OSM-based map slam real-time positioning optimization system

Abstract

An OSM-based SLAM real-time positioning optimization system, comprising: a prior OSM map preprocessing module, a SLAM positioning data processing module, a data alignment module and a SLAM positioning map optimization module.The present application extracts the map environment information of the Open Street Map type map, extracts the topological information and semantic information of the prior map, and aligns the positioning information of the SLAM in real time, so that the SLAM can adjust the positioning based on the OSM map information, reduce the drift and improve the positioning accuracy.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of intelligent driving, and particularly relates to an SLAM (Simultaneous Localization and Mapping) real-time positioning optimization system based on an Open Street Map (OSM) type map. BACKGROUND

[0002] In the case of long-time work, the running result of the SLAM (Simultaneous Localization and Mapping) technology inevitably has an increasingly large accumulated error; and in the case of different working conditions represented by underground, field and construction site, the positioning effect of the SLAM technology is poor in robustness, and the work is not stable enough. The existing road network map is a mainstream map used for vehicle driving and navigation, and a type of road network map such as the Open Street Map (OSM) for environment description is relatively reliable in accuracy, rich in semantic information, and very suitable for serving as a benchmark for SLAM positioning optimization. Since the SLAM positioning data and the OSM type map data are not of the same source, the alignment of heterogeneous data needs to be solved for the SLAM positioning optimization based on the OSM map, so that the technology for optimizing the SLAM positioning effect based on the OSM map still has a technical gap, but an SLAM positioning optimization algorithm based on such an environment description map is a cornerstone for further development of the SLAM positioning technology. SUMMARY

[0003] The application aims at the problems that the existing SLAM technology inevitably has an increasingly large accumulated error in the case of long-time work, and the positioning effect of the SLAM technology is poor in robustness and the work is not stable enough in the case of different working conditions represented by underground, field and construction site, and proposes an SLAM real-time positioning optimization system based on an OSM type map. The map environment information of the Open Street Map type map is extracted, the topological information and semantic information of the prior map are extracted, and the positioning information of the SLAM is aligned in real time, so that the SLAM can adjust the positioning based on the OSM map information, reduce the drift, and improve the positioning accuracy.

[0004] The application is implemented by the following technical scheme:

[0005] The application relates to an OSM-based SLAM real-time positioning optimization system, comprising a prior OSM map preprocessing module, a SLAM positioning data processing module, a data alignment module and a SLAM positioning map optimization module, wherein: the prior OSM map preprocessing module extracts a prior topological map with semantic nodes meaningful for driving from original map data in an offline stage; the SLAM positioning data processing module generates a SLAM topological positioning information map by key points after sparse processing and key point extraction through SLAM algorithm results; the data alignment module performs alignment processing according to the prior topological map and the SLAM topological positioning information map, and filters out topological node pairs meeting map optimization conditions; and the map optimization module constructs a map optimization system based on the aligned map node data, is used for optimizing and adjusting original SLAM positioning data, obtains a SLAM topological positioning information map with reduced SLAM positioning drift, and obtains updated SLAM positioning data after inverse processing of the SLAM positioning data processing module, so that optimized positioning is realized.

[0006] The prior topological map G of the semantic node map ={V map , E map} wherein V map and E map represent nodes and edges in the topological map respectively.

[0007] The SLAM topological positioning information map G SLAM (T) = {V SLAM (T), E SLAM (T)} wherein V SLAM (T) and E SLAM (T) represent nodes and edges in the SLAM topological positioning information map at the current T moment respectively.

[0008] The original map data refers to a prior OSM road network map, that is, map data in the OpenStreetMap format, comprising latitude and longitude coordinates of a point in the map and semantic information Nodes, marked as Road or contour information Ways composed of the Nodes, marked as

[0009] and Relations for storing possible association information between various information.

[0010] The topological node pair (V match map (T), V match sLAM (T)) wherein V matchmap (T) and V match SLAM (T) two node sets have the same number of elements and one-to-one matching relationship.

[0011] The data alignment module and the SLAM positioning data processing module operate at the same frequency.

[0012] The extraction refers to that the prior OSM map processing module analyzes the longitude and latitude coordinates and semantic information of the map points in the original map data and road or contour (such as a fence) information According to TAGS N seg and TAGS W seg Extract the road on which the vehicle can travel After that, extract the as V map , retain the semantic information and establish the topological relationship E between each other according to the associated information Relations map , form a topological semantic map G map ={V map , E map}.

