GB2517320A

GB2517320A - Communication aid

Info

Publication number: GB2517320A
Application number: GB1418390.9A
Authority: GB
Inventors: Barnaby Hawes
Original assignee: SENSORY SOFTWARE INTERNAT Ltd
Current assignee: SENSORY SOFTWARE INTERNAT Ltd
Priority date: 2014-10-16
Filing date: 2014-10-16
Publication date: 2015-02-18
Anticipated expiration: 2034-10-16
Also published as: GB2517320B; GB201418390D0

Abstract

The communication aid (1, figure 1) has an user controlled selection input device 11, such as an eye tracking peripheral or touch sensor, a display 3, one or more processing devices 8, and data storage 9 in which is stored a plurality of phrases 10a each associated with time data representative of when the phrase was last used and/or location data representative of the location of the communication aid when the phrase was last used. The communication aid is used by differently abled users for communication by synthesizing speech using phrases individually selected by the user. The communication aid includes software for weighting individual phrases in the data storage with reference to one or both of the time data and the location data associated with each phrase and ordering the phrases with respect to their weighting values on the display. In this way, the phrases most likely to be selected by the user at a particular time and/or a particular location are made easily accessible to the user by being listed at or near the top of the displayed list of available phrases for selection by the user. The communication aid may be an augmentative and alternative communication (AAC) device.

Description

COMMUNICATION AID

FIELD OF THE INVENTION

[0001] The present invention relates to a communication aid, a communication system and a communication method for enhancing the communication with others of individuals with speech and language challenges and to an adaptive search program and methodology for use in the communication aid, system and method. In particular, but not exclusively, the present invention relates to an augmentative and alternative communication (AAC) device for use by a differently abled person and to an adaptive search program and methodology for use in an AAC device.

BACKGROUND OF THE INVENTION

[0002] Augmentative and alternative communication (AAC) generally means any communication method which is used by individuals with impairments to supplement or replace speech or writing in the production or comprehension of spoken or written language. AAC is useful in relation to a wide range of speech and language impairments including, but not limited to, congenital impairments such as cerebral palsy, intellectual impairment and autism, and acquired conditions such as motor neurone disease, aphasia, and conditions arising, for example, from a brain or spinal injury and Parkinson's disease.

[0003] One type of augmentative and alternative communication is a voice output communication aid (VOCA) which generates synthesized speech from an alternative input. Examples of current voice output communication aids include, but are not limited to, the Sensory Software International Ltd's GridTM Pad and I12TM ofTobiiAB.

DESCRIPTION OF THE RELATED ART

[0004] In US 2011/0257977 an AAC system is described which associates graphics and audio files with topic information in vocabularies that can be accessed by users to enable them to create outlines for speech specific to different events and situations.

[0005] In WO 2010/1 20559 an AAC system is described which allows users to create and add personalized content and to convey such personalized content to others. The AAC system also enables users to customize the content selection interfaces of the system.

[0006] Many VOCA systems can be time-consuming to use. Usually, a user is presented with one or more pages of ordered selectable content which must be navigated by the user to locate and then select the item of content they wish to use in their next speech segment. Especially for users with limited mobility or limited mobility control, scrolling through a large amount of ordered selectable content introduces substantial time delays in any communication exchange with another individual and can frustrate a user's spontaneity.

SUMMARY OF THE INVENTION

[0007] The present invention seeks to mitigate or reduce the difficulties mentioned above in relation to the use of vocal output communication aids.

[0008] In particular, the present invention seeks to provide a communication aid, system and methodology which re-orders selectable content so as to present first content assessed to be the most likely to be needed by the user.

