KR20210029323A - Apparatus and method for improving cognitive performance of sensor fusion using precise map - Google Patents

Abstract

Disclosed are an apparatus and a method for improving the cognitive performance of a sensor fusion using a precise map. According to an aspect of the present invention, the apparatus for improving the cognitive performance of the sensor fusion using the precise map comprises: a position search unit which acquires the position data of a vehicle; a camera unit which takes a surrounding video of the vehicle; a surrounding situation collection unit which collects the surrounding situation data of the vehicle; a storage unit which stores the precise map including the navigation information and a geographic feature on the driven road; and a control unit which map-matches the position data of the vehicle and the camera lane data received from the camera unit with the precise map, estimates the precise position of the vehicle, pre-processes the surrounding situation data received from the surrounding situation collection unit, outputs integrated target information, maps the integrated target information on the precise map with special emphasis on the precise position of the vehicle, and estimates the position based on the lane of the precise map. The present invention aims to provide an apparatus and a method for improving the cognitive performance of a sensor fusion using a precise map, which is able to delete an incorrectly recognized target and to trace an un-recognized target.

Description

Device and method for improving cognitive performance of sensor fusion using precision map {APPARATUS AND METHOD FOR IMPROVING COGNITIVE PERFORMANCE OF SENSOR FUSION USING PRECISE MAP}

The present invention relates to an apparatus and method for improving the cognitive performance of sensor fusion using a precision map, and more particularly, a sensor using a precision map capable of improving the performance of sensor fusion by combining precision map and precision positioning technology. It relates to an apparatus and a method for improving the cognitive performance of fusion.

An autonomous vehicle refers to a vehicle that independently determines a driving route by recognizing the surrounding environment through a function of detecting and processing external information during driving, and driving independently using its own power. Autonomous vehicles can travel to their destination by preventing collisions with obstacles on the driving path and adjusting the vehicle speed and driving direction according to the shape of the road, even if the driver does not operate the steering wheel, accelerator pedal, or brake. have. For example, acceleration may be performed on a straight road, and deceleration may be performed while changing a driving direction in response to a curvature of the road on a curved road.

The positioning system applied to autonomous vehicles is based on road map information built using GPS (Global Positioning System) and various sensors (Radar, LiDAR, Camera, etc.). The current position of the vehicle is determined through sensor data acquired through the sensor. In order to secure the stability of autonomous driving, it is important to accurately grasp the current position of the vehicle, and a function for detecting and avoiding obstacles in the driving section in the entire driving route from the starting point to the destination is required.

As described above, the autonomous vehicle integrates data recognized by various sensors (for example, radar, lidar, image sensor, etc.), and tracks each object to be recognized by one sensor. It uses sensor fusion technology that outputs information as if it were information. Since autonomous vehicles judge driving conditions and control the vehicle based on information output through sensor fusion technology, highly reliable sensor fusion technology is a key technology that must precede the development of autonomous driving technology.

However, conventionally, sensor fusion has been difficult to improve reliability due to unrecognized or misrecognized by relying only on perceived sensor data.

Accordingly, there is a need to develop a technology capable of improving the performance of sensor fusion by incorporating precision maps and precise positioning technologies into sensor fusion logic.

Background art of the present invention is disclosed in Korean Patent Application Publication No. 10-2017-0098071 (published on August 29, 2017).

The present invention was conceived to improve the above problems, and an object of the present invention is to improve the cognitive performance of sensor fusion using a precision map that can improve the performance of sensor fusion by combining precision map and precision positioning technology. It is to provide an apparatus and method.

The problem to be solved by the present invention is not limited to the problem(s) mentioned above, and another problem(s) not mentioned will be clearly understood by those skilled in the art from the following description.

