WO2015167077A1

WO2015167077A1 - Navigation apparatus and control method therefor

Info

Publication number: WO2015167077A1
Application number: PCT/KR2014/007072
Authority: WO
Inventors: 최재영
Original assignee: 현대엠엔소프트 주식회사
Priority date: 2014-04-30
Filing date: 2014-07-31
Publication date: 2015-11-05
Also published as: KR20150125798A; KR102128030B1

Abstract

The present invention relates to a navigation apparatus comprising: a voice input unit for receiving an input of voice from a user; and a control unit for recognizing the received voice, extracting candidates according to a result of the voice recognition, determining whether the number of extracted candidates is greater than or equal to a predetermined number, and immediately searching for a path using the extracted candidates when the number of the extracted candidates is less than the predetermined number.

Description

Navigation device and its control method

The present invention relates to a navigation device and a control method thereof, and more specifically, to a navigation device supporting a path search function using voice recognition, a voice recognition result by applying a weight and a threshold value using the input frequency by time and day of the search target. The present invention relates to a navigation apparatus and a method for controlling the same, which can improve path search performance using speech recognition.

In general, the navigation device maps its location measured using GPS (GPS) to map data, receives a waypoint and / or a destination from a user, and searches for a route to the destination, and searching for the search. It is a device for displaying the route in the map data and performing a road guide along the route.

A method of retrieving map data of a waypoint and / or a destination input by a user in the navigation device, comprising: a name retrieval method and an address for retrieving a name by inputting a name of a representative building or a subway station of a corresponding destination; There is an address search method.

Currently, most navigation devices have a problem of inputting a key (button) displayed on a touch screen screen several times to several tens of times in order to input the name or address. Accordingly, in order to improve the above-mentioned inconvenience, a navigation device for performing a route search using a voice has been recently developed.

However, the speech recognition engine applied to the navigation device using the speech recognition is difficult to maintain a constant performance because the speech recognition rate is different according to the type, and also when applying the expensive speech recognition engine with high speech recognition rate to improve the performance There is a problem that the unit cost of the navigation device increases. Therefore, even in the case of using a low-cost speech recognition engine with low speech recognition rate, it is necessary to develop a navigation device that can improve path search performance.

Background art of the present invention is disclosed in Republic of Korea Patent Registration 10-1063607 (2011.09.01. Registered, navigation system having a name search function using voice recognition and its method).

The present invention was created to solve the above problems, and corrects the voice recognition result by applying a weight and a threshold value using the input frequency by hour and day of the search target in a navigation device that supports a path search function using voice recognition. Accordingly, an object of the present invention is to provide a navigation device and a control method thereof, which can improve path search performance using speech recognition.

According to an aspect of the present invention, the present invention provides a voice input unit for receiving a voice from a user; And extracting candidates according to the voice recognition result by recognizing the received voice, and determining whether the extracted candidates are greater than or equal to a predetermined number, and immediately after the extracted candidates are less than the predetermined number, the extracted candidates. It provides a navigation device comprising a; control unit for searching for a route using.

In the present invention, if the extracted candidates are equal to or greater than a predetermined number, the controller corrects the speech recognition result by applying a preset weight to the extracted candidates according to the speech recognition result, and converts the ranks of the candidates into scores. It is characterized by changing according to.

In the present invention, if the extracted candidates are equal to or greater than a preset number, the controller corrects the speech recognition result by applying a threshold to the extracted candidates according to the weighted speech recognition result and presets the number of candidates. It is characterized by compressing the number below.

In the present invention, if the number of candidates compressed by the threshold is one, the controller performs a path search for the corresponding one candidate immediately.

In the present invention, the control unit, to give a higher weighted score to the candidate with a high frequency of the user's search word input according to at least one of time, daily, day, month, season, holiday, holiday, event day. It features.

According to another aspect of the present invention, the present invention provides a control method of a navigation device capable of speech recognition, comprising: a control unit recognizes a voice input from a user and extracts a candidate according to the speech recognition result; Determining whether the extracted candidate is greater than or equal to a predetermined number; And if the number of the extracted candidates is less than the predetermined number, immediately searching for a path using the extracted candidates, and performing road guidance of the searched paths.

In the present invention, the candidate is a candidate having a high possibility of matching a name that the user wants to input for the route search in the navigation device.

The control method may further include correcting the speech recognition result by applying a predetermined weight to the extracted candidates according to the speech recognition result when the extracted candidates are equal to or greater than a preset number. The ranking of the candidates is characterized by changing according to the conversion score.

In the step of correcting the voice recognition result by applying the weight, the control unit, the user's search frequency input frequency is high according to at least one of time, day, day, month, season, holiday, holiday, event day Higher weight scores for the candidates.