[0009] The present invention therefore provides a communication aid comprising: an input port adapted for communication with a user controlled selection input device; a display; an output port adapted for communication with a sound reproduction device; one or more processing devices; and a memory in which is stored a plurality of phrases each associated with: one or both of: i) time data representative of when the phrase was last used, and ii) location data representative of the location of the communication aid when the phrase was last used, and further comprising stored within said memory a set of instructions for performing the following steps: determining a group of potential candidate phrases for selection by the user; determining a weighting value for each phrase in the group of potential candidate phrases, the weighting value being determined with respect to one or both of the time data associated with the phrase and the location data associated with the phrase; arranging the phrases in the group of potential candidate phrases in order with respect to the weighting value of each phrase; displaying to the user the arranged potential candidate phrases; receiving at the input port a user selection input of a phrase selected from the arranged potential candidate phrases; and communicating the user selected phrase to the output port for onward communication to a sound reproduction device.

[001 0] In a further aspect the present invention provides a method of synthesizing speech substantially in real time using one or more phrases selected by a user, the method comprising the steps of: determining from a plurality of phrases stored in a memory of a communication aid a first group of potential candidate phrases for selection by a user; determining a weighting for each phrase in the first group of potential candidate phrases, the weighting value being determined with respect to one or both of time data associated with the phrase representative of when the phrase was last used and location data associated with the phrase representative of a location of the communication aid when the phrase was last used; arranging the phrases in the group of potential candidate phrases in order with respect to the weighting value of each phrase; displaying to a user the arranged group of potential candidate phrases; receiving at an input port of the communication aid a user selection input of a phrase selected from the arranged group of potential candidate phrases; and synthesizing speech corresponding to the user selected phrase.

[0011] Preferably the communication aid and method further includes instructions for calculating for each phrase in the group of potential candidate phrases a time difference between either a current time or a user selected time and the time data associated with the phrase wherein the weighting value is based upon the calculated time difference and ideally the weighting value is proportional to the inverse of the calculated time difference.

[001 2] Also the communication aid and method may further include instructions for comparing for each phrase the calculated time difference to a threshold and assigning a minimum weighting value to phrases that have a calculated time difference at least exceeding the threshold.

[001 3] Also in a preferred embodiment the communication aid and method further includes instructions for calculating for each phrase in the group of potential candidate phrases a distance between either the current location of the communication aid or a user selected location and the location data associated with the phrase wherein the weighting value is based upon the calculated distance and ideally the weighting value is proportional to the inverse of the calculated distance.

[0014] Also the communication aid and method may further include instructions for comparing for each phrase the calculated distance to a threshold and assigning a minimum weighting value to phrases that have a calculated distance at least exceeding the threshold.

[001 5] In a particularly preferred embodiment the step of determining a group of potential candidate phrases for selection by the user further includes receiving a user input of one or more words whereby the group of potential candidate phrases is determined by identifying potential candidate phrases each with a non-zero weighting value based upon the presence of the user inputted one or more words in the phrase.

[0016] For a user input of a plurality of words in a user inputted order, the set of instructions may further include instructions for determining a weighting value for each phrase in the group of potential candidate phrases based upon the user inputted words appearing in the phrase in the same order as the user inputted order.

[001 7] Also the communication aid and method may further include a reference source stored within said memory and the set of instructions may further include instructions for determining the number of occasions the user inputted one or more words appears in the reference source.

[001 8] Optionally the communication aid and method may further include instructions for determining for each phrase in the group of potential candidate phrases the number of occasions the user inputted one or more words appears in each phrase and instructions for calculating a weighting value for each phrase corresponding to the ratio of the number of occasions the user inputted one or more words appears in the reference source to the number of occasions the user inputted one or more words appears in the phrase.

[001 9] The communication aid and method may further include instructions for adding new time data and / or location data to a phrase in the phrase library each time the phrase is selected by the user.

[0020] Also the communication aid and method may further include instructions for storing new phrases inputted by the user in association with one or both of time data and location data.

[0021] In a further aspect the present invention provides a communication system including a communication aid as described above and a sound reproduction device connected to the output port of the communication aid.

[0022] The communication system may further comprise a user selection input device. The user selection input device may be one or more of: an eye tracking device, a contact sensor and a tube incorporating a pressure sensor.

