WO2005090918A1

WO2005090918A1 - Method and device for precisely indicating driving instructions of a navigation system

Info

Publication number: WO2005090918A1
Application number: PCT/EP2005/050865
Authority: WO
Inventors: Antonius Johannes Hendriks; Thorsten KÖHLER; Hans-Wilhelm Ruehl
Original assignee: Siemens Aktiengesellschaft
Priority date: 2004-03-18
Filing date: 2005-03-01
Publication date: 2005-09-29
Also published as: DE102004013440A1

Abstract

The invention relates to a method and device for precisely indicating driving instructions of a navigation system, particularly for motor vehicles. According to the invention, data, which is obtained by a camera system and which concerns the driving surroundings, particularly in the front area, of the motor vehicle are fed into the navigation system. The precisely indicated driving instructions are based, for example, essentially on the type (solid line, dashed line, changing from a solid line to a dashed line and vice versa) of marking of a traffic lane or of other marked objects/marks identified by the camera system. The invention advantageously minimizes the output of unnecessary and/or misleading driving instructions of a navigation system to the drive of a motor vehicle. At the same time, it advantageously reduces the risk of accidents in road traffic.

Description

Procedure and. Device for specifying the driving instructions of a navigation system

The invention relates to a method and a device for specifying the driving instructions of a navigation system, in particular for motor vehicles.

Navigation systems in vehicles have been widely used in recent years. The navigation data provide a global view of the driving environment from which the behavioral instructions are generated (e.g. "turn right now"). The position of the vehicle in the map data is also often not known with sufficient accuracy. As a result, the exact time of the driving instructions can in many cases not be selected correctly by the navigation system.

Another problem that arises from the coarse assignment between map data and vehicle position is the generation of helpful intermediate instructions that point to the actual driving instruction. Example: the vehicle is traveling in the middle of three lanes, the next driving instruction would be: "now turn right". In this example, the intermediate message "Please change to the right lane" would be very helpful. Ideally, the instruction "now turn right" would only come after this driving maneuver has been carried out.

So far, information that is available to the navigation system (GPS position and map material), as well as vehicle data (odometry) such as wheel speeds, speed, steering angle and the like have been used. The inaccuracy of the position determination with respect to the actual driving environment can be up to 50m. In future vehicles, in addition to navigation systems, camera systems for driver assistance functions such as lane detection or monitoring, for blind spots and lane changes, for obstacle warnings, for pre-crash sensing, for traffic sign recognition or the like will also play an important role. In contrast to the map data, these camera systems allow a precise local view of the driving environment.

In this regard, DE 199 37 489 A1, for example, discloses the combination of an environment detection with a GPS system. However, the generation of driving instructions is not disclosed, nor is the effect of image processing on them.

No. 5,922,036 AI also discloses a method in which the number of lanes and the number of one's own lane are determined by means of a camera. Driving instructions are then generated from this together with a GPS-supported navigation system. A disadvantage of this method, however, is the determination of the number of lanes, since this can sometimes not be clearly determined by the perspective projection. There is also the risk of confusing the lane markings of the outer lanes with e.g. the guardrail. Another disadvantage is the possible one

Outer lines hidden by other vehicles. The type of lane marking (solid, dashed) is not recognized.

Finally, a method is known from US Pat. No. 6,577,334 B1, in which a branch from a "highway" and the rough number of the lane are recognized by means of a camera. This is used together with a GPS-supported navigation system to control the vehicle (engine control, transmission control, 4WD, ABS, suspension, ...). The generation of driving instructions is described, but not in detail executed. In contrast to US 5,922,036 AI, lane recognition consists of three components:

A junction detection system that is based on the measurement of the widening of a lane marking and additionally or alternatively the widening of the lane. This is described in great detail.

A track position detection, which, however, does not recognize the number of tracks as in US 5,922,036 AI, but bypasses the main disadvantage and uses the type of marking. If one line is solid, the other is dashed, the vehicle is on an edge lane. If both lines are dashed, the vehicle is in a middle lane.

A measurement of a lane change.

It is an object of the present invention to provide an improved method and a device for specifying the driving instructions of a navigation system, in particular for motor vehicles, which avoids the aforementioned disadvantages and detects the aforementioned circumstances more reliably. In particular, technical measures must be specified to achieve the following goals: preparation and improvement of the driving instructions and improvement of the timing of the driving instructions.