[0013] The sparse processing refers to that the density of the obtained SLAM positioning trajectory is related to the frequency of the SLAM sensor, and generally does not require such dense data during map post-processing, especially in places with little information (such as straight lines). Therefore, fewer key points are extracted in places without curvature changes in the obtained SLAM positioning data to achieve the effect of sparse processing and reduce the amount of data.

[0014] The key point extraction refers to that based on the SLAM positioning information, a fixed time length D is selected as a sliding window, the orientation of the first positioning point X0 in D time is taken as the positive direction, the positioning points in D time are fitted as a curve, the extreme points of the curve are marked as key points V SLAM (T), that is, the inflection points, and the data is sparsely selected according to the curvature size to reduce the computational burden of later optimization.

[0015] The alignment processing includes topological information alignment and semantic information alignment.

[0016] The topological information alignment refers to that the degrees of V map and V SLAM (T) are compared, and the structure of a certain level of subtree expanded from itself as the root node is compared, and the difference in degrees and the similarity of the subtree structure (detected using the WL-Test method) are used as the first layer judgment of alignment;

[0017] Semantic information alignment means that in G map With semantic information, match V SLAM (T) to determine whether it is possible to have this semantic information, such as V map More likely to match V SLAM (T) with greater curvature. In addition, the alignment node before T time already exists, so this module focuses on processing new D SLAM (T) at T time, and then combines with the previous matching point and outputs.

[0018] The map node data refers to the topological node pair (V match map (T), V match SLAM (T))).

[0019] The original SLAM positioning data refers to the initially constructed SLAM topological information map G SLAM (T) = {V SLAM (T), ES LAM (T)}.

[0020] The optimization adjustment refers to the graph optimization system G match (T) = {V match (T), E match (T)} in which V match (T) = V SLAM (T) ∪ V match map (T), and keep the node attributes in V match map (T) set fixed as one of the optimization references; E match (T) = Relation (V match map (T), V match SLAM (T)) ∪ E SLAM (T), the topological node pair output by the data alignment module and the edge in the SLAM topological positioning information graph are again input to the graph optimization module for iterative optimization until the maximum number of iterations I max or the optimization error is less than the error threshold Error th , the optimized SLAM topological positioning information graph is obtained.

[0021] The optimization error refers to the value of the loss function with E match (T) as the optimization objective. If there is no error, it means that there is a set of solutions that can perfectly satisfy E matchAll the relations of (T) but generally cannot be achieved. The optimization error is the value of the loss function at the current iteration.

[0022] The inverse processing refers to restoring the optimized SLAM topological information map into the original SLAM positioning trajectory format.

[0023] Technical effects

[0024] The present application extracts the prior map, that is, the map environment information of the Open Street Map type map, extracts the topological information and semantic information of the prior map, and aligns the positioning information of the SLAM in real time. The structure of the graph is used to match the OSM type road network map and the SLAM positioning trajectory two heterogeneous map data, and the graph optimization algorithm is combined to optimize the SLAM positioning trajectory. Compared with the prior art, the present application can adjust the positioning based on the OSM map information, reduce the drift, and improve the positioning accuracy. Through the optimization algorithm, the problem of invalid SLAM positioning or poor robustness in some special scenarios can be well solved. BRIEF DESCRIPTION OF DRAWINGS

[0025] Figure 1 It is a schematic diagram of the system of the present application;

[0026] Figure 2 It is a schematic diagram of the effect of the embodiment;

[0027] In the figure: (a) is the visualization of the prior topological map; (b) is the visualization of the SLAM trajectory graph, wherein the yellow thickened nodes are key nodes; (c) is the visualization result after alignment. DETAILED DESCRIPTION

[0028] As shown in Figure 1 , it is a SLAM real-time positioning optimization system based on an OSM type map, comprising: a prior OSM map preprocessing module, a SLAM positioning data processing module, a data alignment module and a SLAM positioning graph optimization module.

[0029] The input of the prior OSM type map preprocessing module is the original map data D map , and the output is a prior topological map G map ={V map , E map} with semantic nodes meaningful for driving, wherein V map and E map represent the nodes and edges in the topological map, respectively. The prior OSM type map preprocessing module is used to extract the topological information and semantic information of the prior map, and align the positioning information of the SLAM in real time. Figure 1The data is typically in XML format and supports various data processing methods. The main content of the raw data includes three categories of map information: Nodes, Ways, and Relations. Nodes are the most basic data, recording the latitude and longitude coordinates and semantic information of a point on the map, and are tagged as... Ways contain information about roads or outlines (such as walls) composed of Nodes, tagged as Relations stores the possible relationships between various pieces of information. The prior OSM map processing module will primarily analyze... and Category information, based on TAGS N seg and TAGS W seg Extracting the road where the vehicle can drive Then on the road Extracted, as V map Preserve semantic information and establish based on Relations The topological relationship between them, as E map To form a topological semantic map G map ={V map E map}