[0023] Additionally the communication system may further comprise a support for mounting the communication system on a wheelchair.

[0024] With the present invention, a communication aid and method is provided which applies improved weightings to user selectable phrases so that the words and phrases most likely to be selected by the user at a particular time and/or a particular location are most easily accessible to the user by being listed at or near the top of the displayed list of user selections.

[0025] Further benefits and advantages provided by the communication aid will be understood from the following detailed description of a preferred embodiment.

BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Embodiments of the invention will now be described, by way of example only, with reference to the accompanying drawings, in which: Fig. I illustrates a communication aid (VOCA) according to an embodiment of the present invention; Fig. 2 is a functional diagram of key components of the processing system of the communication aid of Fig. 1; Fig. 3 is a functional diagram of key components of the communication interface of the communication aid of Fig. 1; Fig. 4 is a flow diagram of the basic operational steps of the communication aid of Fig. 1; and Fig. 5 is a flow diagram of the functional steps of the selectable candidate search performed by the communication aid of Fig. 1.

DETAILED DESCRIPTION

[0027]As illustrated in Figure 1, in overview a voice output communication aid 1 comprises: a processing system 2 which includes a display 3, a communication interface 4 (referred to herein as a back box), and one or more user input peripherals 11; and, optionally, a stand or support 5 for positioning the processing system 2 so as to be aligned with and/or accessible to the user. For example, the stand 5 may be in the form of a tubular arm which includes a clamp 6 at one end for engagement with the arm or frame of a wheelchair, as illustrated in Fig. 1. In an alternative embodiment the communication interface 4 may be omitted and replaced by suitable programming of or integration within the processing system 2. In a further alternative embodiment the display 3 may comprise a touch screen display in addition to or as a substitute for separate user input peripherals 11.

[0028] The processing system 2 may be a conventional mobile processing device such as, but not limited to, laptops, personal digital assistants (ADA5), tablets and mobile phones with disability enabled functionality including smart phones. Alternatively, the processing system 2 may be a mobile electronic device adapted specifically for use by differently abled individuals having speech or language impairments.

[0029]The processing system 2 includes program storage 7 in the form of any suitable read only or read/write tangible memory. Such memory is preferably, but not limited to, solid state drives (SSD5) such as flash memory having one or more sets of instructions stored therein which control the implementation and operation of the communication aid 1. The one or more sets of instructions for controlling implementation and operation of the voice output communication aid I are preferably adapted to operate on commercially available computer operating systems including, but not limited to, Microsoft Windows1M, WPTM, iOSTM, OS XTM, Mac O5TM, Linux1M, QNXTM and AndroidlM.

[0030] The processing system 2 further includes one or more processors 8 which perform the operational steps defined by the one or more sets of instructions stored in the program storage 7. The processors 8 may be any processor suitable for use in a mobile device including, but not limited to, ARMTM processors and x86 processors. The processing system 2 also includes a clock 21, a positioning system 22 which may, for example, identify location using cellular telecoms network and/or a global position system (GPS), input/output (I/O) interfaces 11; an onboard power supply 12; the display 3, which may be touch sensitive, as well as a graphics driver and a graphics buffer; and data storage 9 in the form of read/write tangible memory. The data storage 9 is preferably, but not limited to, solid state drives (SSD5), such as flash memory and within the data storage 9 is stored a phrase library ba and an images library lob.

[0031] The data storage 9 contains one or more relational databases in the form of the phrase library ba and the images library lOb. The phrase library bOa contains a plurality of different individually selectable phrases, each containing two or more separate words. Ideally, the program storage 7 contains conventional Text-To-Speech (TTS) instruction set(s) so that the phrases stored in the phrase library ba can each be reproduced as synthesized speech.

Alternatively each of the selectable phrases may be associated with a pre-recorded audio file. The selectable phrases are also associated with one or more respective thumbnail' images or symbols stored in the image library bOb.