The present invention may further include correcting the speech recognition result by applying a threshold to the speech recognition result to which the weight is applied, if the extracted candidate is equal to or greater than a preset number, and setting the number of candidates to be less than or equal to the preset number. It is characterized by compressing.

In the applying of the threshold value, the threshold value may be a lower limit score from a score converted by applying a weight for each candidate.

The present invention improves the path search performance using voice recognition by correcting the voice recognition result by applying weights and thresholds using the input frequency by hour and day of the search target in a navigation device that supports the path search function using voice recognition. It is possible to maintain constant path search performance without being greatly affected by the speech recognition rate of the recognition engine.

1 is an exemplary view showing a schematic configuration of a navigation device according to an embodiment of the present invention.

2 is a flowchart illustrating a control method of a navigation device according to an embodiment of the present invention.

3 is an exemplary view showing a screen displaying a result of applying weights to candidates extracted according to a voice recognition result of a navigation device according to an embodiment of the present invention.

FIG. 4 illustrates an example of a screen displaying a result of applying a threshold to candidates extracted by applying a weight to a voice recognition result in FIG. 3.

Hereinafter, an embodiment of a navigation device and a control method thereof according to the present invention will be described with reference to the accompanying drawings.

In this process, the thickness of the lines or the size of the components shown in the drawings may be exaggerated for clarity and convenience of description. In addition, terms to be described below are terms defined in consideration of functions in the present invention, which may vary according to the intention or convention of a user or an operator. Therefore, the definitions of these terms should be made based on the contents throughout the specification.

As shown in FIG. 1, the navigation device according to the present embodiment includes a GPS receiver 110, a voice inputter 120, a controller 130, a voice recognition data storage 140, and a map data storage 150. And guide information output unit 160.

The GPS receiver 110 receives current location information of the navigation device by using a Global Positioning System (GPS) module (not shown).

The voice input unit 120 receives a user's voice through a microphone mounted on a navigation device, or a microphone mounted on an external device (eg, a mobile phone, a handsfree, a remote controller, etc.) connected to the navigation device via a wired / wireless connection. Receive input.

The controller 130 recognizes a user's voice input through the voice input unit 120 and searches for a route by setting the voice recognized target area as a waypoint or a destination.

In addition, the controller 130 internally performs a separate module (eg, a software module) that performs functions of the voice signal processor 131, the voice recognizer 132, the best candidate selector 133, and the path searcher 134. ) May be included. However, in the present embodiment, the functions of the modules 131 to 134 are described separately for convenience, but the control unit 130 may also integrate the functions of the modules 131 to 134.

The voice signal processing unit 131 converts the voice received through the voice input unit 120 into a digital signal and processes the voice signal into a voice signal capable of performing voice recognition.

The voice recognition unit 132 analyzes the voice signal processed by the voice signal processing unit 131 and recognizes the voice. For example, the voice recognition unit 132 may recognize a voice input by a user in a phoneme (single syllable) unit, a word unit, or a sentence unit.

In this case, the voice signal processing unit 131 and the voice recognition unit 132 may be replaced by a separate voice recognition processor (not shown) or a voice recognition engine (not shown).

The best candidate selector 133 is a preset number of a plurality of candidates that can be input as a region name or a target name in the navigation apparatus among the information recognized by the speech recognizer 132 (eg, a search term for a route search). (E.g., six) best candidates are selected (see Fig. 3A).

In this case, the order in which the candidates selected as the best candidates are arranged may be determined by the voice recognition rate or the voice recognition algorithm by the voice signal processing unit 131 and the voice recognition unit 132 as shown in FIG. It may vary.

In this embodiment, the voice signal processing unit 131 may be given a higher priority to the voice recognition data (ie, weighted voice recognition data) previously stored in the voice recognition data storage 140. And the best candidates are selected in a predetermined order irrespective of the speech recognition rate or the speech recognition algorithm by the speech recognition unit 132 (see FIG. 3B).

That is, in the present embodiment, the voice recognition results arranged in different order (that is, at different priorities each time) according to the voice recognition rate or the user's pronunciation by the voice signal processing unit 131 and the voice recognition unit 132 are obtained. In addition, the navigation device is calibrated based on information frequently used for searching (ie, information having a high priority applied to weighted time and day).

In other words, as shown in (b) of FIG. 3, the best candidates are rearranged from the highest priority candidates to the lowest candidate candidates in the navigation apparatus. At this time, the weight is applied differently for each hour / day / day / month / month / season / holiday / holiday / event day. Naturally, the top candidate among the best candidates is likely to be the name of a region or a subject for path search.

However, even when the best candidates are rearranged in a more probable order as described above, the number of candidates is still large. Therefore, the user should look at all the best candidates and finally select one best candidate by touching them. When displaying a plurality of candidate lists on one screen, the display interval of the candidates is narrow so that an error may occur when the user selects them by the touch method. Can be. Therefore, it is desirable to reduce the number of candidate lists on the screen as much as possible.