[0030] The SLAM positioning data processing module runs synchronously and in real time with the SLAM positioning algorithm. The input of this module is the result D of the SLAM algorithm. SLAM (T), the output is the SLAM topology localization information map G after processing the SLAM localization data at the current time T (T∈{0,1,2......}). SLAM (T)={V SLAM (T), E SLAM (T)}, where V SLAM (T) and E SLAM (T) represent the nodes and edges in the SLAM topology localization information graph at time T, respectively. This module performs D... SLAM (T) Perform sparsity processing and key point extraction, and use the key points as V SLAM (T) Constructing a topology map. One of the implementation logics for key point extraction is based on SLAM localization information. A fixed time length D is selected as the sliding window. The orientation of the first localization point X0 within time D is taken as the positive direction. The localization points within time D are fitted as a curve, and the extreme points of the curve are marked as key points V. SLAM (T), representing the true inflection point, is used, and data is sparsely selected based on the curvature magnitude to reduce the computational burden of later optimization. Additionally, in D... SLAM(T) Extract semantic descriptors from it as global alignment operators, such as intersections, buildings, etc.

[0031] The input to the data alignment module is the output of the first two modules {G}. map G SLAM (T)}, this module will align the input data and filter out the topological node pairs (V) that satisfy the graph optimization conditions. match map (T), V match SLAM (T)), V match map (T) and V match SLAM (T) Two node sets have the same number of elements and a one-to-one correspondence. The alignment implementation logic mainly includes two parts: topological information and semantic information. Topological information alignment involves comparing the nodes to be matched (V). map and V SLAM The degree of (T) and the structure of subtrees of a certain level extending from itself as the root node are used as the first level of alignment judgment based on the difference in degree and the similarity of subtree structures (detected using the WL-Test method); semantic information alignment is then performed on G. map When semantic information is included, the treatment of matching V SLAM (T) Determine whether the semantic information is likely to be present, such as V with intersection semantic labels. map More likely to match a V with a larger curvature SLAM (T). Furthermore, since the alignment nodes prior to time T already exist, this module at time T focuses on processing the newly added D. SLAM (T), then combined with the previously matched point set and output. This module is a real-time module, operating in sync with the SLAM positioning data processing module.

[0032] The SLAM localization map optimization module constructs a graph optimization system G based on the graph node data to be aligned. match (T)={V match (T), E match (T)} optimizes and adjusts the original SLAM positioning data. Where V match (T)=V SLAM (T)∪V match map (T), and maintain The node attributes in the set are fixed, serving as one of the benchmarks for optimization; E match (T)=Relation(V match map (T), V match SLAM (T))∪E SLAM(T), the output of the above alignment module and the edges of the SLAM topology graph as the edges of the graph optimization participate in the optimization process. The optimization process will iterate until the maximum number of iterations I max or the optimization error is less than the error threshold Errorth. Finally, a new set of SLAM topology positioning information graph G' SLAM (T) = {V' SLAM (T), E' SLAM (T)} is output. Then, through the inverse process of the logic in the SLAM positioning data processing module, the topology positioning information graph is converted into new SLAM positioning data, which is added to the subsequent SLAM positioning algorithm.

[0033] The SLAM positioning data processing module, the data alignment module, and the SLAM positioning graph optimization module are real-time running modules and are combined with the SLAM positioning algorithm to form the real-time technical part of the application. The running frequency of the real-time part is consistent or in a multiple relationship.

[0034] The embodiment relates to a global map alignment method for the SLAM technology of the above system, which comprises the following steps:

[0035] Step 1: processing the prior OSM type map and the SLAM real-time positioning data to obtain the corresponding topology graph;

[0036] Step 2: performing node matching on the topology graph obtained in step 1;

[0037] Step 3: in the process of real-time running of the SLAM algorithm, the node information in the SLAM topology graph is synchronously and real-timely optimized through the graph optimization, the positioning information of the SLAM is updated and then returned to the SLAM algorithm, and the next step of real-time positioning is performed.

[0038] Compared with the prior art, the method can utilize the prior information of the OSM type road network map to assist the SLAM positioning algorithm in stable positioning in various working conditions.