Preferably symbols are associated with individual words. In this embodiment a phrase may contain more than one word that has an associated symbol and so the phrase may be represented by the collective group of symbols associated with a plurality of words contained in the phrase. Alternatively a single symbol from the collective group of symbols may be selected based, for example, on a frequency of use analysis to identify the most distinctive word contained in the phrase. The images may be photographic pictures e.g. of individuals where, for example, a selectable phrase includes the name of the individual or may be computer generated images or symbols each representative of a different phrase. In addition, each phrase has associated with it one or both of time data and location data (which is discussed in greater detail below).

[0032] Turning to Fig. 3, the back box 4 includes connection means 4a, in the form of a mounting plate 4a or other co-operable connection, which is adapted to for engagement to a stand or supportS. The back box 4 manages and routes communication of the user with i) the processing system 2, ii) other people and ii) other electronic devices. To perform these functions in a preferred embodiment the back box 4 includes a first connection 13 in the form of a jack for connection with the processing system 2; an eye tracker peripheral connection 14, preferably a USB port, adapted for communication with commercially available eye tracker peripherals such as, but not limited to, the TobU PCEyeTM; one or more audio peripheral connections 15; and one or more input peripheral connections 16 such as, but not limited to, one or more connections for headrest contact sensors for use in a wheelchair headrest.

[0033] In addition, the back box 4 optionally includes an infra-red (IR) sensor 17 and an IR transmitter 18 (two are shown in Fig. 3); and a radio transceiver 19, such as, but not limited to, EasyWaveTM transceiver or a BluetoothlM transceiver.

Both the IR sensor and transmitter(s) 17, 18 and the radio transceiver 19 are provided for wireless communication with and control of one or more physically remote peripheral devices such as, but not limited to, a television control, an environmental heating/cooling control, window curtain controls, environmental lighting controls and door and security controls. The back box 4 additionally includes a power socket 20 for connection to an external power supply and/or an on-board power source such as a battery (not shown).

[0034] The audio peripheral connections 15 are provided to enable connection with a sound reproduction unit (not shown) in the form of, for example, one or more speakers. A sound reproduction unit may be used to replace or augment in a conventional manner any sound reproduction system integral to the processing system 2.

[0035] As mentioned earlier, the one or more sets of instructions stored within the program storage 7 control the implementation and operation of the voice output communication aid 1. In overview, the functional steps performed by the voice output communication aid 1 are illustrated in Fig. 4. At step SI the voice communication aid software is launched and the one or more sets of instructions are read from the program storage 7. The launch may be in response to a launch command input by the user via one of the input peripherals of it may be selected for an automatic launch whenever the processing system 2 is started.

An initial or start up screen is displayed 52 on the display 3 and a search of the data storage 9 for selectable phrase candidates is performed 53 based on i) any search criteria input by the user, ii) previous usage, Di) the current time (minute/hour/day/month) accessed from the clock 21, and iv) the user's current location accessed from the navigation system 22. The selectable phrase candidates are then ordered based upon a weighting assigned respectively to each phrase candidate and displayed in their assigned order of relevance to the user 54.

[0036] An input from the user is then received 55, the input is in the form of a selection from the ordered list of candidate phrases. The selected phrase is temporarily stored or buffered in construction of a text string 56 and the selected phrase or its associated image is displayed to the user in a statement construction window 57. The user then has the option 58 to perform a new search for another phrase to be added to the end of the current text string under construction or to communicate the buffered text string. If the user wishes to continue construction of the current statement the program returns to step 53. If, on the other hand, a speak' command is received from the user S9, the temporarily stored one or more selected phrases are then processed using conventional TTS software, such as but not limited to AcapelaTM software, to generate audio data. The audio data is used to drive a sound reproduction unit to emit the user selected statement SlO as sounds. The functional steps illustrated in Fig. 4 may then be repeated from step 53 for the user to construct a

new statement.

