Route generation method based on crossroad lane-level map acquisition

Abstract

The invention relates to the technical field of lane-level map acquisition, and discloses a route generation method based on crossroad lane-level map acquisition, which is characterized in that all lanes at an intersection are driven according to a generated driving route, and effective GPS lane data are recorded, and comprises the following steps: step 1, obtaining the length of each lane at the current intersection, and calculating the connection relationship between the lanes; 2, the shortest distance between every two lanes is calculated, and a weighted complete directed graph of the necessary lanes is constructed; 3, calculating the shortest path for accessing all the necessary lanes by using a dynamic rule, and forming a driving sequence of the necessary lanes; and step 4, filling a transition lane, generating a complete optimal path sequence, and outputting a driving route. According to the invention, the vehicle can acquire the map data of all lanes within the shortest distance, so that the driving of repeated lanes is reduced, the data acquisition efficiency is greatly improved, the congestion risk is reduced, and the energy consumption is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of lane-level map acquisition, and particularly to a route generation method for lane-level map acquisition based on intersections. Background Art

[0002] Currently, in the map acquisition method of the non-lane-level navigation technology, due to the lack of precise lane-level path planning, the acquisition vehicle often needs to rely on the driver to judge the driving route, resulting in a long and random driving distance. This not only increases the probability of ineffective acquisition, but also may reduce the overall data acquisition efficiency due to repeated driving on the same lane. In addition, this inefficient acquisition mode will significantly extend the operation time, not only wasting manpower and material resources, but also easily causing fatigue of the acquisition personnel due to long-time driving, affecting operation safety. At the same time, in busy traffic areas, especially at key nodes such as intersections, the repeatedly driving acquisition vehicles may interfere with the normal traffic order and exacerbate road congestion. Moreover, the redundant driving mileage not only increases the consumption of energy such as gasoline, but also accelerates the wear of the acquisition vehicle, further increasing the operation cost. Therefore, the existing technology urgently needs to design a lane-level acquisition path planning method to reduce ineffective driving, improve data acquisition efficiency, and reduce the impact on traffic and the environment. Summary of the Invention

[0003] The present invention aims to provide a route generation method for lane-level map acquisition based on intersections, so as to solve the problems of low efficiency in map acquisition of the existing navigation technology, easy formation of repeated acquisition paths, time-consuming and laborious while affecting traffic, and causing energy waste.

[0004] To achieve the above object, the present invention adopts the following technical solution. A route generation method for lane-level map acquisition based on intersections drives through all lanes of the intersection according to the generated driving route and records the effective GPS lane data, including:

[0005] Step 1, obtain the length of each lane of the current intersection and calculate the connection relationship between each lane;

[0006] Step 2, calculate the shortest distance between each lane and construct a weighted complete digraph of the necessary lanes;

[0007] Step 3, use the dynamic rule to calculate the shortest path to visit all necessary lanes and form the driving order of the necessary lanes;

[0008] Step 4, fill in the transition lanes, generate the complete optimal path order, and output the driving route.

[0009] The principle and advantages of this solution are:

[0010] Traditional map data collection is usually based on road centerlines or fixed trajectory planning, unable to identify the connection relationships and turning constraints between lanes, resulting in repeated driving. Moreover, existing navigation algorithms rely on static road network data and cannot handle lane-level micro-topologies (such as solid and dashed lines, variable guide lanes), with fixed weights, leading to the invalidation of the planned results due to lane attribute conflicts during actual collection, making the driving paths obtained in the prior art have a high degree of repetition and difficult to meet the accuracy requirements.

[0011] This solution provides a route generation method based on lane-level map collection at intersections, creatively constructing a three-layer architecture of "lane - connection relationship - dynamic weight". Each lane is abstracted as a directed graph node, and the edge weights are dynamically calculated through lane length, turning restrictions, etc., accurately quantifying the driving cost. By constructing a lane-level directed graph model and applying a dynamic programming optimization algorithm, a global optimal plan for the data collection path of multiple lanes at intersections is achieved. At the same time, this solution adopts a two-stage method of preferential coverage of necessary lanes + filling of transition lanes. In the first stage, the problem is transformed into the Hamiltonian path problem of a weighted complete directed graph, and the minimum-cost necessary lane sequence is generated through dynamic programming; in the second stage, transition lane constraints are introduced to ensure that the path complies with traffic regulations and is physically reachable.