To this end, in this embodiment, each weighted best candidate is converted into a score, and the number of candidate lists to be displayed on the screen can be reduced by applying a threshold (that is, a lower limit score) to the candidates converted into the score. (See Figure 4). For example, two or more final best candidates may be selected according to the threshold (eg, 90 or more points). In this case, as shown in (b) of FIG. 3, the threshold is applied to the selected best candidates by applying a weight.

In conclusion, as shown in (a) of FIG. 3, candidates randomly displayed according to the voice recognition result are weighted as shown in (b) of FIG. 3, so that the candidates are converted into scores in order of score. The best candidates are selected and displayed, and a threshold is applied to the selected best candidates, so that one or two having the highest probability (i.e., converted scores) to be input as a path search target, as shown in FIG. The final best candidates are compressed and displayed.

At this time, either the candidate display screen or the best candidate display screen to which only a weight is applied as shown in (b) of FIG. 3 or the best candidate display screen to which a threshold is applied as shown in (b) of FIG. 4 is displayed. Can be. If there is only one final best candidate on the best candidate display screen to which the threshold is applied, the path search may be immediately performed.

The path search unit 134 may search for a path using one best candidate selected from a user among the best candidate display screen to which the weight is applied only or the best candidate display screen to which the threshold is applied; The threshold is applied to search the path using one candidate selected as the final best candidate.

The voice recognition data storage 140 stores voice recognition data related to map search or route search.

In the present embodiment, the voice recognition data includes a search word (eg, a destination) and search information (eg, current location, etc.) input by the user of the navigation device by hour / day / day / month / month / season / holiday / holiday / event day. Route, etc.) (for example, the current location and destination when a route search is performed).

The search word (for example, the destination input by the user) and the search information (for example, the current location and the path automatically inputted by the navigation device without the user's input) stored in the voice recognition data storage 140 provide an update service. A server (not shown) may be collected when updating data (eg, map data, voice recognition data, etc.) of each navigation device (a navigation device owned by several users).

If the navigation device according to the present embodiment further includes a communication unit (eg, a mobile communication module) (not shown), the data (eg, map data, voice) in real time in a server (not shown) that provides the update service. Update services such as recognition data, etc., and online search for the search word (e.g., the destination entered by the user) and search information (e.g., the current location and route automatically input from the navigation device without user input). Can also be collected through.

In this case, the information collected by the server (not shown) providing the update service (for example, a search word and search information of each navigation device) includes not only information input by a user but also information input by a key (button).

In addition, the server (not shown) that provides the update service analyzes and processes search terms and search information of the individual users (ie, individual navigation devices) collected as the voice recognition data related to map search or route search. Provided as update information to the individual navigation device.

At this time, the server (not shown) providing the update service inputs the collected information (eg, user's search terms and search information) hourly, daily, weekly, monthly, seasonal, holiday, holiday, and event day. Higher weights are given to higher search terms, and the weight information is provided together. That is, the collected information (eg, a search word and search information of an individual navigation device) is converted into a weighted score.

In addition, the controller 130 of the navigation device may apply a threshold (ie, a lower limit score) to information (eg, a search word and search information) converted into the score. That is, the controller 130 displays information (ie, candidates) having a score equal to or greater than the threshold, and when the information (ie, candidates) is equal to or more than a predetermined number, the user searches for one of the candidates for path search. If the information (i.e., candidate) is one, the route can be directly searched using the one candidate to improve the route search performance.

The map data storage unit 150 stores map data.

In this case, the stored map data includes geographic coordinates indicating latitude and longitude in units of degrees / minutes / seconds. In addition to the geographical coordinates, UTM (Universal Transverse Mercator) coordinates, UPS (Universal Polar System) coordinates, TM (Transverse Mercator) coordinates, etc. may be used as the map data.

The guide information output unit 160 outputs road guidance (or route guidance) information and rest information by using an audio signal, a video signal, and / or a vibration signal under the control of the controller 130. Accordingly, the guide information output unit 160 may include a display module (not shown), a sound output module (not shown), and a vibration module (not shown).

The display module may include at least one of a liquid crystal display (LCD), a thin film transistor liquid crystal display (TFT LCD), an organic light emitting diode (OLED), a flexible display, a field emission display (FED), and a 3D display. It may include one. The sound output module may include a receiver, a speaker, a buzzer, and the like.

As illustrated in FIG. 2, when the user speaks, that is, when voice information (or a search word) for route search is input from the user (S101), the controller 130 processes the input voice. An audio signal is extracted (S102). The extracted voice signal is analyzed to recognize voice information (or search word) input by the user (S103).

