DE102011119762A1 - Positioning system for motor vehicle, has processing unit that determines localized position of vehicle using vehicle movement data measured based on specific location data stored in digital card - Google Patents

Abstract

The position system has a digital card (DLK) that stores data over specific location such as landmarks. An environmental identifier device such as camera detects specific location data near the vehicle. A processing unit in location module (F) compares detected specific location data and stored location data to locate the vehicle position. An inertial measurement unit (IMU) determines vehicle movement data based on specific location data stored in digital card. The processing unit determines localized position of vehicle using vehicle movement data. An independent claim is included for method of determining position of motor vehicle.

Description

The invention relates to a system and a method for the high-precision position determination of a motor vehicle, preferably to its autonomous or automated driving operation. The system and method are also suitable for use in areas where satellite-based systems are not or only partially functional.

Satellite-based global navigation systems (GNSS), most notably the US Global Positioning System (GPS), have been driven by a variety of military and civil positioning and navigation applications as well as card-based services and constant technological advances, as well as cost-effective components many areas of everyday life.

In the case of car navigation systems, a GPS measurement of the position is used at regular intervals, which often allows a location of a few meters to locate the vehicle and provides the basis for route calculation, navigation and map-based services. For the localization of the vehicle, a free route to the satellites is required.

The position determination of a vehicle can take place by a combination of absolute position determination with the aid of GNSS, GPS, Glonass or Galileo supported by an interpolating coupling measurement with the aid of odometers and / or the inertial sensor system; In sum, it is often possible to determine the position to a few meters under the open sky. An odometer for the mechanical measurement of a traveled distance is an incremental measuring device, which consequently has a systematic measurement error, ie a cumulative drift, and is therefore unsuitable for longer-lasting applications if its data are not supported by third-party sensors and good heading information.

The use of inertial sensors, which detect the positional changes by integrating the accelerations twice, leads to significant inaccuracies due to integration errors without regular supportive GPS support after several seconds or minutes.

If high-precision positioning or navigation in the range of a few centimeters is required, such as in the case of control tasks for autonomous or automated vehicles with a high trajectory accuracy requirement, a high-precision differential GPS survey is currently used (so-called RTK-GPS). With RTK (Real Time Kinematic) surveying, points are measured or staked to approximately 1 to 2 cm using satellite-based navigation systems such as GPS or GLONASS or GALILEO. Prerequisite is again an undisturbed reception of the signals from at least five satellites, preferably at least six satellites. The differential GPS, which has an update rate in the second range, is interpolated with the help of low-drift inertial sensors with an update rate of 10 milliseconds.

However, there are areas in which, for example, GPS / GNSS signals due to shadowing, reflections, multiple propagations, etc. are not available in sufficient quality and provide inaccurate results. This must be expected with position inaccuracies in the range of several tens of meters. Therefore, this technology is limited for a variety of locations (tunnels, parking garages, between tall buildings, in valley cuts in mountainous areas and mountains).

Furthermore, a disturbance of the GPS signals or their reception is not excluded, so that the safe operation of sensitive systems, eg. B. autonomous or automated vehicles, not alone on this basis can be guaranteed. For safety-relevant applications with high reliability requirements, in addition to the GPS navigation, a second independent source of information is needed; For this purpose, it is preferable to use a system which (diversified for GPS orientation to distant satellites) is preferably oriented to local earth-related features.

For example, vehicles such as those developed by Google ^™ , which operate autonomously in the United States for several thousand kilometers under real traffic conditions, can use not only (differential) GPS signals but also environmental plans and landmarks for their localization, orientation and trajectory planning. created with the help of a Lidar system and stored in digital maps and recognized while driving with the same Lidar system in combination with various vehicle sensors for environmental detection in real time in the environment. The localization of the vehicle is then carried out by a comparison of the landmarks. However, the use of the high-value, therefore cost-intensive and thus for a wide application unprofitable lidar sensors excludes the use for the automotive mass market.