An apparatus for improving perception performance of sensor fusion using a precision map according to an aspect of the present invention includes: a location search unit for obtaining location data of a vehicle, a camera unit for photographing an image around the vehicle, and a surrounding situation of the vehicle. The surrounding situation collecting unit to be collected, a storage unit for storing a precision map including navigation information and topographical features on the driving road, and the location data and camera lane data of the vehicle received from the location search unit and the camera unit are accurately processed. Map and map matching to estimate the precise location of the vehicle, pre-processing the surrounding situation data received from the surrounding situation collecting unit to output the integration target information, and the integration target information based on the precise location of the vehicle After mapping on the map, it includes a control unit for estimating a location based on the lane of the precision map.

In the present invention, the control unit is a first pre-process of estimating the precise location of the vehicle by mapping the location data of the vehicle and the camera lane data with the precision map, and converting the surrounding map data of the vehicle into the host vehicle center coordinate system. Second, a second preprocessing unit that preprocesses the surrounding situation data to output the integration target information, displays the preprocessed integration target information on a precision map around the host vehicle coordinate system, and selects a tracking target from the integration target information And a sensor fusion unit that predicts the location of the tracking target based on the lane of the precision map and estimates the location of the tracking target based on the predicted location and the measured location.

In the present invention, the second preprocessor may match the coordinate system and unit of the information to be integrated with the coordinate system and unit of the first preprocessor.

In the present invention, the sensor fusion unit converts the lane and the lane center line of the precision map into the own vehicle center coordinate system, displays the integration target information on the converted precision map, and integrates displayed on the converted precision map. A mapping module for clustering the same target from target information, a tracking target selection module for selecting at least one tracking target from among the clustered targets and classifying the tracking target based on the lane of the precision map, and the precision map lane And a position prediction module that predicts a location of the tracking target based on a trajectory and an estimated location, and estimates an actual location of the tracking target based on the predicted location and the measured location.

In the present invention, the mapping module may cluster the same target within a predetermined distance from the integrated target information displayed on the converted precision map as a single target.

In the present invention, the tracking target selection module includes a left lane target, a right lane target, a vehicle rear target, and a vehicle preceding the selected tracking target based on the lane of the precision map and the relative position in the longitudinal/transverse direction with the vehicle. It can be classified as at least one of a target and an out-of-range deletion target.

In the present invention, the position prediction module may predict the current position using an existing estimated position under the assumption that the target to be tracked moves on a lane center line.

In the present invention, the location prediction module may predict a location by applying a weight for a horizontal direction on the precision map of the tracking target.

In the present invention, the sensor fusion unit further includes a post-processing module for selecting at least one of a top priority target required for longitudinal speed control, a cut-in target, and a target to be considered when changing lanes based on the estimated position of the tracking target. can do.

According to another aspect of the present invention, a method for improving recognition performance of sensor fusion using a precision map is provided by a control unit receiving vehicle location data, camera lane data, and surrounding situation recognition data, estimating the precise location of the vehicle, and an integration target. Generating information, the controller mapping the integrated target information on the precision map based on the precise location of the vehicle, and estimating the location based on the lane of the precision map.

In the present invention, the step of generating the integrated target information includes, by the control unit, matching the location data of the vehicle and the camera lane data with the precision map to estimate the precise location of the vehicle, and to determine the surrounding map data of the vehicle. Converting to a vehicle center coordinate system, the control unit pre-processing the surrounding situation data may include the step of outputting the integration target information.

In the present invention, the step of estimating a location based on the lanes of the precision map may include, by the control unit, displaying the preprocessed integration target information on a precision map around the host vehicle coordinate system, and the control unit Selecting a tracking target from among, the control unit predicting the location of the tracking target based on the lane of the precision map, and estimating the location of the tracking target based on the predicted location and the measured location can do.

In the step of displaying the preprocessed integration target information of the present invention, the control unit may cluster the same target within a predetermined distance from the integration target information displayed on the converted precision map into a single target.

In the step of selecting the tracking target of the present invention, the control unit determines the selected target to be tracked based on the lane of the precision map and the relative position of the vehicle in the vertical/transverse direction as a left lane target, a right lane target, and a vehicle tail. It may be classified as at least one of a target, a vehicle preceding target, and an out-of-range deletion target.