[0012] Implementing the present invention has the following advantages:

[0013] 1. Through lane length weighting and turning cost modeling, the total path mileage is reduced by more than 30% compared to conventional collection, directly reducing fuel consumption and labor costs, and greatly saving resources;

[0014] 2. By setting a weighted complete directed graph and setting the legality verification of U-turn points, the risk of illegal U-turns in the traditional method is avoided. At the same time, the number of lane changes is minimized, reducing the probability of traffic accidents caused by frequent lane changes and improving safety; and the data integrity can be ensured through forced full coverage of the directed graph, avoiding missed collection;

[0015] 3. The output driving route can avoid continuously occupying the same lane, avoiding the congestion chain reaction caused by the low-speed driving of the collection vehicle and reducing the impact on traffic;

[0016] 4. The topological modeling and dynamic weights of this solution can adaptively generate feasible paths, and can quickly generate the optimal driving path for deformed intersections, improving the universality for complex intersections. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 It is a schematic flow chart of a route generation method based on lane-level map collection at intersections according to the present invention;

[0018] Figure 2 It is a schematic diagram of the intersection structure in a route generation method based on lane-level map collection at intersections according to the present invention;

[0019] Figure 3 It is a schematic diagram of the structure of a single intersection in a route generation method based on lane - level map collection at intersections of the present invention;

[0020] Figure 4 It is a schematic diagram of the structure of lane route identification in a route generation method based on lane - level map collection at intersections of the present invention;

[0021] Figure 5 It is a fully - directed graph with the right of 4 necessary lanes in a route generation method based on lane - level map collection at intersections of the present invention;

[0022] Figure 6 It is a display diagram of the driving route generated in a route generation method based on lane - level map collection at intersections of the present invention. Detailed implementation manners

[0023] The following is a further detailed description through specific implementation manners:

[0024] Embodiment 1

[0025] A route generation method based on lane - level map collection at intersections in this embodiment is used to generate an optimal driving path for all lanes at an intersection, ensuring that the vehicle collects map data of all lanes within the shortest distance to form lane - level map data, achieving less repetition of driving lanes, and making U - turns, turning left and right at the best positions, etc., so as to reduce the collection working hours and energy consumption and improve the data collection efficiency. In this embodiment, as shown in the appendix Figure 1 The vehicle drives through all lanes at the intersection according to the generated driving route and records the effective GPS lane data to form lane - level map data, including:

[0026] S1. Obtain the length of each lane at the current intersection and calculate the connection relationship between each lane.

[0027] In this embodiment, for the convenience of description, the intersection information to be collected is sorted into a display diagram as shown in the appendix Figure 2 The roads that need to be collected are defined as inspection roads, and the roads that do not need to be collected are non - inspection roads. Each intersection of the inspection roads is marked with a number. At the same time, to simplify the description of intersection data, in this embodiment, a single cross - road will be described, specifically taking the basic cross - road shown in the appendix Figure 3 As shown. Generally, each single intersection includes 4 road segments, and the end nodes of each road segment are represented by 1, 2, 3, and 4 respectively, and 0 represents the intersection center node of the current inspection road. Each road segment includes 6 lanes in both directions, so a single cross - road consists of 24 lanes. The turning directions of each lane are as shown, including all possible driving directions such as going straight, turning left, turning right, and making U - turns. Combining with the appendix Figure 4As shown, section 1-0 means traveling from end node 1 to center node 0.

[0028] Regarding each lane at the current intersection as a node and the feasible connections between lanes as edges, the edge weights are calculated, where the edge weight = actual distance, thereby calculating the length of each lane. Meanwhile, the connection relationships between lanes are calculated. In this embodiment, the connection relationship refers to the feasibility of a vehicle traveling from one lane to another, that is, according to traffic rules and intersection geometric constraints, legal connection relationships between lanes are defined, such as not being able to turn right immediately after turning left, prohibited lane changes, etc. Among them, when there are multiple feasible lanes for a path, the vehicle traveling order is the U-turn lane > straight lane > right-turn lane. At the same time, when encountering an intersection of non-inspected roads, a U-turn is made, thus forming constraint conditions to filter out invalid path combinations. By dynamically calculating edge weights such as lane length and turning restrictions, the driving cost is accurately quantified and the repetition rate is reduced.

[0029] S2. Calculate the shortest distance between each lane and construct a weighted complete directed graph of the necessary lanes.