The controller 130 extracts a candidate (ie, a candidate having a high possibility of matching a desired name) according to the voice recognition result. Then, it is determined whether the extracted candidates are equal to or greater than a preset number (eg, two) according to the speech recognition result (S104).

According to the determination result, if the extracted candidates are less than two (that is, one) (NO in S104), the path is immediately searched using the extracted candidates (S108).

However, if there are two or more extracted candidates according to the determination result (YES in S104), the speech recognition result is corrected by applying a weight to the candidates extracted according to the speech recognition result (S105).

At this time, the weight applied to the candidates is given a higher weight score to the search word having a high frequency of input by time / day / day / day / month / season / holiday / holiday / event day.

Then, a threshold is applied to the speech recognition result to which the weight is applied to correct the speech recognition result (S106).

In this case, the threshold value applied to the candidates means a lower limit score from a score converted by applying the weight for each candidate, and has a probability of matching the highest best candidate (ie, a search word input by the user) according to the applied threshold value. The number of high candidates) can be compressed to two or less.

The controller 130 displays the final best candidate extracted by applying a threshold to the speech recognition result (S107).

If the number of the extracted final best candidates is one, the path is searched using the extracted candidates immediately (S108). If the number of the extracted final best candidates is two or more, one of the candidates is selected by the user and the path is searched.

The controller 130 displays the searched route and performs a road guide (S109).

As described above, the present embodiment sets the ranking of the candidates that can be input in the navigation apparatus by applying a weight using the input frequency by hour and day to the candidate extracted according to the voice recognition result (that is, the candidate of the searchable input words for the path search). The number of final best candidates can be compressed to at least two or less (or one) by (or changing) and applying a threshold to the set (or changed) candidate. Accordingly, even if the speech recognition rate of the speech recognition engine (or the speech recognition algorithm, the speech recognition processor, etc.) is reduced, the user can extract the candidate closest to the search word to be input by the user to perform the path search.

Although the present invention has been described with reference to the embodiments illustrated in the drawings, this is merely exemplary, and various modifications and equivalent other embodiments are possible for those skilled in the art to which the art pertains. I will understand the point. Therefore, the technical protection scope of the present invention will be defined by the claims below.

Claims

A voice input unit configured to receive a voice from a user; And

Recognizing the received voice and extracting candidates based on the voice recognition result, and determining whether the extracted candidates are greater than or equal to a predetermined number, and if the number of the candidates extracted is less than the predetermined number, the extracted candidates are immediately extracted. And a controller for searching for a route using the navigation device.
The method of claim 1, wherein the control unit,

When the extracted candidates are equal to or more than a predetermined number, the voice recognition result is corrected by applying a preset weight to the extracted candidates according to the voice recognition result, and the rank of the candidates is changed according to the conversion score. Navigation device.
The method of claim 2, wherein the control unit,

When the extracted candidates are equal to or greater than a predetermined number, the speech recognition result is corrected by applying a threshold to the extracted candidates according to the weighted speech recognition result, thereby compressing the number of candidates to a predetermined number or less. Navigation device.
The navigation apparatus of claim 3, wherein the controller immediately performs a path search for the one candidate when the number of candidates compressed by the threshold is one.
The method of claim 2, wherein the control unit,

The navigation device, characterized in that a higher weighted score is assigned to a candidate having a high frequency of search term input by the user according to at least one of hourly, daily, day of the week, monthly, seasonal, holiday, holiday, and event day.
As a control method of a navigation device capable of speech recognition,

Recognizing, by the controller, a voice received from the user and extracting a candidate based on the voice recognition result;

Determining whether the extracted candidate is greater than or equal to a predetermined number;

If the number of the extracted candidates is less than the predetermined number, immediately searching for a path using the extracted candidates, and performing a path guidance of the searched paths. .
The method of claim 6, wherein the candidate,

And a candidate having a high possibility of matching a name desired by the user for path search in the navigation device.
The method of claim 6,

If the extracted candidates are greater than or equal to a predetermined number, further comprising the step of correcting the speech recognition result by applying a preset weight to the extracted candidates according to the speech recognition result;

The control unit controls the navigation apparatus, characterized in that for changing the rank of the candidates according to the conversion score.
The method of claim 8, wherein in the correcting of the voice recognition result by applying the weight,

The controller may be configured to assign higher weight scores to candidates having a higher frequency of inputting a user's search word according to at least one of time, daily, day of the week, monthly, season, holiday, holiday, and event day. Control method of the device.
The method of claim 8,

If the extracted candidate is more than a predetermined number,

And correcting the voice recognition result by applying a threshold to the weighted speech recognition result, compressing the number of candidates to a predetermined number or less.
The method of claim 10, wherein in applying the threshold:

The threshold value is a control method of the navigation device, characterized in that for applying the weight for each candidate means a lower limit score in the converted score.