In the step of estimating the location of the tracking target according to the present invention, the controller may predict the current location using the existing estimated location on the assumption that the tracking target moves on a lane center line.

In the step of estimating the location of the tracking target according to the present invention, the control unit may predict the location by applying a weight in the horizontal direction on the precision map of the tracking target.

In the present invention, the control unit may further include selecting at least one of a top priority target required for longitudinal speed control, a cut-in target, and a target to be considered when changing lanes, based on the estimated position of the tracking target. .

An apparatus and method for improving the cognitive performance of sensor fusion using a precision map according to an aspect of the present invention can delete a misrecognized object by correcting the surrounding situation data based on the lane of the precision map, and track an unrecognized object. can do.

An apparatus and method for improving the cognitive performance of sensor fusion using a precision map according to another aspect of the present invention may improve the performance of sensor fusion by combining a precision map and a precision positioning technology.

Meanwhile, the effects of the present invention are not limited to the above-mentioned effects, and various effects may be included within a range that is apparent to a person skilled in the art from the contents to be described below.

1 is a block diagram showing an apparatus for improving cognitive performance of sensor fusion using a precision map according to an embodiment of the present invention.
FIG. 2 is a block diagram for explaining in detail the control unit illustrated in FIG. 1.
3 is a block diagram illustrating in detail the sensor fusion unit shown in FIG. 2.
4 is an exemplary diagram for explaining an operation of a control unit according to an embodiment of the present invention.
5 is an exemplary view for explaining the operation of the fapping module according to an embodiment of the present invention.
6 is an exemplary diagram for explaining the operation of the tracking target selection module according to an embodiment of the present invention.
7 is an exemplary diagram for explaining an operation of a position prediction module according to an embodiment of the present invention according to an embodiment of the present invention.
8 is an exemplary diagram for explaining lane-based position prediction according to an embodiment of the present invention.
9 is an exemplary diagram for explaining an operation of a post-processing module according to an embodiment of the present invention.
10 is a flowchart illustrating a method for improving cognitive performance of sensor fusion using a precision map according to an embodiment of the present invention.
11 is a flowchart illustrating a method of estimating a location of information to be integrated according to an embodiment of the present invention.

Hereinafter, an apparatus and method for improving cognitive performance of sensor fusion using a precision map according to an embodiment of the present invention will be described with reference to the accompanying drawings. In this process, the thickness of the lines or the size of components shown in the drawings may be exaggerated for clarity and convenience of description.

In addition, terms to be described later are terms defined in consideration of functions in the present invention, which may vary according to the intention or custom of users or operators. Therefore, definitions of these terms should be made based on the contents throughout the present specification.

Further, the implementation described herein may be implemented in, for example, a method or process, an apparatus, a software program, a data stream or a signal. Although discussed only in the context of a single form of implementation (eg, only as a method), the implementation of the discussed features may also be implemented in other forms (eg, an apparatus or program). The device may be implemented with appropriate hardware, software and firmware. The method may be implemented in an apparatus such as a processor, which generally refers to a processing device including, for example, a computer, a microprocessor, an integrated circuit or a programmable logic device, or the like. Processors also include communication devices such as computers, cell phones, personal digital assistants ("PDAs") and other devices that facilitate communication of information between end-users.

1 is a block diagram showing an apparatus for improving cognitive performance of sensor fusion using a precision map according to an embodiment of the present invention, FIG. 2 is a block diagram for explaining in detail the control unit shown in FIG. 1, and FIG. 3 is A block diagram for explaining the sensor fusion unit shown in 2 in detail, FIG. 4 is an exemplary view for explaining the operation of a control unit according to an embodiment of the present invention, and FIG. 5 is an operation of a fapping module according to an embodiment of the present invention. 6 is an exemplary view for explaining the operation of the tracking target selection module according to an embodiment of the present invention, and FIG. 7 is an exemplary view for explaining the operation of the tracking target selection module according to an embodiment of the present invention. An exemplary diagram for explaining the operation of the position prediction module, FIG. 8 is an exemplary diagram for explaining lane-based position prediction according to an embodiment of the present invention, and FIG. 9 is an operation of a post-processing module according to an embodiment of the present invention. It is an exemplary diagram for explaining

Referring to FIG. 1, an apparatus for improving the cognitive performance of sensor fusion using a precision map according to an embodiment of the present invention includes a location search unit 110, a camera unit 120, a vehicle state collection unit 130, and a surrounding area. It includes an environment collection unit 140, a storage unit 150, and a control unit 160.

