DE102011013755B4 - Method and device for alphanumeric voice input in motor vehicles - Google Patents

Abstract

Method for the verbal input of a data sequence in the form of data blocks into an input and recognition device of a motor vehicle, wherein recognized data blocks are combined into a current data sequence and the current data sequence is shown in a display of the input and recognition device, characterized in that
the data block last entered and recognized is displayed in a first format for verification and, if necessary, correction and, after verification, in a second format,
a linguistic correction of the current data sequence causes the last recognized data block to be deleted, with the previous data block being converted back to the first formatting for renewed verification after the last recognized data block has been deleted
a haptic correction of the recognition result causes the last character of the current data sequence to be deleted and transfers the current data sequence to the second formatting, the haptic correction having a higher priority than the linguistic correction.

Description

The invention relates to a method and a device for alphanumeric voice input in motor vehicles.

Due to the increasing number of comfort components in motor vehicles that can be operated and adjusted via multifunction control devices, it is becoming more and more difficult for the vehicle driver to have all operating options and their setting modalities at hand. Furthermore, access to the required commands or the input of necessary information, for example in the case of infotainment systems with navigation and / or telephone facilities, can be relatively cumbersome and thus time-consuming. In particular, when entering information such as telephone numbers while driving, this can lead to a distraction for the driver.

To minimize these problems, voice input devices are increasingly being developed and used so that the driver does not have to take his hands off the steering wheel to enter commands and data. This type of command and data entry often leads to other problems in practice. With many conventional components with voice control, for example, there is a fixed set of commands that the user basically has to learn by heart. Therefore, the problem often arises while driving that the driver has forgotten the necessary command and is not able to look it up in the manual.

From the DE 100 28 869 A1 a method for supporting command and data input is known in which a user is shown at least one input menu on a display unit in which possible command and menu fields of the currently permissible voice command and voice inputs are optically highlighted. The visual highlighting can be achieved through color design, the use of different brightness and contrast levels or other measures. However, the procedure is based on a fixed set of commands.

From the DE 199 42 871 A1 a method for operating a voice-controlled command input unit in a motor vehicle is known in which the operability is acoustically menu-driven by first voice-controlled the command level that is to be activated before device- or function-related command input, whereby so-called shortcut commands are used for better operation with which you can switch between device-specific menus.

The pamphlet DE 10 2007 052 345 A1 describes a method and a device for operating a device of a vehicle with a voice control, in which an operating step is selected and a function assigned to the operating step is carried out. In this case, by selecting the operating step, voice control is automatically activated for an input following the operating step, the input being able to be made verbally or haptically.

The pamphlet EP 0 389 514 B1 discloses an input method for data sequences by means of voice input into a speech recognition system, a data sequence being understood here to be a number block consisting of a number of digits. After a spoken data sequence, the speech recognition generates a pause signal and the recognized data sequence is repeated. Furthermore, the recognized data sequence is shown visually on a display. The user can then enter the next data sequence. If the user recognizes a recognition error on the part of speech recognition either visually or on the basis of the verbal confirmation, then the verbal input of a corresponding keyword, such as “delete”, deletes the data sequence last entered. A mixture of verbal and haptic input is not intended.

• - Erkennen von Schlüsselworten, die durch einen Benutzer deutlich gesprochen werden, und in einer oder mehreren Sprachäußerungskette(n) mit einer variablen Anzahl von Schlüsselworten enthalten sind, wobei die Anzahl von deutlich gesprochenen Schlüsselworten bei Sprachäußerungsketten durch den Benutzer während des Aussprechens dieser Schlüsselworte bestimmt wird und jede Sprachäußerungskette gefolgt ist von einem Pausenzeitraum;
• - Schaffen eines Pausensignals, nachdem eine Sprachäußerungskette für die Dauer eines vorbestimmten Pausenzeitraums aufgehört hat;
• - Schaffen einer Anzeige von bestimmten erkannten Schlüsselworten für den Benutzer in Reaktion auf das Pausensignal, welche Anzeige der vor dem Pausensignal aufgetretenen Sprachäußerungskette entspricht; und
• - Korrigieren bestimmter Schlüsselworte in Reaktion auf das Pausensignal und in Reaktion auf das Erkennen eines vorbestimmten Fehlerbefehl-Schlüsselwortes, das in einer nach dem Pausensignal auftretenden Sprachäußerungskette enthalten ist.