This object is achieved by the features of the independent claims.

Advantageous refinements and developments, which can be used individually or in combination with one another, are the subject of the dependent claims.

In the method for specifying the driving instructions of a navigation system, in particular for motor vehicles, data of the driving environment, in particular in the front area, of the Motor vehicle fed into the navigation system, which were obtained by means of a camera system, the precise driving instructions being essentially based on the type of a lane identified by the camera system (solid line, dashed line, change from solid line to dashed and vice versa).

In contrast to US Pat. No. 5,922,036 AI and in particular US Pat. No. 6,577,334 B1, the present invention advantageously makes an important simplification. The lane recognition and the driving instructions generated from it are essentially based on the recognized type of lane marking. As a result, a lane model in particular is not required, which advantageously saves considerable computing time.

Alternatively or cumulatively, driving instructions can be specified essentially on the basis of turning arrows ^• recognized by the camera system, stop lines, crosswalks, etc. of one's own and / or on an adjacent lane.

In contrast to non-changing markings of lanes, zebra crossings, arrows, stop lines, intersections, road junctions, traffic lights, and the like have a defined longitude and latitude, which advantageously allow a precise comparison with a corresponding entry in a map set, as a result of which a motor vehicle is more accurate can be localized and consequently navigated more precisely.

The mere recognition of a traffic sign, a zebra crossing, a traffic light, a stop line or the like thus allows an exact comparison of the driving environment of the motor vehicle with stored map data.

However, this applies not only to stationary objects but also to other road users, i.e. moving objects. For example, the position of a crossing vehicle in the image of the camera gives an indication of the exact one Position of your own vehicle relative to a distinctive road event entered in the navigation maps or data, such as intersections, junctions, exits or the like.

Therefore, alternatively or cumulatively, driving instructions can essentially be specified on the basis of traffic signs, traffic lights, oncoming and cross traffic, etc. recognized by the camera system, ie, when listed, this advantageously allows driving instructions to be generated depending on the current traffic situation.

For example, driving instructions can be generated that relate to the type of lane currently being traveled (straight ahead, right turn, left turn, etc.).

Alternatively or cumulatively, driving instructions can be generated that relate relatively to the lane number currently being used (e.g. try to change two lanes to the left).

According to the invention, unnecessary driving instructions can advantageously be suppressed for the first time. For example, if the vehicle is already in a correct lane, or if the type of lane marking does not allow maneuvers, ie no "please turn if possible" with a solid center line or the like.

As an alternative or cumulatively, driving instructions or reminders of a previous driving instruction can be issued for the first time as soon as a change in the lane marking allows a driving maneuver (eg "The median now allows turning" or the like) and / or as soon as the type of lane marking changes a planned maneuver allows or requires (eg "Now right on the

Change turn lane "after the start of a turn track is recognized from the beginning of a broken line).

The invention also relates to a device for carrying out the methods described above, namely a navigation system which comprises:

- A camera system that detects the driving environment, especially in the front area, of the motor vehicle;

an evaluation unit for the image data recorded by the monitoring unit;

a computing unit in which driving instructions are generated on the basis of prepared map data, a position determination of the motor vehicle e.g. by means of GPS, odometry and / or the like, and the evaluation results of the image data recorded by the camera system.

The means for carrying out the methods according to the invention have the advantages listed in this regard.

In a first development, the evaluation unit is part of the computing unit, the image data from the camera system preferably being transmitted to the evaluation unit via an LVDS connection.

Alternatively or cumulatively, an evaluation unit is provided in the camera system, which partially or completely (pre-) evaluates the captured image data, with partially or fully evaluated image data using a CAN, RS232, MOST, LIN or other suitable connection to the computing unit be transmitted.

The present invention is based on the knowledge of making use of the precision of the driving instructions of a navigation system, in particular for motor vehicles, of a camera system, for example the type of marking a lane (solid line, dashed line, change from solid line to dashed and vice versa) reliably detect.

The invention thus advantageously allows the position of a motor vehicle to be compared to markings stored in map data, such as lane changes, arrows, crosswalks, traffic lights, etc., to the map data, and thus more precise navigation than is known in the prior art.

