Image- or voice-guided destination input on a vehicle navigation system

Abstract

The invention relates to a system for specifying a destination to a navigation system in a vehicle, comprising an input interface (1) and a computing unit (3), wherein the computing unit (3) is configured to analyze digital image information or image-descriptive speech information received at the input interface (1) using at least one machine learning model and to assign a destination based on an implicit or explicit database comparison and to transmit the assigned destination to the navigation system.

Description

[0001] The invention relates to a system for a vehicle for specifying a destination for a navigation system, and to a vehicle with such a system.

[0002] The conventional method of entering destinations into vehicle navigation systems, such as those used in cars, typically requires manually entering an address or selecting a saved destination. This can be time-consuming and cumbersome, especially in unfamiliar cities or locations that are difficult to describe, for example, if the exact address of the desired destination is unknown. Furthermore, there are situations where the user doesn't know the exact address but has a general idea of ​​the place they want to reach.

[0003] DE 10 2012 219 329 A1 relates to a method for input with a pen, in particular on a satellite navigation device, comprising: receiving a first signal from a first electrode of the pen, wherein the pen is configured to communicate wirelessly with a device via a touch sensor of the device; receiving a second signal from a second electrode of the pen, wherein the second signal is a reference signal and the second electrode is a reference electrode; and comparing the first signal with the second signal.

[0004] To increase the user-friendliness and efficiency of navigation, there is a need for a system that makes it possible to simply set the destination for the navigation system by showing an image of a place.

[0005] DE 10 2015 003 769 A1 relates to a navigation device with an input device for command input for a vehicle, wherein the input device comprises at least one image capture unit arranged in a vehicle interior for capturing text information and / or image information in a capture area and an evaluation unit coupled to the image capture unit for analyzing and evaluating the text information and / or image information, wherein the evaluation unit is coupled to a control unit for carrying out an automatic input of target data determined on the basis of the text information and / or image information.From DE 10 2015 003 769 A1 it is further known in particular that a user transmits handwritten, printed or displayed text information and / or image information, which includes target data for route planning of the navigation device, to a camera located in a vehicle interior, which is permanently installed in the vehicle interior or can be installed as a mobile device within it.

[0006] The object of the invention is to provide a system to an occupant of a vehicle which enables a simplified specification of a destination for a navigation system.

[0007] The invention is defined by the features of the independent claims. Advantageous further developments and embodiments are the subject of the dependent claims.

[0008] A first aspect of the invention relates to a system for a vehicle for specifying a destination for a navigation system, comprising an input interface and a computing unit, wherein the computing unit is configured to analyze digital image information or image-descriptive speech information received at the input interface using at least one machine learning model and to assign a destination based on an implicit or explicit database comparison and to transmit the assigned destination to the navigation system.

[0009] The system allows users to transmit a digital image file or video stream to the input interface instead of a conventional text-based or voice-based input of a target name, such as an address. Alternatively, it can transmit speech information that describes an image of the desired target in natural language. This speech information does not refer to a voice input of a target address, but rather to naturally descriptive language that reflects the target's characteristics, which are generally recognizable in an image. For example, a building can be described with its respective properties, such as its height, color, shape, etc.

[0010] It is sufficient for the user to transmit a digital image file or video data stream to the input interface, for example wirelessly via Bluetooth, WLAN, UWB, mobile networks such as those using the 4G / 5G / 6G standard, or USB, LAN, or similar. Alternatively, the user can describe their digital image file or video data stream to the input interface using natural language, as described above, and the microphone will capture the natural language.

[0011] In both cases, the information received at the input interface is forwarded to the processing unit, which uses it as input for a machine learning model. This model preferably comprises an artificial neural network that processes the received information and compares it explicitly or implicitly with a database. In the explicit case, the input can include a digital image as well as images from the database, in which case the machine learning model performs a similarity analysis by evaluating the natural image data. However, the machine learning model can also extract a set of features from the received digital image and compare them with abstracted information stored in the database, such as hash tables.

[0012] The latter approach is particularly suitable for obtaining language information that cannot be directly compared with images from a database; rather, at least one machine learning model serves to make these multimodal inputs comparable, especially through abstract feature extraction, which does not necessarily yield features interpretable by a human, but provides machine-comparable results. This corresponds to implicit matching.

[0013] If a suitable match is found in the database, either implicitly or explicitly, it is suggested as the destination and the coordinates assigned to the destination in the database are used for route calculation.

[0014] This allows a user in a vehicle to specify a route to a destination via a navigation system, even if the address or name of the destination is unknown. This is particularly useful when driving to an unfamiliar city with well-known landmarks that can be described by a picture, but whose address or official name is unknown.

[0015] According to an advantageous embodiment, the input interface is a data interface designed to receive digital image files.

[0016] According to a further advantageous embodiment, the database is an image database, wherein the machine learning model is designed to compare features of images from the image database with features from the image information and / or language information, which are each obtained at the input interface, and to determine a most probable match.

[0017] According to another advantageous embodiment, the database comprises abstracted features that are stored in the database as information that is not interpretable by a human.

[0018] According to a further advantageous embodiment, the computing unit is designed to receive feedback from a user regarding the assignment of the computing unit's image information and / or speech information to a destination for the navigation system and, if necessary, to adjust the assignment based on the feedback.