The pamphlet DE 38 55 164 T2 relates to a user-interactive method for controlling an electronic device, with the steps:

• Recognition of key words which are clearly spoken by a user and which are contained in one or more speech utterance chain (s) with a variable number of key words, the number of clearly spoken key words in speech utterance chains being determined by the user during the utterance of these key words and each utterance chain is followed by a pause period;
• - providing a pause signal after a chain of utterances has ceased for the duration of a predetermined pause period;
• Providing the user with an indication of certain recognized key words in response to the pause signal, which indication corresponds to the speech utterance chain that occurred before the pause signal; and
• Correcting certain key words in response to the pause signal and in response to the recognition of a predetermined error command key word which is contained in a speech utterance chain occurring after the pause signal.

The pamphlet EP 1 136 984 A2 relates to a method for voice control of an electrical Device, wherein information to be entered is entered by spelling, in which a recognized letter or a recognized character or a recognized letter or character sequence is output to acknowledge the letter or character input. The acknowledgment enables the user to immediately verify the voice input and thus, if necessary, to correct an input during the input process. In the event that error corrections are required, this results in a considerable acceleration of speech input processes.

The pamphlet DE 100 28 869 A1 relates to a device and a method for supporting the input of commands and data in motor vehicles by means of at least one display unit and a control device which controls the display unit and to which at least one command input device is assigned, wherein the control device uses the display unit to display to a user at least one input menu in which in principle Command and data inputs are represented by menu fields, the menu fields of the currently permissible voice command and / or voice data inputs or display areas in which the menu fields are arranged being highlighted.

The invention is therefore based on the object of creating an input method for the haptic and verbal input of data sequences with improved verification of the input data sequences.

The object is achieved by a method for inputting data sequences into an input and recognition device of a motor vehicle with the features of claim 1 and by a corresponding device with the features of claim 6. Preferred embodiments of the invention are the subject of the subclaims.

• - eine sprachliche Korrektur der aktuellen Datensequenz ein Löschen des zuletzt erkannten Datenblocks bewirkt, wobei nach dem Löschen des zuletzt erkannten Datenblocks der vorherige Datenblock zur erneuten Verifikation wieder in die erste Formatierung überführt wird,
• - eine haptische Korrektur des Erkennergebnisses das Löschen des letzten Zeichens der aktuellen Datensequenz bewirkt und die aktuelle Datensequenz in die zweite Formatierung überführt, wobei die haptische Korrekturvorrichtung (4) eine größere Priorität als die sprachliche Korrekturvorrichtung aufweist.

In the method according to the invention for the verbal input of a data sequence as a sequence of data blocks into an input and recognition device of a motor vehicle, with recognized data blocks being shown as the current data sequence in a display of the input and recognition device, the last input and recognized data block is used for verification and possibly Correction shown in a first format and after verification in a second format, wherein

• - a linguistic correction of the current data sequence causes the last recognized data block to be deleted, with the previous data block being converted back to the first formatting for renewed verification after the last recognized data block has been deleted,
• - a haptic correction of the recognition result causes the deletion of the last character of the current data sequence and converts the current data sequence into the second formatting, the haptic correction device (4) having a higher priority than the linguistic correction device.

By displaying the last verbally entered data block in different formats depending on the verification status, the user is shown clearly more perceptibly and more quickly what has been recognized and what is to be confirmed by him. This leads to a relief and thus an increase in security.

A linguistic correction of the current data sequence causes the last recognized data block, i.e. the current data block, to be deleted. If the user has recognized that the recognition of the current data block shown in the first formatting was incorrect, a linguistic correction, for example by entering a correction command, causes the current data block to be deleted.

Furthermore, after the last recognized current data block has been deleted, the previous data block is transferred back to the first formatting for renewed verification. In this way, a linguistic correction of the previous data blocks is also possible.

A haptic correction of the current data sequence including the current data block represented in the first formatting causes the last character to be deleted and the entire data sequence is transferred to the second formatting. The correction of the last character is thus understood as a transition from the verbal input to haptic operation, with a further haptic correction of the new last character being possible in the usual way. However, it is also possible to continue the verbal input by speaking another data block.

The haptic correction has a higher priority than the linguistic correction. In other words, a haptic correction has priority over a linguistic correction and the device changes to the haptic operating mode.

A data block preferably has at least one alphanumeric character, alphanumeric characters being understood to mean letters and digits. In other words, a data block comprises one or more characters, it being possible for the number of characters per data block to be different. In particular, a data block can therefore be a data sequence from a sequence of digits exist, as is the case, for example, when entering a telephone number by voice. The sequence of data blocks of different lengths then results in the overall summary of the data sequence, the current data sequence being understood to mean the summary of the input data blocks including the current data block.