The location search unit 110 may obtain the location data of the vehicle and provide it to the control unit 160. Here, the location search unit 110 may search for location data of the vehicle through one or more of satellite navigation and guessing navigation.

At this time, satellite navigation is to acquire vehicle location data based on GNSS (Global Navigation Satellite System), such as GPS (Global Positioning System), GLONASS (Global Navigation Satellite System), Galileo, Beidou, etc. It is possible to receive the absolute coordinates of the vehicle from the system of, and based on this, it is possible to generate location data (latitude and longitude coordinates, direction, speed, quality, etc.). In addition, speculative navigation creates vehicle location data based on the speed acquired from the vehicle's speedometer (not shown), gyro sensor (not shown), and geomagnetic sensor (not shown), and the vehicle's DR (Dead Reckoning) information. can do.

The camera unit 120 may capture an image around the vehicle and detect camera data such as lanes, traffic lights, surrounding vehicles, pedestrians, and obstacles. The camera unit 120 provides the detected camera data to the controller 160, and the camera data may include camera lane data.

The vehicle state collection unit 130 collects the driving state of the vehicle and provides it to the control unit 160. For example, the vehicle state collection unit 130 may collect and provide a steering state, a braking state, an acceleration state, and a driving state of the vehicle.

The surrounding environment collection unit 140 collects the surrounding situation of the vehicle and provides it to the control unit 160. Here, the vehicle state collection unit 130 may include various sensors such as a radar (not shown) and a rider (not shown), and detect surrounding situation awareness data including lanes, traffic lights, surrounding vehicles, pedestrians, and obstacles. And, the detected surrounding situation recognition data may be provided to the control unit 160.

The storage unit 150 stores a precision map including navigation information and topographic features for a driving road for autonomous driving. Here, the navigation information may include a map version, a link ID, a link distance, whether a rotating link, a road grade, and a prescribed speed for each road grade.

In addition, the storage unit 150 may store a high-precision map of a road and provide information on lanes, traffic lights, signs, curbs, road marks, and various structures.

The controller 160 estimates the precise location of the vehicle by matching the map with the precision map stored in the storage unit 150 with the location data of the vehicle and the camera lane data transmitted from the location search unit 110 and the camera unit 120, Pre-processing the surrounding situation perception data received from the vehicle state collection unit 130 to output the integrated target information, map the integrated target information on a precision map centering on the precise location of the vehicle, and then based on the lanes of the precision map. Estimate the location.

That is, the control unit 160 estimates the precise location of the vehicle by matching the vehicle location data with the precision map, maps the surrounding situation data around the precise location of the vehicle, and recognizes the surrounding situation based on the lane of the precision map. Correct the data. At this time, the control unit 160 can delete the falsely recognized object and track the unrecognized object by correcting the surrounding situation data based on the lane of the precision map. For example, as illustrated in FIG. 4, the controller 160 may improve sensor fusion performance by mapping the surrounding situation perception data to a high-precision map.

As shown in FIG. 2, the control unit 160 includes a first preprocessor 161a, a second preprocessor 161b, and a sensor fusion unit 162.

The first preprocessor 161a estimates the precise location of the vehicle by matching the vehicle location data and the camera lane data with a precision map, and converts the surrounding map data of the vehicle into the host vehicle center coordinate system.

That is, the first preprocessor 161a may accurately estimate the location of the vehicle by using the vehicle location data, camera lane data, and a precision map. Thereafter, the first preprocessor 161a may convert the map data around the vehicle into a vehicle center coordinate system and provide it to the control unit 160.

