation path planning method in the embodiments of this application. The processor 810 executes various functional applications and data processing of the server by running the non-volatile software programs, instructions, and modules stored in the memory 820, thereby implementing the navigation path planning method of the above-described method embodiments.

[0197] The memory 820 may include a program storage area and a data storage area. The program storage area may store the operating system and applications required for at least one function; the data storage area may store data created based on the use of the navigation path planning device. Furthermore, the memory 820 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 820 may optionally include memory remotely located relative to the processor 810, and these remote memories may be connected to the navigation path planning device via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0198] Input device 830 can receive input digital or character information and generate signals related to user settings and function control of the navigation path planning device. Output device 840 may include display devices such as a display screen.

[0199] The one or more modules are stored in the memory 820, and when executed by the one or more processors 810, they execute the navigation path planning method in any of the above method embodiments.

[0200] The above-described product can perform the methods provided in the embodiments of this application, and has the corresponding functional modules and beneficial effects for performing the methods. Technical details not described in detail in this embodiment can be found in the methods provided in the embodiments of this application.

[0201] The computer device in this application embodiment exists in various forms, including but not limited to:

[0202] (1) Mobile communication devices: These devices are characterized by their mobile communication capabilities and primarily aim to provide voice and data communication. These terminals include: smartphones (e.g., iPhones), multimedia phones, feature phones, and low-end phones, etc.

[0203] (2) Ultra-mobile personal computer devices: These devices fall under the category of personal computers, possessing computing and processing capabilities, and generally also have mobile internet access features. These terminals include PDAs, MIDs, and UMPCs, such as the iPad.

[0204] (3) Portable entertainment devices: These devices can display and play multimedia content. This category includes audio and video players (such as iPods), handheld game consoles, e-book readers, as well as smart toys and portable car navigation devices.

[0205] (4) Server: A device that provides computing services. The components of a server include a processor, hard disk, memory, system bus, etc. Servers are similar to general computer architectures, but because they need to provide highly reliable services, they have higher requirements in terms of processing power, stability, reliability, security, scalability, and manageability.

[0206] (5) Other electronic devices with data interaction functions.

[0207] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0208] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0209] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for planning a navigation path, comprising: obtaining a current position and a destination position of a mobile platform; determining a corresponding memory map according to the current position and the destination position of the mobile platform, the memory map comprising a lane-level topological map from the current position to the destination position; performing a global topological path search according to at least the lane-level topological map to determine a global topological path from the current position to the destination position; performing a local geometric path search according to at least the global topological path and road traffic information collected in real time by the mobile platform to determine a local geometric path.

2. The method of claim 1, wherein, The memory map further comprises a driver historical driving trajectory. The step of planning a lane-level navigation path from the current position to the destination position according to at least the lane-level topological map comprises: performing a global topological path search according to the lane-level topological map and the driver historical driving trajectory to determine a global topological path from the current position to the destination position.

3. The method of claim 1, wherein, The memory map further comprises a driver historical driving trajectory. The step of performing a local geometric path search according to at least the global topological path and road traffic information collected in real time by the mobile platform to determine a local geometric path comprises: performing a local geometric path search according to the global topological path, the driver historical driving trajectory and road traffic information collected in real time by the mobile platform to determine a local geometric path.

4. The method according to any one of claims 1 to 3, characterized in that, The memory map further comprises a cost map, the cost map comprising cost values corresponding to each lane segment.

5. The method of claim 4, wherein, The method further comprises: preprocessing weight values of lane segments before and after a corresponding intersection point in the cost map to reduce the probability of lane changing of a vehicle before and after the intersection point; and / or, preprocessing weight values of lane segments corresponding to the driver historical driving trajectory in the cost map to generate a navigation path in line with the driver's habits; and / or, preprocessing weight values of lane segments where a split point and a merge point are located in the cost map to control the situation of a vehicle passing through the split point or the merge point; and / or, preprocessing weight values of corresponding most lateral lane segments in the cost map according to road information and navigation information to reduce the probability of a vehicle driving in the most lateral lane; and / or, preprocessing weight values of at least non-passable lane segments in the cost map to limit the passing of a vehicle.

6. The method according to any one of claims 1-3, characterized in that, The memory map further comprises curvature values of lane segments, and the method further comprises: determining whether a corresponding lane is a curved lane or a U-turn lane according to the curvature values; for a curved lane or a U-turn lane, calculating a longitudinal distance from the curved lane or the U-turn lane and / or calculating a maximum speed limit of a vehicle according to a corresponding curvature value.

7. A computer device comprising a memory, a processor, and a computer program stored on the memory, wherein the computer program comprises instructions that, when executed by the processor, cause the processor to perform the method of any one of claims 1-6. The processor executes the computer program to implement the steps of the method of any one of claims 1-6.

8. A computer readable storage medium having stored thereon computer programs / instructions, characterized in that, The computer program / instruction is executed by the processor to implement the steps of the method of any one of claims 1-6.

9. A computer program product comprising computer programs / instructions, characterized in that, The computer program / instruction is executed by the processor to implement the steps of the method of any one of claims 1-6.

10. A mobile platform, characterized by The computer device of claim 7 is installed. The computer device of claim 7 is installed.

Images