[0037] Whenever a user selected statement is emitted 510, each sentence of the statement (or where the statement does not grammatically form a sentence, the entire statement) are saved as separate new selectable phrases in the phrase library ba. Individual sentences are automatically identified by the processing system 2 through the presence of conventional closing sentence punctuation e.g. a full stop, exclamation mark or question mark within the statement. The newly saved phrase is then stored in the data storage 9 for use in future searches.

[0038] In Fig. 5 details of the functional steps of the selectable candidate search (53) and the weighting of the selectable candidates for ordering purposes (54) are illustrated. As mentioned earlier, the search at step 53 is based upon four separate criteria: i) any textual input by the user; U) previous usage; Ui) current time; and iv) current location. This has the effect of ensuring that selectable phrases are more likely to appear closer to the beginning of the selectable candidate list when ordered in step 54 where a candidate at least partially matches one or more of the search criteria. Scrolling through a list of selectable content is laborious and time consuming for users with limited or impaired mobility. With the voice input communication aid 1 described herein, these difficulties are reduced by intelligent ordering of candidates automatically so that the most likely selectable content appears close to the beginning of the list.

[0039] Turning to the first of the search criteria, the user may enter text in the form of the starting letter of a desired word. For example, the user may wish to say the phrase hello, how are you?' and so inputs the letter H'. In the search at step 511 all selectable phrases containing words beginning with the letter H' are treated as partial matches and each is assigned an individual first criteria weighting value between 0 and 1. For each potential candidate each separate occurrence of a partial word match within the phrase is assigned a separate weighting value with the first criteria weighting value for the phrase being the summation of each separate weighting value. Phrases not meeting the first search criteria are each assigned a weighting value of 0 and are discarded from subsequent steps in the search. To reduce the number of potential candidates in the search, the user may further refine the search by entering two or more of the initial letters of the desired word so that for a partial match a word must have both letters adjacent each other in the same order at the beginning of the word. To further limit the number of potential candidates the user may enter a full word or a plurality of keywords in the desired phrase. Ideally, for the purposes of the search the keywords do not have to be entered in the order in which they appear in the desired phrase. Optionally, correct ordering of the keywords can be assessed and can contribute an additional weighting value.

[0040] In a particularly preferred embodiment, to significantly reduce the number of potential candidate phrases following a search using the first criteria the communication aid may be adapted to require two or more words to be entered by the user for the first search criteria.

[0041] For all candidate phrases having non-zero first criteria weighting values the second search criteria is then applied at step S12. This second criteria performs a local Term Frequency Inverse Document Frequency (TFIDF) analysis using a predetermined standard reference book for the words contained in each potential candidate and applies a weighting to each potential candidate according to the result of the analysis. The local TFIDF analysis identifies the number of occasions the word appears within the phrase (Df) and for each candidate phrase divides this number by the total number of times the individual words of each candidate phrase appear in the reference book (Tf). This analysis, DfITf, produces a weighting value for each potential candidate phrase whereby the highest weighting values are generated for phrases containing words which occur multiple times within the phrase but which occur infrequently within the standard reference book. In contrast the lowest weighting values are generated for very common words which appear only once within a phrase. The weighting value determined by applying the second criteria is rescaled to a range of 0 to 1 S13 and is then added to the first criteria weighting value. The overall effect of the second criteria is to lower the weighting value of phrases where the partial/full match word is a common word such as "the" and to raise the weighting value of phrases where the partial/full match word is unusual.

[0042] Although not illustrated in Fig. 5, the second criteria of the search may be extended to apply rules common to predictive texting techniques. For example where two phrases are commonly used together by a user, one immediately following the other, the weighting of the second phrase of the pair is increased in a search where the first phrase has already been selected by the user.

[0043] There may be occasions in which a user will not enter any textual input. In such situations all words and phrases are assigned an equal weighting of 1 and the second criteria is not applied.