Starting from this prior art, it is an object of the present invention, a cost-effective, suitable for large-scale use system and an economical and robust method for accurate absolute and / or relative Provide localization or self-positioning in areas and areas without or with insufficiently accurate GPS reception.

This object is achieved by a system having the features of claim 1 and by a method having the features of claim 6. Further developments of the system and the method are set forth in the subclaims.

A first embodiment of a positioning system according to the invention of a motor vehicle comprises a digital map in which data about site-specific features, also called landmarks, are located, ie. h., assigned to a geographical position, are recorded. Furthermore, the position determination system has one or more environment recognition devices, with which the location-specific features or landmarks in the vicinity of the vehicle can be detected. A location module having access to the digital map as well as being coupled to the environment detection devices includes a processing unit configured to match data about the detected location-specific features with the data of landmarks recorded in the digital map, and the vehicle location based on localized in the digital map localized location-specific characteristics to locate. The system further comprises an inertial measurement unit, also coupled to the location module, of the vehicle, which detects the translational movements and the rotational movements of the vehicle in the form of movement data, wherein the location module is configured to determine the vehicle position by means of the vehicle movement data based on the position based on the site-specific features was located.

The vehicle positions, which are determined by landmarks, thus serve as bases for determining the vehicle position by means of the inertial sensor, or the inertial measuring unit. The inaccuracies that arise when using the inertial measurement unit due to integration errors are corrected by the landmark positions used as landmarks. Thus, by the combined use of landmarks for position determination and inertial sensors, a highly accurate position determination of the vehicle without satellite-based navigation is possible.

In general, the environment detection device which can be used for this purpose is not limited to expensive lidar systems; it is also cost-effective to simply use one or more cameras, for example a stereo camera or a multi-camera. Ultimately, however, lidar systems or radar systems can also be used. Also, the combination of several environment recognition devices is conceivable. Common prerequisite is a high spatial and temporal resolution of the environment detection device. For processing the environmental information recorded using the environment recognition device, the environment recognition device itself or else the localization module coupled thereto can comprise a data processing unit. In the case of a camera as environment detection device, for example, this may be an image processing unit.

In a preferred embodiment, the localization module processing unit is configured with a Kalman filter, which is a set of mathematical equations which takes into account the measurement uncertainties of the various incoming measurements in determining vehicle position.

The position determination system may comprise at least one further sensor device, which may be an odometer, a radar system, a laser scanner, an ultrasonic sensor, a lidar system; alternatively or additionally, the system may include a satellite receiver module of the motor vehicle. Both the further sensor device (s) and the satellite receiver module are then coupled to the localization module, respectively.

The digital map used is preferably a high-resolution digital map, such as Google Maps with Street View in question, as well as high-resolution satellite images and exact maps of specific infrastructures and areas, such as tunnels or airports or parking garages, etc. Also, the use of building information of land registries for the digital map with registered landmarks conceivable.

A corresponding method for determining the position of a motor vehicle which uses the position determination system of the invention therefore comprises firstly providing the digital map in which the data about the site-specific features or landmarks are recorded in a localized manner. With the environment recognition device the environment of the vehicle is recorded and with a data processing unit belonging to the environment recognition device or the localization unit, data about the location-specific features are acquired from the environment data. The processing unit of the localization module performs the matching of the acquired data and the data recorded in the digital map via the location-specific features and locating the vehicle position on the basis of the location-specific features recorded in the digital map. After detecting the translational movement and the rotational movement of the vehicle As vehicle movement data by means of the inertial measurement unit, the vehicle position is determined by means of the vehicle movement data based on the position localized on the basis of the location-specific features by the localization module. Thus, a high-precision position determination and high Trajektoriengenauigkeitsanforderungen be ensured by the landmark-based localization provides the bases for the inertial sensor.