The second preprocessing unit 161b preprocesses the surrounding situation-aware data and outputs integration target information. That is, the second preprocessor 161b may pre-process the surrounding situation perception data and output integrated target information required for sensor fusion. Here, the information to be integrated may include surrounding vehicles, pedestrians, and obstacles.

In addition, the second pre-processing unit 161b may convert the surrounding situation perception data into data that can be used in sensor fusion. In this case, the second preprocessor 161b may match the coordinate system and unit of the information to be integrated with the coordinate system and unit of the first preprocessor 161a.

For example, since the vehicle state collection unit 130 is composed of various sensors, the size, length, and coordinate system of the detected object may be different for each sensor. Accordingly, the second preprocessor 161b may match units such as size and length, and a coordinate system so that the surrounding situation perception data received from various sensors can be utilized in sensor fusion.

The sensor fusion unit 162 displays the integrated target information preprocessed by the second preprocessor 161b on a precision map around the own vehicle's coordinate system, selects a tracking target from the integrated target information displayed on the precision map, and The location of the target to be tracked is predicted based on the lanes of the map, and the location of the target to be tracked is estimated based on the predicted and measured locations.

The sensor fusion unit 162 may estimate the location of the tracking target based on the lane after mapping the integrated target information on the precision map.

As shown in FIG. 3, the sensor fusion unit 162 includes a mapping module 163, a tracking target selection module 164, a location prediction module 165, and a post-processing module 166.

The mapping module 163 converts the lanes and the lane center lines of the precision map into the vehicle center coordinate system, displays the integrated target information on the converted precision map, and selects the same target from the integrated target information displayed on the converted precision map. Cluster.

Specifically, the mapping module 163 converts the lane and the lane center line of the precision map into the vehicle center coordinate system based on the precise location of the vehicle. For example, the mapping module 163 may display the lane and the lane center line of the precision map around the precise location A of the vehicle as shown in FIG. 5A.

Then, the mapping module 163 displays the information to be integrated on the precision map centering on the precise location of the vehicle. For example, when the mapping module 163 displays the integration target information as shown in FIG. 5(b) in FIG. 5(a), it may be displayed as in FIG. 5(c).

Thereafter, the mapping module 163 clusters the same object within a predetermined distance from the integrated object information displayed on the precision map as a single object based on the precise location of the vehicle. Here, clockstering may mean grouping the same objects and classifying them into a single object. For example, the mapping module 163 may cluster the same object as shown in (c) of FIG. 5.

The tracking target selection module 164 selects at least one tracking target from among targets clustered by the mapping module 163 and classifies the tracking target based on a lane of the precision map. At this time, the tracking target selection module 164 considers sensor characteristics (eg, measurement range, vertical/horizontal measurement accuracy, tracking reliability, etc.), and selects the tracking target based on the tracking target information before a predetermined time. Can be selected. Here, the tracking target information before a preset predetermined time may mean, for example, tracking target information 0.5 seconds before.

The tracking target selection module 164 may select a target that may pose a risk to the host vehicle as a tracking target, centering on the host vehicle. For example, the tracking target selection module 164 may select a preceding vehicle, a pedestrian, a left vehicle, a right vehicle, and the like as a tracking target.

When the tracking target is selected, the tracking target selection module 164 may classify the selected tracking target based on the lane of the precision map and the relative position of the vehicle in the longitudinal/lateral directions. For example, the tracking target selection module 164 may classify a tracking target of a predetermined distance or more (a target with a low possibility of giving a risk to the host vehicle) as an out-of-range target. In addition, the tracking target selection module 164 may classify each of a left lane target, a right lane target, a vehicle rear target, and a vehicle preceding target within a predetermined distance.

For example, as shown in FIG. 6, when the host vehicle A and the integrated target information are displayed, the tracking target selection module 164 may delete the target at a distance E or more from the target to be tracked because the target is outside the range. In addition, the tracking target selection module 164 may classify the tracking target into a left lane target (B), a right lane target (C), a vehicle rear target (D), and a vehicle preceding target (not shown).