[0044] For all remaining potential candidate phrases a third search criteria is now applied S14. For each potential candidate phrase a third criteria weighting value is calculated based on how long ago the phrase was last used. To determine the third criteria weighting value for each potential candidate phrase, the inverse of the summation of time elapsed since each previous use of the phrase (identified from the time data) is calculated.

[0045] An elapsed time threshold may be applied, e.g. 1 year, so that all potential candidate phrases last used prior to the threshold are assigned a weighting value of 0. The elapsed time threshold may be different from the 1 yr example given above. Also the elapsed time threshold may be pre-determined and fixed for the voice output communication aid or may be adjustable by the user.

[0046] In a further alternative embodiment, instead of calculating an elapsed time from the current time, the third criteria weighting value may be based upon the minimum relative time difference between the time data when a phrase was last used and a selected earlier time event or regular repeating event. With this alternative embodiment the voice output communication aid is adapted to enable the user to select the time event from which relative time difference is to be determined. This alternative embodiment is particularly useful if a user intermittently but regularly needs to say the same thing. For example, the user may need to say similar things whenever a particular doctor, engineer or service company employee visits. Where such visits take place at regular intervals, e.g. once a month or even once a year, the user is able to set the time event to the last visit so that the calculated relative time difference will be minimal for phrases commonly used only during such visits. This in turn maximizes the weighting value (the inverse of the elapsed time) for phrases used during such visits.

[0047] Having calculated the third criteria weighting value for each potential candidate phrase, the third criteria weighting value is rescaled so that all third criteria weighting values are between 0 and 1 and the rescaled additional weighting value is then added to the combined first and second weighting value S15.

[0048] Where current location data is available by means of the navigation system 22, the fourth search criteria is applied to all remaining potential candidate phrases S16. For each potential candidate phrase a fourth criteria weighting value is determined based on the distance between the user's current location, as determined by the navigation system 22, and location data stored in association with the potential candidate phrases. To determine the fourth criteria weighting value for each potential candidate phrase, the inverse of the distance is calculated so that potential candidate phrases which have been used before in the same location will have a higher fourth criteria weighting value than potential candidate phrases that have only been used in different locations. It will, of course, be apparent that many phrases may be used in many different locations and so will have a plurality of different locations stored as location data associated with the phrase. In such circumstances a weighting value based upon the calculated distance is assigned for each occasion the phrase has been used and the cumulative weighting value is then rescaled (see below). This has the effect of applying a higher weighting value for phrases used in specific locations such as in a coffee bar or at the hospital and reducing the overall weighting value of phrases that are used in a variety of different locations.

[0049] Here too, a distance threshold may be applied when calculating distances.

For example any potential candidate phrases having calculated distances greater than 500 m may be assigned a weighting value of 0.

[0050] Having calculated the fourth criteria weighting value for each potential candidate phrase, the fourth criteria weighting values are rescaled so that the values fall within 0 and 1 and the rescaled fourth criteria weighting value is then added to the combined first, second and third weighting value 817.

[0051] Finally, all surviving candidate phrases are displayed to the user in descending order with respect to its combined and rescaled weighting value 84.

[0052] In the above description the rescaled range of weighting values for each search criteria is the same. However, it is alternatively envisaged that different ranges of weighting values may be employed for the different search criteria. For example the partial matching and TFIDF may be assigned a wider range of weighting values than the time data and location data weighting values so that the results of these search criteria dominate. In a further alternative embodiment the voice output communication aid 1 may be adapted to enable the user to select one or more of the four search criteria which are to be assigned a greater range of weighting values. For example, where particular words or phrases are used at a particular time or at a particular location the user has the option to increase the weighting value range for the relevant search criteria to ensure that words and phrases common to that time and location dominate at the top of the list of potential candidates.

[0053] For the avoidance of doubt, the order in which the search criteria are applied to potential candidate words and phrases is not limited to the particular order described herein. Furthermore, to prevent uncontrolled growth of the phrase library, optionally the voice output communication aid 1 may be programmed to delete from data storage phrases which are not used more than once during a predetermined time period, e.g. 2 years. Similarly if the data contents exceed a threshold, e.g. 100,000 entries, the phrases with the longest elapsed time may be deleted.