With the method, the data stream from which the landmarks are extracted, and the features extracted therefrom, can be processed so economically that they can be robustly tuned while driving and in real time with the digital maps.

In this case, the vehicle position is determined by means of the vehicle movement data based on the position localized on the basis of the location-specific features using the Kalman filter configured in the processing unit of the localization module. The Kalman filter takes into account the measurement uncertainties of the various incoming measured variables when determining the vehicle position. It is applicable when the mathematical structure of the underlying dynamic system and the measurement distortions are known.

The data processing unit performing the data on the location-specific features from the environment data is carried out by directly recognizing the location-specific features from the environment data or by transforming environment feature data into a feature space of the site-specific features recorded in the digital map.

Furthermore, the method may comprise steps such as specifying the determination of the vehicle position by taking into account further detection variables which the localization module or its processing unit receives from further sensor devices, or predicting an approximate vehicle position by means of a GPS receiver, although only partially available, receiving it from the satellite receiver module. Signal to restrict the search space. Thus, the processing unit can be limited to a few landmarks for comparison. Furthermore, such a GPS signal can be used to verify consistency of location of the vehicle position by means of the location-specific landmarks.

The digital map, in which the data on the site-specific features are located localized, can finally be provided by the environment data is detected by means of at least one environment detection device and detection positions of the environment data, extracted from the data on site-specific features and localized with the detection positions in the digital Map to be recorded. Such a card present in the system can be updated, for example automatically or according to user request, if necessary regularly. It is also possible to use a digital map provided on the Internet, which is transmitted to a vehicle as required before driving through a corresponding area via a data transmission channel, for example via mobile radio.

As landmarks are preferably object in question, which can be well detected by the expected driving position of the vehicle from the environment detection device. These are in particular those objects which have also been set up for human orientation and information on the roadside and are essentially immobile. These are in particular traffic signs, city signs, traffic lights, street beacons, but also lampposts, other piles or poles and tree trunks. Furthermore, attached to the road markings are well suited as landmarks, such as. B. lane markings, restricted areas, stop lines and other permanent markings and objects. The latter can also be changes in the road surface, manhole covers, water inlets. Furthermore, come from the road from well identifiable characteristics of buildings as landmarks in question, such. As curbs, edges of traffic islands, surfaces and edges of facades (including windows and doors) and surfaces and edges of walls.

• – Marken mit sich vom Hintergund abhebenden Farben und/oder Formen,
• – Marken mit speziellen Mustern,
• – reflektierende Marken,
• – leuchtende Markierungen.

In addition to these "natural" landmarks that are already present in the environment, artificial markings specially designed for this purpose may facilitate localization in particularly difficult cases (eg in multi-storey car parks). B.

• - trademarks with colors and / or shapes that contrast with the background,
• - brands with special patterns,
• - reflective marks,
• - luminous markings.

In order to enable a clear localization of constantly recurring patterns (eg in tunnels), the artificial markings can be provided with a clearly identifiable pattern, so that the mark can be unmistakably assigned to a position. This can be done for example via a barcode or other known signatures in principle. Orientation based on the landmarks in a digital map requires the landmarks to be retrieved from the localization unit using the data of the Recognized environment detection device. If as a landmark z. B. a street beacon is present in the map, but which is no longer available in reality, would degrade the quality measure for the localization at this point. The method therefore provides that the recognition frequency of a landmark be recorded in the digital map, and this information is taken into account in the assessment of the localization quality. In this way, changes in reality against the update status of the card can be adequately taken into account.

These and other advantages will be set forth by the following description with reference to the accompanying figure.

The reference to the figure in the description is to aid in the description and understanding of the subject matter. The figure is merely a schematic representation of an embodiment of the invention.

The single figure shows a schematic representation of an embodiment of the system according to the invention.

