Construction method of unmanned vehicle semantic map model and its application method in unmanned vehicle

Abstract

The invention discloses an unmanned vehicle semantic map model building method and an application method thereof to an unmanned vehicle. Extraction of a conceptual structure indicates that key map elements such as road networks, road traffic participants and traffic rules related in the running process of the unmanned vehicle are reasonably abstracted into different conceptual types, establishment of the semantic relation between concepts refers to establishment of map concept semantic hierarchical relations and incidence relations, and living examples of the conceptual types and the semantic relation among the living examples are established in an instantiated manner to finally obtain a semantic map for the unmanned vehicle. A map data structure applicable to the unmanned vehicle is built, the sufficient semantic relation among the map elements is designed, the semantic map is generated, semantic reasoning is performed according to the semantic map, a globally planned route, the current position and orientation of the unmanned vehicle and peripheral real-time obstacle information to obtain local scene information of the unmanned vehicle, scene understanding of the unmanned vehicle is realized, and the unmanned vehicle is assisted in behavior decision.

Description

technical field [0001] The present invention mainly relates to the technical field of unmanned vehicles, in particular to a method for constructing a semantic map model of an unmanned vehicle and its application method on an unmanned vehicle. Background technique [0002] In recent years, unmanned vehicles have received extensive attention from domestic and foreign academic and industrial circles, and their related supporting technologies have developed rapidly. From the perspective of system composition and information flow, the unmanned vehicle system can generally be divided into modules such as environmental perception, decision planning, and motion control. The environmental perception uses various sensors to obtain real-time scene information of the traffic environment and generate an environmental model (ie perception map); on this basis, on the basis of the decision-making and planning environment model, make behavioral decisions that comply with traffic rules and sa...

AI Technical Summary

Problems solved by technology

[0004] The Chinese patent with the publication number CN104535070A (application number 20141083873.5), which provides a high-precision map data structure, collection and processing system and method, divides the map data structure into four layers: road network, lane network, lane line Information and special information data, although database-level associations are defined between several levels, due to the lack of semantic information, it is difficult for unmanned vehicles to establish a complete semantic relationship between various map elements and traffic participants in this map data structure , distinguish real-time scene information of unmanned vehicles, and realize scene understanding

At the same time, information such as intersections and U-turns is difficult to reflect in its data structure, and the relationship between lane lines and lanes is not accurate enough. For example, a certain section of road may change from two lanes to three lanes. In this case, the lane in the middle of the three lanes and the lane It is difficult to express the relationship between the lines

[0005] The Chinese patent with publication number CN104089619A (application number 201410202876.4) provides a GPS navigation map precise matching system for unmanned vehicles and its operation method. By obtaining road information, determining the starting point, obtaining vehicle positioning information, information The process of matching and screening completes the precise matching of the navigation map, but the matching method is mainly to search through discrete points, without utilizing the correlation between map elements, which will lead to the problem of low matching efficiency

Images