[0054] As mentioned earlier, the voice output communication aid may be adapted to separately store a phrase each time a user instructs the phrase to be synthesized. Whenever a new phrase is saved as a search candidate phrase it is associated with the time data representative of the time when the phrase was used and, if available, location data. In addition each phrase is associated with a usage count which is automatically increased to 1 the first time the phrase is stored. Each time a previously stored phrase is re-used new additional time data is associated with the phrase as a separate usage and, optionally, location data, is added, if available, and the associated usage count is increased by 1.

[0055] The functional steps described above are performed substantially in real-time with the list of candidate phrases being refreshed in the event the user inputs additional search criteria.

[0056] The communication aid 1 described herein may be integrated in a bespoke mobile electronic device. Alternatively, the communication aid I may be retro-fitted to existing mobile electronic devices through the provision of suitable user input peripherals and installation of the set(s) of program instructions described herein.

[0057] Although reference is made herein to an eye tracker peripheral for the entry of selections by the user. The communication aid is not limited to this type of user input peripheral. Suitable alternative user input peripherals include but are not limited to: a joystick, suck and puff! blowing pipes e.g. Sip/Puff1M Tunstall Healthcare and head rest touch sensors.

[0058] With the communication aid describe above the program includes text to speech (TTS) functionality for synthesizing speech in real time based upon the phrase(s) selected by the user. In an alternative embodiment the communication aid may include both a phrase library and an audio clip library. The audio clip library includes an audio clip for each phrase stored in the phrase library so that when a user selects a particular phrase, the processing system retrieves from the audio clip library the audio clip associated with the selected phrase and temporarily stores the audio clip until the user is ready to issue the speech.

Additionally, these two embodiments may be combined so that the communication aid includes a TTS functionality for generating a spoken version of any newly created phrase. The synthesized speech is then saved by the communication aid in the audio clip library as an audio clip for future use.

[0059]Although the methodology of the search process refers to the weighting values being rescaled to a range of 0 to 1, it will be apparent that alternative rescaled ranges may be used. Also, as mentioned earlier the rescaled weighting value ranges may differ for different search criteria, so that one or more of the search criteria has a greater impact in the overall weighting value assigned to the phrase. Variation in the weighting value range may be predetermined or adjustable by the user.

[0060] It is to be understood that the communication aid is not limited to the particular embodiment described herein. Individual functional components described herein may be omitted or substituted with alternative components capable of performing similar functions. Moreover, not all components mentioned herein are essential to the proper functioning of the communication aid and may be omitted. Instead the communication aid is limited only to the scope defined by the accompanying claims.