A verbal verification of the current data block is preferably carried out by entering a further data block or by verbally entering a verification command. If the user has recognized that the data block entered last was correctly recognized, the input of the next data block automatically results in the verification of the still current data block. It is also possible to carry out the verification with a separate voice command.

The second formatting is preferably carried out by means of bold display, italic display, underlining, selection of a different font with regard to the last recognized data sequence and / or the insertion of special characters before the last recognized data sequence. This conspicuous first formatting enables simple identification and verification of the current data block.

• - eine Einrichtung zur sprachlichen Eingabe und Erkennung von Datenblöcken,
• - eine Einrichtung zur Verifikation und Korrektur der sprachlich eingegebenen Datenblöcke,
• - eine Anzeigeeinrichtung zur Anzeige der aktuellen Datensequenz, und
• - eine Formatierungseinrichtung zur Formatierung der sprachlich eingegebenen und erkannten Datenblöcke in mindestens zwei unterschiedliche Formate.

The device according to the invention for carrying out the method described above for the verbal input of a data sequence in the form of data blocks into an input and recognition device of a motor vehicle comprises:

• - a device for linguistic input and recognition of data blocks,
• - a device for the verification and correction of the linguistically entered data blocks,
• a display device for displaying the current data sequence, and
• - A formatting device for formatting the linguistically entered and recognized data blocks in at least two different formats.

The device preferably further comprises an input device for haptic input of the data sequence and a device for correcting the haptic input.

• 1 einen Freitext-Speller für Telefonnummern in schematischer Darstellung,
• 2 den Freitext-Speller nach Eingabe eines ersten Ziffernblocks,
• 3 den Freitext-Speller nach der Eingabe eines zweiten Ziffernblocks, und
• 4 den Freitext-Speller nach der Korrektur der zweiten Eingabe.

A preferred embodiment of the invention is described below with reference to the drawings. It shows

• 1 a free text speller for telephone numbers in a schematic representation,
• 2 the free text speller after entering the first number block,
• 3 the free text speller after entering a second number block, and
• 4th the free text speller after correcting the second input.

The 1 to 4th explain the method according to the invention and the corresponding device using a so-called free text speller, whose characteristic is the fact that the user is always offered all possible input modes, ie haptic and linguistic input options. In the following, digits are regarded as alphanumeric characters, as is the case, for example, when entering a telephone number.

1 FIG. 11 shows a representation of a menu for a free text speller on a touch-sensitive screen 1 , in other words a touch screen. A display block is shown 2 to display a current data sequence, here the digits "1234", a numeric keypad 3 with the digits 0 to 9 for haptic input, a correction key 4th as well as a verification button 5 . There is also a button 6th intended to close the current menu, ie an actuation of the button 6th usually takes the menu back to the next higher menu level.

Due to the different modalities, namely haptic on the one hand and linguistic on the other, different correction mechanisms are used. For example, while a haptic "correction" mostly refers to the deletion of the last character, which is effected here by the "DEL" key 4, a linguistic "correction" can be used to delete several previously entered characters, for example the current data block, to lead.

Since the haptic-visual representations, formatting and verification mechanisms for a purely haptic operation are coherent and sufficiently known, only the linguistic representation and verification will be discussed here, as well as - taking into account a fully multimodal system - the transition from linguistic operation towards a haptic operation and back.

A verbal input of digits is usually made in the form of n data blocks, in which each of the individual data blocks can in turn contain 1 to k digits. This makes it possible to enter a sequence of digits in the form of whole numbers or as a sequence of individual digits. The standard case will generally be found in the middle area - made up of a sequence of several smaller blocks.

Assuming that the user has not yet entered a sequence of digits haptically or verbally and now begins to verbally enter the number block “1234”, this results in 2 example of the visualization shown.

The user says "1234". The system replies with the recognized sequence of digits, for example “1234”, and places the recognized result in the display block 2 represent.

However, since the result was "written" into the speller by voice control, a different type of representation is selected for the verbal input. The formatting of the recognized data block "1234" shown here is done in bold font with an underscore. This result can now be corrected by the user by means of a suitable dialog, for example the commands “Correct”, “Delete”, etc.

However, if the user confirms the input implicitly, for example by naming a data block with additional single digits or explicitly, i.e. by means of a corresponding command, the display form will now be converted to the standard formatting of the haptic-visual input, which is shown here by normal writing without an underscore . Ultimately, the user has verified and confirmed the previous entry.