The location prediction module 165 predicts the location of the tracking target classified in the tracking target selection module 164 based on the trajectory and the estimated location of the precision map lane, and the actual tracking target is based on the predicted location and the measured location. Estimate the location. At this time, the location prediction module 165 may predict the current location based on the existing estimated location on the assumption that the tracking target moves on the lane center line.

That is, the location prediction module 165 may predict a location after a specific time of the tracking target based on the trajectory of the precision map lane and the existing estimated location. At this time, the calculation time can be shortened by excluding the object outside the range from the estimation object. Then, the location prediction module 165 may estimate the location of the actual target based on the predicted location and the measured location.

For example, a case of estimating the actual position of F in FIG. 7A will be described. In this case, the location prediction module 165 may predict a location b after a specific time by using the existing estimated location a of F as shown in FIG. 7B. In this case, the position prediction module 165 may predict the position based on the existing position a on the assumption that F moves on the lane center line. Then, the location prediction module 165 may estimate the location as shown by d by using the predicted location b and the measured location c.

In addition, the position prediction module 165 may estimate the actual position by correcting the predicted position using the measured position. In this case, the location prediction module 165 may predict a location by applying a weight for a horizontal direction on a precision map of the tracking target. That is, since the position prediction module 165 predicts the location of the target to be tracked using the path of the lane centerline, the error variance in the lateral direction based on the lane centerline may be greater than the error variance in the vertical direction. Accordingly, the position prediction module 165 may predict the location of the target to be tracked by adding a weight for the horizontal direction on the precision map in order to reduce the error variance in the horizontal direction.

In addition, the position prediction module 165 may reduce the weight of inaccurate lateral information on a curved road or the like and predict a position using a path of a lane centerline. In particular, the location prediction module 165 may predict a location using a road path when the curvature change of the road is severe.

For example, in FIG. 8, if e is an existing location, if the location is predicted using the measured value, it may be the location of g, and if predicted using the path of the map, it may be the location of f.

In addition, the location prediction module 165 may estimate the location and speed of the target to be tracked by integrating the predicted current location and the measured current location.

On the other hand, the apparatus for improving the cognitive performance of sensor fusion using a precision map according to the present invention needs to provide tracking information of surrounding objects in a desired form and information from a controller (eg, ECU, etc.).

Accordingly, the post-processing module 166 selects at least one of a top priority target required for longitudinal speed control, a cut-in target, and a target to be considered when changing lanes, based on the position estimated by the position prediction module 165. .

That is, the post-processing module 166 may transmit selected information that is difficult to determine with the surrounding situation recognition data alone, based on the estimated information of the tracking target. For example, as shown in FIG. 9, the post-processing module 166 includes information on the highest priority target (preceding vehicle) required for longitudinal speed control, information on the target for interruption, target information necessary for determining whether a lane change is possible (target information on the left/right lane), etc. Can be selected.

As described above, by correcting sensor data based on the lane of the precision map, the controller 160 can delete the falsely recognized object and track the unrecognized object. Also, the controller 160 may correct the recognition target information based on the lane information.

10 is a flowchart illustrating a method for improving cognitive performance of sensor fusion using a precision map according to an embodiment of the present invention.

Referring to FIG. 10, when receiving vehicle location data, camera lane data, and surrounding situation recognition data (S1000), the controller 160 estimates the precise location of the vehicle using the received data and generates integrated target information. (S1010). In this case, the control unit 160 may estimate the precise location of the vehicle by matching the vehicle location data and the camera lane data with the precision map, and may pre-process the surrounding situation recognition data to generate the integrated target information.

When step S1010 is performed, the controller 160 maps the integrated target information on the precision map based on the precise location of the vehicle, and then estimates the location of the integrated target information based on the lane of the precision map (S1020).

For a detailed description of how the control unit 160 estimates the location of the information to be integrated, refer to FIG. 11.

11 is a flowchart illustrating a method of estimating a location of information to be integrated according to an embodiment of the present invention.

