KR-20260063805-A - Apparatus for generating autonomous vehicle precision map including infrastructure information and method using the same

KR20260063805AKR 20260063805 AKR20260063805 AKR 20260063805AKR-20260063805-A

Abstract

The present invention relates to a precision map generation device for an autonomous vehicle including infrastructure information and a method using the same, characterized by comprising: a sensing information collection unit installed along a roadside to collect sensing data from infrastructure; a fixed information recognition unit that recognizes fixed information from the sensed data; a dynamic information recognition unit that recognizes dynamic information from the sensed data; a precision map fusion unit that reflects road conditions recognized from the recognized fixed information and dynamic information into a precision map; and a map server transmission unit that transmits the precision map reflecting the road conditions to a map server.

Inventors

오광만
추연호
안웅

Assignees

(주)테슬라시스템

Dates

Publication Date: 20260507
Application Date: 20241031

Claims (7)

A sensing information collection unit installed along the roadside that collects sensing data from infrastructure; A fixed information recognition unit that recognizes fixed information from the above-mentioned sensed data; A dynamic information recognition unit that recognizes dynamic information from the above-mentioned sensed data; A precision map fusion unit that reflects road conditions recognized from the above-mentioned recognized fixed information and dynamic information into a precision map; and It is composed of a map server transmission unit that transmits a precision map reflecting the above road conditions to a map server, A precision map generating device for an autonomous vehicle characterized by the above road conditions being at least one of road construction, traffic accidents, jaywalking pedestrians, or vehicles driving against the flow of traffic.
In paragraph 1, The above precision map fusion unit includes a risk level determination unit that recognizes the fixed information and the dynamic information to determine the risk level; A global coordinate conversion unit that converts the location of the above-recognized road conditions into global coordinates and It is composed of a precision map reflection unit that reflects the precision map according to the above risk level, A precision map generation device for autonomous vehicles characterized by the above risk levels being normal, caution, and warning.
In paragraph 1, A precision map generation device for autonomous vehicles characterized by further including an unstructured information recognition unit that recognizes frequently changing unstructured information such as traffic light colors, traffic facilities, variable lanes, and speed limits.
In paragraph 1, A precision map generation device for an autonomous vehicle, characterized by further including a V2X communication unit that transmits a precision map reflecting the above road conditions to an autonomous vehicle located within the coverage range of the above infrastructure using V2X communication.
In paragraph 1, The above fixed information is at least one of a lane, crosswalk, traffic light, and traffic sign that is fixedly located in a certain place, and A precision map generating device for an autonomous vehicle, characterized in that the above dynamic information is at least one of a vehicle, motorcycle, bicycle, kickboard, or pedestrian moving on a road.
In a method of using a precision map generation device for autonomous vehicles, A first step of recognizing fixed information and dynamic information located on surrounding roads from sensors equipped in the infrastructure; A second step of determining the risk level from the above-mentioned recognized information; A method for generating a precision map for an autonomous vehicle, characterized by comprising a third step of generating recognized information into a packet and transmitting it to a map server to reflect it in a precision map for an autonomous vehicle when the risk level is caution or warning among the above risk levels.
In paragraph 6, A method for generating a precision map for an autonomous vehicle, characterized by using the recognized information in the third step to generate a precision map for an autonomous vehicle independently within the infrastructure and transmitting the generated precision map to at least one of a map server or an autonomous vehicle.

Description

Apparatus for generating autonomous vehicle precision map including infrastructure information and method using the same The present invention relates to a precision map generation device for an autonomous vehicle comprising infrastructure information that reflects road conditions recognized from infrastructure installed along a roadside into a precision map for an autonomous vehicle for use in autonomous driving, and a method using the same. Positioning, which accurately estimates a vehicle's current location, is one of the essential core technologies for implementing autonomous driving or Advanced Driver Assistance Systems (ADAS). To this end, sensor fusion positioning technology is being widely researched to precisely correct the vehicle's current position by fusing sensors such as GPS, inertial measurement units (IMUs), cameras, or LiDAR with high-precision maps and comparing the distance of an object measured by the sensors with location information from a pre-stored map. Typically, the most realistic approach being considered is location correction technology utilizing image recognition sensors such as cameras, which necessarily requires high-precision maps containing various road information, attributes of road facilities, and absolute coordinate data. However, the road facilities that cameras and LiDAR sensors must recognize to extract vehicle location information include traffic lights, lanes, and curbs that separate roads from sidewalks; if these physical sensors misidentify them, there is a very high possibility of causing serious traffic accidents. Furthermore, sensors equipped in vehicles are affected by environmental factors such as fog or rainfall; in particular, there is a growing need to recognize road conditions that cannot be perceived from a distance or while a vehicle is in motion through infrastructure and reflect this in precision maps for autonomous vehicles. Generally, a precision map for autonomous driving refers to a 3D digital map constructed with an accuracy of 25 cm, including lane information, regulatory and safety information, and various road and signage facilities. Once such a precision map is constructed, benefits such as reduced data processing errors, improved AI learning capabilities, a decrease in the volume of data that needs to be analyzed in real time, and enhanced eco-friendliness can be achieved. While existing navigation maps include node and link-based network information regarding road centers as primary information, precision maps can be composed of various information with higher accuracy to support the driving, control, and location perception of autonomous vehicles. However, while precision maps can be defined in various ways depending on their type and use and standardization is already underway, there is a problem in that the precision maps currently defined in ISO standards lack information recognized by edge infrastructure systems (hereinafter referred to as 'infrastructure') in their map components. As such, infrastructure equipped with various sensors installed along the roadside is essential to recognize the autonomous driving environment in advance. Infrastructure is gradually shifting toward a driving method that recognizes dynamic elements around the road, information on unexpected situations, road surface conditions, traffic accidents, and road construction, transmits this information to nearby autonomous vehicles via C-V2X communication, and enables autonomous cooperative driving based on this prior information. As such, cooperative autonomous driving is a technology that utilizes not only information recognized by the autonomous vehicle's own sensors but also road environment information received from infrastructure installed around the road or traffic centers for driving. In addition, providing autonomous vehicles with information on various unexpected situations occurring in their blind spots (traffic accidents, road construction, vehicle breakdowns, jaywalking, etc.) enables safer and accident-free operation. As such, a device that generates real-time precision maps for autonomous vehicles to effectively utilize information from infrastructure has not yet been developed, but it is an essential technology. Patent Document 1 relates to a method and apparatus for generating an autonomous driving path including infrastructure information, characterized by obtaining information regarding a target section and a high-definition road map, obtaining node information, link information, and first infrastructure information from a part corresponding to the target section in the high-definition road map, generating a first autonomous driving path based on the node information, link information, and first infrastructure information, obtaining second infrastructure information from at least one sensor or traffic information intermediary server included in the autonomous driving device, generating a second autonomous driving path based on the first autonomo