In our example, the user now implicitly continues the dialog by naming additional single digits, as shown in 3 is shown. The sequence of digits mentioned is "56", but this time the system recognizes "567". This recognized data block is now also initially displayed in the language-specific representation, namely in the formatting "bold" and "underlined". The first data block “1234”, which is implicitly verified by the continuation of the dialog, is displayed in the haptic-visual formatting, ie the second formatting.

The user now has several options for correcting this recognition result, which is displayed as the current data sequence. A verbal "correction" would, for example, delete the last recognized number block "567" and - the user is back in voice control - the previously entered number sequence "1234" could be marked again in a language-specific manner, namely "bold" and "underlined ". For a haptic correction by pressing the button 4th with the designation "DEL" the last digit of "567" would be deleted.

What is decisive here is that when a haptic correction is detected, for example by means of a proximity sensor system or detection of the pressure of the correction button 4th , the voice control-specific formatting is transferred to the haptic visual formatting, as the user no longer continues the dialog under voice control. In addition, the speech dialogue is even paused or ended in the case of haptic operation.

4th shows that in the example the user enters the last digit by pressing the correct key 4th has deleted. Since this sequence of digits has now also been verified, which is shown by the representation of the sequence “123456” in the second formatting, the formatting can also be started when the voice dialog is started again, for example by resuming the voice dialog after a previous pause, by pressing a PTT button (push-to-talk, not shown), naming a corresponding command "Continue", etc., are retained.

Optionally, the voice control-specific formatting (bold) of the last entry - this would then only concern the digit sequence "56" - but can also be resumed if the voice dialog is started again or no further haptic control, for example by a proximity sensor, is recognized.

This interplay of the various formatting, which can be supported by animations if necessary, is now continued for as long as the user also enters digits if the input modality is used differently.

• - kursiv: „1234567“
• - unterstrichen: „1234567”
• - Einfügen von Sonderzeichen (zum Beispiel Leerzeichen): „1234 567“
• - andere Schriftart

und verschiedene Kombinationen aus den jeweiligen Formatierungen:

• - Kursiv und fett: „1234567'
• - mit Leerzeichen und fett: „1234 567“
• - mit Leerzeichen, kursiv und fett: „1234 567“

Alternative formatting of the recognized sequence of digits, compared to the "bold" already shown, would be, for example:

• - italic: "1234567"
• - underlined: "1234567"
• - Insertion of special characters (for example spaces): "1234 567"
• - different font

and various combinations of the respective formatting:

• - Italic and bold: "1234567"
• - with spaces and bold: "1234 567"
• - with spaces, italics and bold: "1234 567"

List of reference symbols

1

screen

2

display

3

Numeric keys

4th

Delete button

5

Confirmation button

6th

"Close menu" button

Claims (6)

Method for the verbal input of a data sequence in the form of data blocks into an input and recognition device of a motor vehicle, with recognized data blocks being combined into a current data sequence and the current data sequence being shown in a display of the input and recognition device, characterized in that the last input and a recognized data block for verification and, if necessary, correction is displayed in a first format and after the verification in a second format, a linguistic correction of the current data sequence causes the last recognized data block to be deleted, the previous data block being repeated after the last recognized data block has been deleted Verification is transferred back to the first formatting, a haptic correction of the recognition result causes the last character of the current data sequence to be deleted and the current data sequence is transferred to the second formatting the haptic correction has a higher priority than the linguistic correction. Procedure according to Claim 1 , characterized in that a data block has at least one alphanumeric character. Method according to one of the Claims 1 or 2 , characterized in that a verbal verification of the current data block takes place by entering a further data block or by verbally entering a verification command. Method according to one of the preceding claims, characterized in that the first formatting takes place by means of a bold display, italic display, underlining, selection of a different font with regard to the last recognized data block and / or insertion of special characters before the last recognized data block. Device for carrying out the method for the verbal input of a data sequence in the form of data blocks in an input and recognition device of a motor vehicle according to one of the preceding claims, with a device for verbal input and recognition of data blocks, a device for the verification and correction of the linguistically entered data blocks, a display device (2) for displaying the current data sequence, and a formatting device for formatting the linguistically entered and recognized data blocks in at least two different formats. Device according to Claim 5 , characterized in that the device has an input device (3) for haptic input of the data sequence, and a device (4) for correcting the haptic input.

Images