So far, for a high-precision positioning or navigation in the range of a few centimeters, as in the case of control tasks for autonomous or automated vehicles with a high Trajektoriengenauigkeitsanforderung, resorted to a high-precision differential GPS survey (so-called RTK-GPS). With RTK (Real Time Kinematic) surveying, points are measured or staked to approximately 1 to 2 cm using satellite-based navigation systems such as GPS or GLONASS or GALILEO. However, this requires an undisturbed reception of the signals from at least five satellites, preferably at least six satellites.

The system according to the invention for the exact position determination of a vehicle without precise satellite support is shown in the figure, in which a camera KAM for collecting data about the location-specific features in the surroundings of the vehicle is inexpensively provided as environment recognition device. The positioning system relies on a digital map DLK in which data on the site-specific features or landmarks are recorded localized. This may be, for example, a database of landmarks with position reference in a digital map. The localization module F coupled to the digital map DLK and the camera KAM comprises a processing unit for matching the acquired data and the data recorded in the digital map DLK via the site-specific features. The location of the vehicle position is based on the location-specific features recorded in the digital map DLK. An inertial measurement unit IMU of the vehicle that detects the vehicle movement data forms part of the system and is coupled to the location module F whose processing unit is configured to determine the vehicle position using the vehicle movement data based on the location-localized location.

Further variables, detected by further sensor devices α such as odometer, radar system, laser scanner, ultrasonic sensor, lidar system, and / or a satellite signal received by a satellite receiver module GPS of the motor vehicle, can be used to support the system by coupling with the localization module F be used.

The processing result of the localization module F provides a precise and robust position determination of the vehicle, ie information about location x, y, z and position φ, ψ, Ω in the space with accuracy indication σ _xyz .

The fact that GPS / GNSS signals are not required for the exact positioning of the vehicle in the first place, the system can also be used in areas where the GPS / GNSS signals through shadows, reflections, multiple propagations, etc. only limited available or are very imprecise. In tunnels, multi-storey car parks, between tall buildings, in valley cuts in mountainous areas and mountains, the system also precisely determines specific vehicle positions and can therefore also be used for autonomous or automated vehicles.

The position determination in the vehicle takes place by a combination of absolute position determination with the help of the landmarks, supported by an interpolating coupling measurement with the aid of inertial sensors IMU, possibly supported by odometer α. An odometer for mechanically measuring a traveled distance is an incremental measuring device which consequently suffers from a systematic measurement error in the form of a cumulative drift and thus unsuitable for longer-term applications only if its data are not supported by third-party sensors and good heading information. The use of inertial sensors, which detect the position changes by integrating the accelerations twice, leads to significant inaccuracies due to integration errors without the regular supportive landmark determination after several seconds or minutes.

Inertial sensors measure translational and rotational accelerations. By means of three orthogonally arranged acceleration sensors (translation sensors) for detecting the linear acceleration, the translational movement can be calculated. Another three, also orthogonally arranged rotation rate sensors (gyroscopic sensors) permit the calculation of the rotational movement by measuring the angular velocities. The inertial measuring unit IMU comprises these six sensors. To improve the accuracy or to correct the drift of the sensors, for example, magnetic field sensors (compass sensors) can be used.

In order to ensure the system functionality and to ensure continuous even in the absence or limited GPS reception, the positioning method according to the invention is based on a decentralized, locally available, robust, economical and independent of the GPS availability method.

On the one hand, the constant development of new driver assistance systems means that more and more environment recognition devices such as cameras, lidar or radar systems, etc. are being installed in the vehicles. Today available systems allow an accurate mapping of the vehicle environment, which also takes place with a high temporal resolution.

On the other hand, there are high-resolution digital maps enriched with a wealth of information. Examples include Google Maps with Street View and the exact, in digital form available maps of specific infrastructures and areas (eg tunnels, airports, parking garages, etc.), the land use of cadastral offices, high-resolution satellite imagery, etc. It is conceivable the use of a digital map provided on the Internet, the system must have access to the Internet, this can be realized for example by means of the infotainment system of the vehicle.