Referring to FIG. 11, the controller 160 maps integrated target information on a precision map based on a precise location of a vehicle, and clusters the same target (S1100). That is, the control unit 160 converts the lane and the lane center line of the precision map into the vehicle center coordinate system, displays the integrated target information on the converted precision map, and displays the same target from the integrated target information displayed on the converted precision map. Is clustered.

When step S1100 is performed, the controller 160 selects at least one tracking target from among the clustered targets, and classifies the tracking target based on the lane of the precision map (S1110). That is, the controller 160 considers sensor characteristics (eg, measurement range, vertical/lateral measurement accuracy, tracking reliability, etc.), and selects a tracking target based on tracking target information before a predetermined time, Selected tracking targets can be classified based on the lanes of the precision map and the relative position of the vehicle in the longitudinal/transverse directions.

When step S1110 is performed, the controller 160 predicts the location of the tracking target based on the trajectory and the estimated location of the precision map lane (S1120). At this time, the location prediction module 165 may predict the current location based on the existing estimated location on the assumption that the tracking target moves on the lane center line.

When step S1120 is performed, the controller 160 corrects the predicted position using the measured position and estimates the actual position (S1130). In this case, the controller 160 may predict a location by applying a weight for a horizontal direction on the precision map of the tracking target.

When step S1130 is performed, the control unit 160 selects at least one of a top priority target required for longitudinal speed control, a cut-in target, and a target to be considered when changing lanes based on the estimated position of the tracking target (S1140). .

As described above, in the apparatus and method for improving the cognitive performance of sensor fusion using a precision map according to an aspect of the present invention, by correcting the surrounding situation data based on the lane of the precision map, it is possible to delete a misrecognized object, You can track unknown objects.

An apparatus and method for improving the cognitive performance of sensor fusion using a precision map according to another aspect of the present invention may improve the performance of sensor fusion by combining a precision map and a precision positioning technology.

Although the present invention has been described with reference to the embodiments shown in the drawings, this is only illustrative, and those of ordinary skill in the field to which the technology pertains, various modifications and other equivalent embodiments are possible. I will understand.

Therefore, the true technical protection scope of the present invention should be determined by the following claims.

110: location search unit
120: camera unit
130: vehicle state collection unit
140: surrounding landscape collection unit
150: storage unit
160: control unit

Claims (17)