[0030] In this embodiment, each lane in the intersection is defined as a necessary lane. The Floyd algorithm is used to calculate the shortest distance between each lane and record the shortest distance between any two lanes. Taking the necessary lanes as nodes, a complete directed graph is constructed, and the weight of the edge is the shortest distance or travel time between two lanes, thereby transforming the LCPP problem into a weighted complete directed graph that only contains necessary lanes. As shown in the appendix Figure 5 Since the illustration of 24 lanes is too large, only 4 necessary lanes are taken as an example for display here. Thus, by finding the distances between necessary lanes, the entire map is simplified to a map that only contains necessary lanes, greatly simplifying the road information, minimizing the computational amount of the system, and thereby improving the calculation efficiency.

[0031] S3. Use dynamic rules to calculate the shortest path for accessing all necessary lanes and form the driving order of the necessary lanes.

[0032] Through the calculated shortest path of the necessary lanes, an improved algorithm based on dynamic programming is used to find the shortest path passing through all paths of the weighted complete directed graph, that is, to find the order that visits all necessary lanes once and has the minimum total weight in the directed graph. In this embodiment, the improved algorithm formula of the dynamic rule is expressed as

[0033]

[0034] LCPP * = minLCPP({l1, l2, l3,..., l i-1 , l i , l i ) + ll i ;

[0035] Wherein, LCPP * represents the shortest distance passing through all lanes; L represents the set of all lanes to be passed; l i represents the last lane; l j represents the previous last lane; d ji represents the distance from lane j to lane i; ll i represents the length of lane i.

[0036] The dynamic programming algorithm adopted is an improved algorithm based on the state compression method. The shortest distance passing through all roads in the LCPP problem can be obtained through the above formula. During the calculation process, the whole process of passing through the roads is divided into multiple small processes. By ensuring that each process is an optimal solution, the whole path must be an optimal path. At the same time, there will be repeated calculations of multiple paths during the calculation process. Therefore, in this embodiment, the intermediate values of the calculations are saved using space to greatly improve the efficiency, shorten the time, and reduce the output of repeated data, thereby obtaining the driving order of the necessary lanes.

[0037] S4. Fill in the transition lanes, generate the complete optimal path order, and output the driving route.

[0038] In this embodiment, through the above steps, the driving order of the necessary lanes is obtained. It is also necessary to reverse-derive the non-necessary lanes passed through, sort and screen the non-necessary lanes, and finally select the transition lane with the shortest distance for filling to form a complete driving route. In this embodiment, the obtained driving order of the necessary lanes is divided into multiple sections (for example, dividing 1-0-3 into 1-0 and 0-3). The Dijkstra algorithm is used to calculate the shortest path of each section of the road, obtain the sequence of transition lanes included in the shortest path, obtain the shortest transition lane and fill it into the section to form a complete passing sequence.

[0039] In this embodiment, it also includes verifying whether the path order meets all constraint rules, drawing the navigation map and then outputting the final solution. The result of its driving route is as shown in the appendix Figure 6 and is described in the form of "starting node - ending node _ driving lane" for the route. Combining with the appendix Figure 4 shown, then "1-0_1" means driving from node 1 to node 0 and passing through lane 1. Then the shortest navigation route in the appendix Figure 6 is represented as

[0040] "1-0_1" (starting point)

[0041] "0-1_1"

[0042] "1-0_2"

[0043] "0-3_1"

[0044] "3-0_1"

[0045] "0-3_2"

[0046] "3-33_1"

[0047] "33-3_2"

[0048] "3-0_2"

[0049] "0-1_2"

[0050] "1-11_1"

[0051] "11-1_2"

[0052] "1-0_3"

[0053] "0-2_1"

[0054] "2-0_1"

[0055] "0-2_2"

[0056] "2-22_1"

[0057] "22-2_2"

[0058] "2-0_2"

[0059] "0-4_1"

[0060] "4-0_1"

[0061] "0-4_2"

[0062] "4-44_1"

[0063] "44-4_2"

[0064] "4-0_2"

[0065] "0-2_3"

[0066] "2-21_1"

[0067] "21-2_1"

[0068] "2-0_3"

[0069] "0-3_3"

[0070] "3-23_1"

[0071] "23-3_1"

[0072] "3-0_3"

[0073] "0-4_3"

[0074] "4-43_1"

[0075] "43-4_1"

[0076] "4-0_3"

[0077] "0-1_3" (end point)

[0078] It can be seen from the figure that this driving route is the optimal route, which can cover all lanes and reduce the lane driving repetition rate, thereby improving collection efficiency while reducing collection time and energy consumption.

[0079] During the actual data collection process, a GPS positioning system is installed in the middle of the vehicle's roof, so that the vehicle is driven in the center of the lane, and the GPS data of the lane center is recorded as lane map data.