With the help of mechanical methods for extracting features from a variety of sensor sources such as a (stereo) camera, radar, Lidarsystemen or other sensors for environmental detection, it is possible to site specific features - the so-called landmarks - to extract from the environment data and their (to a few centimeters accurate ) Position in a digital map. These landmarks are then used as bases for accurate navigation, eg. B. in a vehicle used.

Either the same features are recognized with the aid of the environment recognition devices of the vehicle and compared with the digital map material, or there is a transformation of detected with the environment detection devices environment features in the feature space of the deposited in the digital maps landmarks, as well as a corresponding comparison with these.

High precision positioning and high trajectory accuracy requirements are achieved by providing landmark-based localization with inertial sensor bases that provide the required position, velocity and orientation data in high temporal resolution.

The combination of these two systems can be based on a Kalman filter. Furthermore, this system can be extended by adding further input variables from sensor devices α such. As odometer, radar systems, laser scanners, ultrasonic sensors, lidar systems, etc.

The Kalman filter is a set of mathematical equations that allow conclusions to be drawn on the state of a dynamic system in the presence of only erroneous observations, and whose special mathematical structure allows use in real-time systems. Thus, the Kalman filter virtually serves to remove the interference caused by the measuring devices. The Kalman filter is applicable if the mathematical structure of the underlying dynamic system and that of the measurement distortions are known.

An optionally - also limited - available GPS signal can be used to limit the search space for a start of the system, in which the detected by the camera sensors landmarks with those in the database, their localization with the approximate based on the GPS signal Position corresponds, be compared and thus come to a clear result faster. Furthermore, a GPS signal can be used as an independent further source of information with which the consistency of the location can be verified.

Advantageously, the system and method according to the invention enables passive, exact, decentralized georeferencing, which serves as the basis for the necessary drift compensation of the inertial measuring unit.

Another significant advantage consists in the fact that the executed with the existing from the driver assistance systems in the vehicle environment detection sensors system allows use in the automotive mass market, which is a significant difference to the previously used approaches that a variety of non-automotive sensors and Use special systems.

In particular, the system offers the opportunity to use significantly lower cost sensors than both cameras and inertial sensors in comparison to D-GPS systems with inertial sensors on the one hand or laser scanner-based systems on the other hand.

• – Verkehrsschilder,
• – Ortsschilder,
• – Verkehrsampeln,
• – Straßenbaken,
• – Laternenpfähle,
• – sonstige Pfähle oder Masten,
• – Baumstämme
• – Spurmarkierungen,
• – Sperrflächen,
• – Haltelinien,
• – sonstige dauerhafte Markierungen,
• – Änderungen im Straßenbelag,
• – Kanaldeckel,
• – Wassereinläufe
• – Bordsteine,
• – Kanten von Verkehrsinseln,
• – Flächen, Kanten von Fassaden (inkl. Fenstern und Türen),
• – Flächen und Kanten von Mauern
• – Marken mit sich vom Hintergund abhebenden Farben und/oder Formen,
• – Marken mit speziellen Mustern,
• – reflektierende Marken,
• – leuchtende Markierungen

As landmarks are from the road from highly visible, immobile objects or attached on the road markers or objects or from the roadway from well identifiable characteristics of buildings or special attached for this purpose artificial markers are selected such. B.

• - Road Signs,
• - place signs,
• - traffic lights,
• - street beacons,
• - lamp posts,
• - other piles or masts,
• - tree trunks
• - lane markings,
• - restricted areas,
• - stop lines,
• - other permanent markings,
• - changes in the road surface,
• - Manhole cover,
• - water inlets
• - curbs,
• - edges of traffic islands,
• - surfaces, edges of facades (including windows and doors),
• - Surfaces and edges of walls
• - trademarks with colors and / or shapes that contrast with the background,
• - brands with special patterns,
• - reflective marks,
• - luminous markings

The artificial markings are preferably equipped with a clearly identifiable pattern, so that a mark can be unmistakably assigned to a position. In the digital map, a measure of the recognition frequency and recognition quality of the landmarks is included in order to assess the availability of the landmark for future localizations.