A location search unit for obtaining location data of the vehicle;
A camera unit for capturing an image around the vehicle;
A surrounding situation collecting unit for collecting the surrounding situation of the vehicle;
A storage unit for storing a precision map including navigation information and topographic features for the driving road; And
Map-matching the location data and camera lane data of the vehicle received from the location search unit and the camera unit with the precision map to estimate the precise location of the vehicle, and pre-process the surrounding situation data received from the surrounding situation collection unit. The control unit outputs the integration target information, maps the integration target information on the precision map based on the precise location of the vehicle, and estimates the location based on the lane of the precision map
Device for improving cognitive performance of sensor fusion using a precision map comprising a.
The method of claim 1,
The control unit,
A first preprocessing unit for estimating the precise location of the vehicle by mapping the location data of the vehicle and the camera lane data with the precision map, and converting the surrounding map data of the vehicle into the own vehicle center coordinate system;
A second preprocessing unit preprocessing the surrounding situation data and outputting information to be integrated; And
Displaying the preprocessed integrated target information on a precision map around the own vehicle coordinate system, selecting a tracking target from the integrated target information, predicting the location of the tracking target based on the lane of the precision map, and the An apparatus for improving cognitive performance of sensor fusion using a precision map, comprising: a sensor fusion unit that estimates the location of the target to be tracked based on the predicted location and the measured location.
The method of claim 2,
The second pretreatment unit,
An apparatus for improving cognitive performance of sensor fusion using a precision map, characterized in that the coordinate system and unit of the integrated target information are matched with the coordinate system and unit of the first preprocessor.
The method of claim 2,
The sensor fusion unit,
Converting the lane and the lane center line of the precision map into the own vehicle center coordinate system, displaying the integrated object information on the converted precision map, and clustering the same object in the integrated object information displayed on the converted precision map Mapping module;
A tracking target selection module for selecting at least one tracking target from among the clustered targets and classifying the tracking target based on a lane of the precision map; And
And a location prediction module that predicts the location of the tracking target based on the trajectory and the estimated location of the precision map lane, and estimates the actual location of the tracking target based on the predicted location and the measured location. A device for improving the cognitive performance of sensor fusion using precision maps.
The method of claim 4,
The mapping module,
An apparatus for improving cognitive performance of sensor fusion using a precision map, characterized in that the same object within a predetermined distance from the integrated object information displayed on the converted precision map is clustered as a single object.
The method of claim 4,
The tracking target selection module,
Classify the selected tracking target into at least one of a left lane target, a right lane target, a vehicle rear target, and a vehicle preceding target, and an out-of-range deletion target based on the lane of the precision map and the relative position of the vehicle in the longitudinal/lateral direction An apparatus for improving cognitive performance of sensor fusion using a precision map, characterized in that.
The method of claim 4,
The position prediction module,
An apparatus for improving cognitive performance of sensor fusion using a precision map, characterized in that the current location is predicted using an existing estimated location under the assumption that the tracking target moves on a lane center line.
The method of claim 4,
The position prediction module,
A device for improving cognitive performance of sensor fusion using a precision map, characterized in that the position is predicted by applying a weight in a horizontal direction on the precision map of the tracking target.
The method of claim 4,
Based on the estimated position of the tracking target, the precision map further comprises a post-processing module for selecting at least one of a top priority target required for longitudinal speed control, a cut-in target, and a target to be considered when changing lanes. Device for improving cognitive performance of sensor fusion using.
Receiving, by a controller, vehicle location data, camera lane data, and surrounding situation recognition data, estimating a precise location of the vehicle, and generating integrated target information; And
Mapping, by the controller, the integrated target information on the precision map based on the precise location of the vehicle, and estimating the location based on the lane of the precision map.
Method for improving cognitive performance of sensor fusion using a precision map comprising a.
The method of claim 10,
Generating the integration target information,
Estimating the precise location of the vehicle by mapping the vehicle location data and camera lane data with the precision map, and converting the surrounding map data of the vehicle into a coordinate system of the host vehicle; And
And outputting, by the control unit, pre-processing the surrounding situation data and outputting information to be integrated.
The method of claim 11,
The step of estimating a location based on the lane of the precision map,
Displaying, by the control unit, the preprocessed integrated target information on a precision map around the coordinate system of the own vehicle;
Selecting, by the control unit, a tracking target from the integration target information; And
The control unit predicts the location of the tracking target based on the lane of the precision map, and estimates the location of the tracking target based on the predicted location and the measured location. A method for improving cognitive performance of sensor fusion using.
The method of claim 12,
In the step of displaying the preprocessed integration target information,
The control unit clusters the same object within a predetermined distance from the integrated object information displayed on the converted precision map into a single object.
The method of claim 12,
In the step of selecting the tracking target
The control unit selects the selected tracking target based on the lane of the precision map and the relative position of the vehicle in the vertical/transverse direction, among a left lane target, a right lane target, a vehicle rear target, and a vehicle preceding target, and an out of range deletion target. A method for improving cognitive performance of sensor fusion using a precision map, which is classified into at least one.
The method of claim 12,
In the step of estimating the location of the tracking target,
The control unit predicts the current position using an existing estimated position on the assumption that the target to be tracked moves on a lane center line.
The method of claim 12,
In the step of estimating the location of the tracking target,
The control unit predicts a location by applying a weight in a horizontal direction on the precision map of the target to be tracked.
The method of claim 12,
The control unit further comprises the step of selecting at least one of a top priority object required for longitudinal speed control, an interruptible object, and an object to be considered when changing lanes, based on the estimated position of the tracking object. A method for improving cognitive performance of sensor fusion using maps.

Images