[0080] This embodiment utilizes innovative lane-level directed graph modeling and a dynamic programming optimization algorithm to achieve globally optimal planning for multi-lane data collection routes at intersections. This innovative approach breaks through the limitations of traditional navigation algorithms that rely on static road networks. It pioneers a three-layer topological model: "lane-connectivity-dynamic weights." This model accurately quantifies the turning constraints and driving costs for each lane. Dynamic planning generates a sequence of required lanes and intelligently fills in transition lanes, addressing existing issues such as duplicate collection, illegal U-turns, and local optimality traps caused by coarse-grained path planning. Compared to traditional approaches, this solution demonstrates significant engineering success: by weighting lane lengths and modeling turn costs, it reduces total route mileage by over 30%. Furthermore, it incorporates U-turn legality verification and lane change minimization mechanisms, significantly reducing safety risks while improving collection efficiency. It is also applicable to complex scenarios such as irregular intersections. This approach overcomes the long-standing technical bottleneck in high-precision map collection, which hinders the coordinated optimization of accuracy, efficiency, and safety. This represents a breakthrough in lane-level map collection and navigation technology.

[0081] The above is only an embodiment of the present invention, and the common knowledge such as the specific technical solutions and / or characteristics in the solution are not described in detail here. It should be pointed out that for those skilled in the art, without departing from the technical solution of the present invention, several variations and improvements can be made, which should also be regarded as the scope of protection of the present invention, and these will not affect the effect of the implementation of the present invention and the practicality of the patent. The scope of protection required by this application shall be based on the content of its claims, and the specific implementation methods and other records in the description can be used to interpret the content of the claims.

Claims

1. A route generation method based on crossroads lane-level map acquisition, characterized in that Drive along all lanes of the intersection according to the generated driving route, and record the effective GPS lane data, including: Step 1: Obtain the length of each lane at the current intersection and calculate the connection relationship between lanes; Step 2: Calculate the shortest distance between each lane and construct a weighted complete directed graph of the necessary lanes; Step 3: Use the dynamic programming method to calculate the shortest path to visit all necessary lanes and form the driving order of the necessary lanes; Step 4: Fill in the transition lanes, generate the complete optimal path order, and output the driving route.

2. The route generation method based on crossroads lane-level map acquisition according to claim 1, characterized in that: The lanes include straight, left turn, right turn and U-turn.

3. A route generation method based on crossroads lane-level map acquisition according to claim 1, characterized in that: The connection relationship refers to the feasibility of a vehicle driving from one lane to another.

4. A route generation method based on crossroads lane-level map acquisition according to claim 1, characterized in that: In Step 3, the improved algorithm formula of the dynamic programming method is expressed as LCPP * = minLCPP({l1, l2, l3,..., l i-1 , l i}, l i ) + ll i ; wherein, LCPP * represents the shortest distance passing through all lanes; Let \(L\) denote the set of all lanes to be traversed; \(l\) i denotes the last lane; \(l\) j denotes the previous last lane; \(d\) ji denotes the distance from lane \(j\) to lane \(i\); \(ll\) i denotes the length of lane \(i\).

5. A route generation method based on intersection lane-level map acquisition according to claim 4, characterized in that: In Step 1, each lane at the current intersection is used as a node, and the feasible connection between lanes is used as an edge, then the edge weight = actual distance.

6. The route generation method based on crossroads lane-level map acquisition according to claim 1, wherein: In Step 4, it includes dividing the driving order of the necessary lanes obtained into multiple sections, using the Dijkstra algorithm to calculate the shortest path of each section, obtaining the sequence of transition lanes included in the shortest path, and obtaining the shortest transition lane and filling it into the section.

7. A route generation method based on intersection lane-level map acquisition according to claim 3, characterized in that: When there are multiple feasible lanes for a path, the vehicle driving order is U-turn lane > straight lane > right turn lane, and when encountering an intersection of non-inspection roads, make a U-turn.

8. A route generation method based on crossroads lane-level map acquisition according to claim 1, characterized in that: Use the Floyd algorithm to calculate the shortest distance between each lane and record the shortest distance between any two lanes.

9. A route generation method based on crossroads lane-level map acquisition according to claim 1, characterized in that: A single intersection includes 4 sections, and each section includes 6 lanes in both directions.

10. A route generation method based on intersection lane-level map acquisition according to claim 1, characterized in that: It also includes installing a GPS positioning system in the middle of the vehicle roof to make the vehicle drive in the center of the lane and record the GPS data of the lane center as lane map data.

Images