Claims (10)

Positioning system suitable for a motor vehicle, comprising - a digital map (DLK) in which data on location-specific features are located localized, - at least one environment detection device (KAM) for detecting the location-specific features in the vicinity of the vehicle, and - a with a localization module (F) coupled to the digital map (DLK) and the environment recognition device (KAM), which comprises a processing unit for matching the acquired data and the data recorded in the digital map (DLK) with the location-specific features and locating the vehicle position according to the method described in US Pat digital local area map (DLK), characterized in that the system comprises an in-vehicle vehicle motion data inertial measurement unit (IMU) coupled to the localization module (F) whose processing unit is configured to position the vehicle by means of Fahrzeugbewegun gsdaten based on the location-specific characteristics localized position to determine. Positioning system according to claim 1, characterized in that in the configuration of the processing unit of the localization module (F) a Kalman filter is included. Position determining system according to claim 1 or 2, characterized in that the system at least one further sensor device (α), in particular an odometer, a radar system, a laser scanner, an ultrasonic sensor, a lidar system, and / or a satellite receiver module (GPS ) of the motor vehicle coupled to the localization module (F). Positioning system according to at least one of claims 1 to 3, characterized in that the environment recognition device (KAM) is a camera, a stereo camera, a multi-camera, a lidar system, and / or a radar system, and / or the environment recognition device ( KAM) or the localization module (F) comprises a data processing unit. Positioning system according to at least one of claims 1 to 3, characterized in that the digital map (DLK) is a high-resolution digital map (DLK) comprising high-resolution, calibrated satellite imagery, exact maps of specific infrastructures and territories, cadastral development information, environmental images localized features. Method for determining the position of a motor vehicle using a system according to at least one of claims 1 to 5, comprising the steps: - providing the digital map (DLK) in which the data about the site-specific features are recorded in a localized manner, Acquiring environmental data of the vehicle with at least one environment recognition device (KAM) and acquiring data on the location-specific features from the environment data by means of a data preparation unit comprised by the environment recognition device or the localization unit, matching the acquired data and that in the digital map DLK) recorded data about the location-specific features and localization of the vehicle position based on the location-specific features localized in the digital map (DLK) by the processing unit of the localization module (F), detecting a translational movement and a rotational movement of the vehicle as vehicle movement data by means of the inertial measurement unit (IMU), - Determining the vehicle position by means of the vehicle movement data based on the localized by the location-specific location position by the processing unit of the localization module (F). The method of claim 6, wherein determining the vehicle position by means of the vehicle movement data is performed based on the location-specific location-localized position by the processing unit of the location module (F) using a Kalman filter. Method according to claim 6 or 7, in which collecting the data about the location-specific features from the environment data by means of the data preparation unit - Detecting the site-specific characteristics directly from the environment data or by - Transforming environment feature data into a feature space of the location-specific features recorded in the digital map (DLK) he follows. Method according to at least one of claims 6 to 8, comprising at least one of the steps: - Specifying the determination of the vehicle position by taking into account further detection variables, the processing unit of the localization module (F) receives from the at least one further sensor device (α), and / or Predetermining an approximate vehicle position by means of a GPS signal received from the satellite receiver module (GPS), and / or By means of a GPS signal, verifying a consistency of the location of the vehicle by means of the location-specific landmarks. Method according to at least one of claims 6 to 9, in which providing the digital map (DLK) in which the data of the site-specific features are recorded in a localized manner, comprising the steps of: Acquiring environmental data by means of at least one environment recognition device (CAM) and acquisition positions, Extracting site-specific features from the environment data, - Locate the location-specific features localized according to the detection positions in the digital map (DLK).

Images