It also advantageously minimizes the output of unnecessary and / or misleading driving instructions from a navigation system to the driver of a motor vehicle in road traffic.

Additional details and further advantages of the invention are described below on the basis of a preferred exemplary embodiment.

A camera system consisting of a camera and an image processing computer is mounted, for example, on the windscreen in the area of the interior mirror. The image area includes the course of the road in front of the vehicle with a total horizontal viewing angle of approx. 50 °. The camera system recognizes its own lane, the type of lane marking, the number of existing lanes, turning arrows, stop lines, traffic signs, crosswalks, traffic lights and / or cross traffic, etc. The information is sent as a message e.g. fed into the CAN bus, ie into the network hardware for data transmission in the vehicle. These messages / data are read out of the CAN by a navigation system and converted into improved driving instructions.

The system typically consists of: - a navigation system, which processes the map data, determines the position of the vehicle, for example using GPS and odometry, and generates driving instructions; - A camera system that detects the driving environment in the front area of the vehicle; such as

- Communication between camera system and navigation system at least in the direction: camera to navigation system.

The navigation system must have an arithmetic unit, the camera system can have its own arithmetic unit. The following cases are alternatively or cumulatively conceivable: - the camera transfers the image data to its own computing unit, the image data is evaluated on this computing unit and the results of the evaluation are transferred to the navigation computer using suitable hardware (for example CAN, RS232, MOST, LIN) , ....); - The camera transfers the image data to the navigation computer (e.g. via LVDS), the image data is evaluated on the navigation unit and incorporated into the results in the driving instructions.

The preparation and improvement of driving instructions by means of image evaluation by a camera is characterized in particular by:

- recognition of the type of lane marking (e.g. solid line, dashed line); - A detection of the change of a lane to the right / wrong side with a corresponding message;

- Detection of turn arrows in one's own lane;

- Detection of turning arrows on the neighboring lanes;

- a possible generation of information relating to the type of lane currently being traveled (straight ahead, right turn, left turn);

- A possible generation of information that relates relatively to the current lane number used (eg "Try to change two lanes to the left."); - suppression of unnecessary information about changing lanes when the vehicle is already on a correct lane;

- ^' ' suppression of information if the type of 'lane marking does not allow maneuvering (eg no "if possible please turn" with solid center line);

- output of information or a reminder of a previous information as soon as a change in the lane marking permits a maneuver (e.g. "The median now allows a turn"); and or

- Outputting information as soon as the change in the type of lane marking permits or requires a planned maneuver (e.g. "Now switch to the right onto the turning lane" after the start of a turning lane is recognized on the right by the start of a broken line).

Would e.g. the next driving instruction is "turn right" and there is a solid line on the right side of the lane, so changing the lane is not allowed and therefore not necessary. However, if there is a dashed line on the right side, a change to the right lane would be possible and very likely - also useful. It could be the message "on the right

Change lane "and only then the message" Turn right. "If a dashed line appears again on the right side after changing the lane in the right direction, the request to change lane could be repeated.

In a further stage of the system, the camera detects turn arrows on its own lane and notifies the driver of a change of lane if the turn arrow (eg straight arrow) does not match the upcoming turning maneuver ("turn right"). In a further expansion stage of the system, it is also possible to take into account which turn arrow is present on one's own and the neighboring lanes.

The improvement in the timing of driving instructions is characterized by:

- a detection of the change in the type of lane marking (e.g. from a solid to a dashed line);

detection of the change in the type of lane marking, in particular at a distance of approximately 25 to 50 m in front of the vehicle;

- Detection of turn arrows;

- Detection of stop lines on the road;

- recognition of traffic signs (e.g. stop sign, observe the right of way, priority street, ...);

- detection of crosswalks;

- detection of traffic lights; and or

- detection of cross traffic.