[0019] According to another advantageous embodiment, the computing unit is designed to automatically adjust parameters of the machine learning model through reinforcement learning depending on user feedback.

[0020] According to another advantageous embodiment, the computing unit is designed to analyze and semantically interpret the user's language information using a Large Language Model.

[0021] According to a further advantageous embodiment, the computing unit is designed to extend the machine learning model and / or the database to include image information and / or speech information about a target that has no corresponding entry in the database.

[0022] The system continuously learns from user input and feedback to improve its recognition accuracy. Each time a new image is used and the destination is confirmed, the system expands its database with additional information to increase accuracy for future queries. The system can also recognize and consider user preferences by, for example, storing preferred routes or frequently visited destinations.

[0023] According to another advantageous embodiment, the database is a central database located outside the vehicle, with the computing unit designed to access the database via an API of the database.

[0024] Another aspect of the invention relates to a vehicle with a system as described above and below.

[0025] Advantages and preferred further developments of the proposed vehicle result from an analogous and substantive transfer of the above statements made in connection with the proposed system.

[0026] Further advantages, features and details will become apparent from the following description, in which - possibly with reference to the drawing - at least one embodiment is described in detail.

[0027] It shows: Fig. 1: A system for specifying a destination for a navigation system according to an embodiment of the invention.

[0028] The representations in the figure are schematic and not to scale.

[0029] Fig.Figure 1 shows a vehicle system for specifying a destination to a navigation system, comprising an input interface 1 and a processing unit 3. The processing unit 3 is configured to receive a smartphone at the input interface 1 and analyze the resulting digital image information using at least one machine learning model. Based on an implicit or explicit database comparison, the processing unit assigns a destination and transmits the assigned destination to the navigation system. The processing unit 3 analyzes the visual features of the digital image, such as building shapes, road layouts, signs, and other characteristic features. By employing deep learning algorithms, the processing unit 3 can interpret complex visual data and recognize patterns to accurately identify the location within the image.Furthermore, a comparison with external databases is performed: Processing unit 3 accesses extensive image databases or similar services to compare the analyzed image with existing street views. By comparing the visual characteristics of the displayed image with the databases, processing unit 3 can determine the exact location where the image was taken. Once the location has been identified, processing unit 3 displays it to the driver as a possible navigation destination suggestion. The driver can confirm the destination or add further details (e.g., exact address or waypoints). After confirmation, the navigation system automatically guides the driver to the recognized destination. All necessary route information is displayed on the vehicle's display, and navigation proceeds as usual.

[0030] Although the invention has been further illustrated and explained in detail by means of preferred embodiments, the invention is not limited by the disclosed examples, and other variations can be derived from them by a person skilled in the art without departing from the scope of protection of the invention. It is therefore clear that a multitude of possible variations exist. It is also clear that the embodiments mentioned as examples are truly only examples and are not to be understood in any way as limiting, for example, the scope of protection, the possible applications, or the configuration of the invention.Rather, the preceding description and the description of the figures enable the person skilled in the art to implement the exemplary embodiments in concrete terms, whereby the person skilled in the art, with knowledge of the disclosed inventive concept, can make various changes, for example with regard to the function or the arrangement of individual elements mentioned in an exemplary embodiment, without leaving the scope of protection defined by the claims and their legal equivalents, such as further explanations in the description. QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0000] DE 10 2012 219 329 A1

[0003] DE 10 2015 003 769 A1

[0005]

Claims

[1] System for a vehicle for specifying a destination for a navigation system, comprising an input interface (1) and a computing unit (3), wherein the computing unit (3) is configured to analyze digital image information received at the input interface (1) or image-descriptive speech information received at the input interface (1) using at least one machine learning model and to assign a destination based on an implicit or explicit database comparison and to transmit the assigned destination to the navigation system. [2] System according to claim 1, wherein the input interface (1) is a data interface designed to receive digital image files. [3] System according to one of the preceding claims, wherein the database is an image database, wherein the machine learning model is designed to compare features of images from the image database with features from the image information and / or language information, which are each obtained at the input interface (1), and to determine a most probable match. [4] System according to one of the preceding claims, wherein the database comprises abstracted features which are stored in the database as information that is not interpretable by a human. [5] System according to one of the preceding claims, wherein the computing unit (3) is configured to receive feedback from a user about the assignment of the computing unit (3) of the image information and / or speech information to a destination for the navigation system and, if necessary, to adjust the assignment based on the feedback. [6] System according to claim 5, wherein the computing unit (3) is configured to automatically adjust parameters of the machine learning model by reinforcement learning depending on the user's feedback. [7] System according to any of the preceding claims, wherein the computing unit (3) is configured to analyze and semantically interpret the user's language information using a Large Language Model. [8] System according to one of the preceding claims, wherein the computing unit (3) is configured to extend the machine learning model and / or the database by the image information and / or speech information when image information and / or speech information is provided about a target which has no equivalent in the database. [9] System according to any of the preceding claims, wherein the database is a central database outside the vehicle, wherein the computing unit (3) is configured to access the database via an API of the database. [10] Vehicle with a system according to any of the preceding claims.

Images