[0039] The above specific embodiments can be adjusted in different ways by those skilled in the art without departing from the principles and purposes of the application. The protection scope of the application is subject to the claims and is not limited by the above specific embodiments. Each implementation scheme within the scope is subject to the constraints of the application.

Claims

1. A real-time SLAM localization optimization system based on OSM-like maps, characterized in that, The system includes: a priori OSM map preprocessing module, a SLAM positioning data processing module, a data alignment module, and a SLAM positioning map optimization module. Specifically: the priori OSM map preprocessing module extracts a priori topological map with semantically meaningful nodes for driving from the original map data during the offline phase; the SLAM positioning data processing module performs sparsity processing and key point extraction on the SLAM algorithm results to generate a SLAM topological positioning information map based on the key points; the data alignment module performs alignment processing based on the priori topological map and the SLAM topological positioning information map, and filters out topological node pairs that meet the graph optimization conditions; the graph optimization module constructs a graph optimization system based on the aligned map node data to optimize and adjust the original SLAM positioning data, obtaining a SLAM topological positioning information map that reduces SLAM positioning drift. This map is then inversely processed by the SLAM positioning data processing module to obtain updated SLAM positioning data, achieving optimized positioning. The prior topology map of the semantic nodes middle, and These represent nodes and edges in the topological map, respectively. The SLAM topology positioning information map middle, and Representing the current Nodes and edges in the time-matter SLAM topology localization information graph; The original map data refers to: a priori OSM-type road network map, i.e., map data similar to OpenStreetMap format, including: latitude and longitude coordinates of a point on the map and semantic information Nodes, labeled as... ;Ways, which contain road or outline information composed of Nodes, are tagged as And the relationships between the various pieces of information; The topology node pair middle, and The two node sets have the same number of elements and are matched one-to-one; The map node data mentioned above refers to: topology node pairs ; The original SLAM localization data refers to the initially constructed SLAM topology map. ; The data alignment module operates in sync with the SLAM positioning data processing module. The alignment process includes: topological information alignment and semantic information alignment; Topological information alignment refers to: comparing the parts to be matched and The degree of the subtree and the structure of the subtrees at a certain level with itself as the root node are used as the first level of alignment judgment based on the difference in degree and the similarity of the subtree structure. Semantic information alignment refers to: in When semantic information is included, the matching process To determine whether the semantic information is likely to be present, use intersection semantic labels. Able to match larger curvature ,in addition, The alignment node before the current time already exists, so This module focuses on handling newly added features. Then combine it with the previous set of matching points and output it.

2. The SLAM real-time localization optimization system based on OSM-like maps according to claim 1, characterized in that, The extraction mentioned above refers to the prior OSM map processing module analyzing the latitude and longitude coordinates and semantic information of points in the original map data. and road or outline information ,according to and Extracting the road where the vehicle can drive Afterwards, extract the road As Preserve semantic information and establish relationships based on relational information (Relations). Topological relationships between each other To form a topological semantic map .

3. The SLAM real-time positioning optimization system based on OSM-like maps according to claim 1, characterized in that, The key point extraction refers to: based on SLAM positioning information, selecting a fixed time length. As a sliding window, The first positioning point within the time period The orientation is positive, and the fitting is... The location points within a time period are represented by a curve, and the extreme points of the curve are marked as key points. The inflection point is determined by selecting data sparsely based on the curvature magnitude to reduce the computational burden of later optimization.

4. The SLAM real-time localization optimization system based on OSM-like maps according to claim 1, characterized in that, The aforementioned optimization and adjustment refers to: graph optimization system middle, and maintain The node attributes in the set are fixed, serving as one of the benchmarks for optimization; The topology node pairs output by the data alignment module and the edges in the SLAM topology localization information graph are then used as input to the graph optimization module for iterative optimization until the maximum number of iterations is reached. Or the optimization error is less than the error threshold Up to this point, the optimized SLAM topology localization information map is obtained.

5. The SLAM real-time localization optimization system based on OSM-like maps according to claim 1, characterized in that, The inverse processing refers to restoring the optimized SLAM topology information map to the original SLAM positioning trajectory format.

6. A global map alignment method for SLAM technology in any of the systems described in claims 1-5, characterized in that, include: Step 1: Process prior OSM-type maps and SLAM real-time positioning data to obtain the corresponding topology map; Step 2: Perform node matching on the topology graph obtained in Step 1; Step 3: During the real-time operation of the SLAM algorithm, the node information in the SLAM topology graph is simultaneously optimized in real time through graph optimization, the SLAM positioning information is updated and then sent back to the SLAM algorithm for the next step of real-time positioning.

Images