In contrast to the prior art cited at the beginning, in particular to US 6,577,334 B1, motorway exits are not recognized by widening the lane marking or the lane itself, but advantageously simply by changing from solid to dashed lines. A further developed version of the system does not explicitly recognize departures, but the driving instruction can be generated simply because the type of lane marking is dashed at a certain point in time (as long as the type is solid, the instruction is suppressed). If a change of the right lane marking from a solid line to a dashed line is recognized on a freeway, the instruction "now turn right" can be given at this point. In a further advantageous embodiment of the system, this change is recognized approximately 25 to 50 m in front of the vehicle and the driving instructions are generated even earlier. For the clarification of the instruction time, all markings can be used which give an indication of an intersection area. These are turning arrows, stop lines, traffic signs, crosswalks and traffic lights. In particular, a comparison can be made with the information from the map data. For example, the position of crosswalks is available in map data.

Another indication of an intersection area is the detection of cross traffic.

The present invention advantageously minimizes the output of unnecessary and / or misleading driving instructions from a navigation system to the driver of a motor vehicle. The current position of the vehicle in the map data of a conventional navigation system can often only be determined to within 50m using GPS (shadowing) and odometry. The camera-based recognition of characteristic features such as turns, zebra stripes, stop lines, traffic signs, etc., which are recorded in the map data, enables the vehicle position to be compared in the map data for the first time, thereby making the timing of the driving instructions much more precise. The present invention thus advantageously reduces the risk of accidents in road traffic.

Claims

claims

1. A method for specifying the driving instructions of a navigation system, in particular for motor vehicles, in which data from the driving environment, in particular in the front area, of the motor vehicle is fed into the navigation system, which data were obtained by means of a camera system, the precise driving instructions essentially based on the type recognized by the camera system (solid line, dashed line, change from solid line to dashed and vice versa) of the marking of a lane.

2. The method, in particular according to claim 1, for specifying the driving instructions of a navigation system, in particular for motor vehicles, in which data of the driving environment, especially in the front area, of the motor vehicle are fed into the navigation system, which were obtained by means of a camera system, wherein the specified driving instructions are essentially based on turning arrows, stop lines, zebra crossings, etc., recognized by the camera system, of one's own and / or on an adjacent lane.

3. The method, in particular according to claim 1 or 2, for specifying the driving instructions of a navigation system, in particular for motor vehicles, in which data of the driving environment, especially in the front area, of the motor vehicle are fed into the navigation system, which were obtained by means of a camera system , the precise driving instructions being based essentially on traffic signs, traffic lights, oncoming and cross traffic, etc. recognized by the camera system.

4. The method according to any one of claims 1 to 3, in which driving instructions are generated which are based on the current the currently used lane type (straight ahead, right turn, left turn, etc.).

5. Method according to one of the preceding claims, in which driving instructions are generated which relate relatively to the track number currently being traveled on (for example, try to change two lanes to the left).

6. The method according to any one of the preceding claims, in which unnecessary driving instructions are suppressed (e.g. if the vehicle is already in a correct lane).

7. The method as claimed in one of the preceding claims, in which driving instructions are suppressed if the type of road marking does not permit the maneuver (for example no "if possible please turn over" with a solid center line).

8. The method according to any one of the preceding claims, in which driving instructions or reminders of a previous driving instruction are issued as soon as a change in the lane marking permits a driving maneuver (e.g. "the median now allows turning").

9. The method according to any one of the preceding claims, in which driving instructions are issued as soon as the change in the type of lane marking permits or requires a planned maneuver (for example "now switch to the right onto the turning lane" after the start of a turning lane based on the start of a right broken line is detected).

10. Navigation system for carrying out the method according to one of the preceding claims, comprising - a camera system which detects the driving environment, especially in the front area, of the motor vehicle; an evaluation unit for the image data recorded by the monitoring unit; a computing unit in which driving instructions are generated on the basis of prepared map data, a position determination of the motor vehicle, for example by means of GPS, odometry and / or the like, and the evaluation results of the image data recorded by the camera system.

11. Navigation system according to claim 10, characterized in that the evaluation unit is part of the computing unit.

12. Navigation system according to claim 11, characterized in that the image data are transmitted from the camera system to the evaluation unit via an LVDS connection.

13. Navigation system according to one of claims 10 to 12, characterized in that an evaluation unit is provided in the camera system which partially or completely evaluates the captured image data.

14. Navigation system according to claim 13, characterized in that partially or completely evaluated image data are transmitted to the computing unit by means of a CAN, RS232, MOST, LIN or other suitable connection.