Claims

CLAIMS1. A communication aid comprising: an input port adapted for communication with a user controlled selection input device; a display; an output port adapted for communication with a sound reproduction device; one or more processing devices; and a memory in which is stored a plurality of phrases each associated with: one or both of: i) time data representative of when the phrase was last used, and ii) location data representative of the location of the communication aid when the phrase was last used, and further comprising stored within said memory a set of instructions for performing the following steps: a) determining a group of potential candidate phrases for selection by the user; b) determining a weighting value for each phrase in the group of potential candidate phrases, the weighting value being determined with respect to one or both of the time data associated with the phrase and the location data associated with the phrase; c) arranging the phrases in the group of potential candidate phrases in order with respect to the weighting value of each phrase; d) displaying to the user the arranged potential candidate phrases; e) receiving at the input port a user selection input of a phrase selected from the arranged potential candidate phrases; and f) communicating the user selected phrase to the output port for onward communication to a sound reproduction device.
2. A communication aid as claimed in claim 1, further including instructions for calculating for each phrase in the group of potential candidate phrases a time difference between either a current time or a user selected time and the time data associated with the phrase wherein the weighting value is based upon the calculated time difference.
3. A communication aid as claimed in claim 2, wherein the weighting value is proportional to the inverse of the calculated time difference.
4. A communication aid as claimed in claim 3, further including instructions for comparing for each phrase the calculated time difference to a threshold and assigning a minimum weighting value to phrases that have a calculated time difference at least exceeding the threshold.
5. A communication aid as claimed in any one of claims 1 to 4, further including instructions for calculating for each phrase in the group of potential candidate phrases a distance between either the current location of the communication aid or a user selected location and the location data associated with the phrase wherein the weighting value is based upon the calculated distance.
6. A communication aid as claimed in claim 5, wherein the weighting value is proportional to the inverse of the calculated distance.
7. A communication aid claimed as claimed in claim 6, further including instructions for comparing for each phrase the calculated distance to a threshold and assigning a minimum weighting value to phrases that have a calculated distance at least exceeding the threshold.
8. A communication aid as claimed in any one of the preceding claims, wherein the step of determining a group of potential candidate phrases for selection by the user comprises receiving a user input of one or more words whereby the group of potential candidate phrases is determined by identifying potential candidate phrases each with a non-zero weighting value based upon the presence of the user inputted one or more words in the phrase.
9. A communication aid as claimed in claim 8, wherein, for a user input of a plurality of words in a user inputted order, the set of instructions further includes instructions for determining a weighting value for each phrase in the group of potential candidate phrases based upon the user inputted words appearing in the phrase in the same order as the user inputted order.
10. A communication aid as claimed in either of claims 8 or 9, wherein the communication aid further includes a reference source stored within said memory and the set of instructions further includes instructions for determining the number of occasions the user inputted one or more words appears in the reference source.
11. A communication aid as claimed in claim 10, wherein the set of instructions further includes instructions for determining for each phrase in the group of potential candidate phrases the number of occasions the user inputted one or more words appears in each phrase and instructions for calculating a weighting value for each phrase corresponding to the ratio of the number of occasions the user inputted one or more words appears in the reference source to the number of occasions the user inputted one or more words appears in the phrase.
12. A communication aid as claimed in any one of the preceding claims, wherein the set of instructions further includes instructions for adding new time data and I or location data to a phrase in the phrase library each time the phrase is selected by the user.
13. A communication aid as claimed in any one of the preceding claims, wherein the set of instructions further includes instructions for storing new phrases inputted by the user in association with one or both of time data and location data.
14. A communication system including a communication aid according to any one of claims ito 13 and a sound reproduction device connected to the output port of the communication aid.
15. A communication system as claimed in claim 14 further comprising a user selection input device.
16. A communication system as claimed in claim 15, wherein the user selection input device comprises one or more of: an eye tracking device, a contact sensor and a tube incorporating a pressure sensor.
17. A communication system as claimed in any one of claims 14 to 16 further comprising a support for mounting the communication system on a wheelchair.
18. A method of synthesizing speech substantially in real time using one or more phrases selected by a user, the method comprising the steps of: a) determining from a plurality of phrases stored in a memory of a communication aid a first group of potential candidate phrases for selection by a user; b) determining a weighting for each phrase in the first group of potential candidate phrases, the weighting value being determined with respect to one or both of time data associated with the phrase representative of when the phrase was last used and location data associated with the phrase representative of a location of the communication aid when the phrase was last used; c) arranging the phrases in the group of potential candidate phrases in order with respect to the weighting value of each phrase; d) displaying to a user the arranged group of potential candidate phrases; e) receiving at an input port of the communication aid a user selection input of a phrase selected from the arranged group of potential candidate phrases; and f) synthesizing speech corresponding to the user selected phrase.
19. A method as claimed in claim 18, further including calculating for each phrase in the group of potential candidate phrases a time difference between either a current time or a user selected time and the time data associated with the phrase wherein the weighting value is based upon the calculated time difference.
20. A method as claimed in claim 19, wherein the weighting value is proportional to the inverse of the calculated time difference.
21. A method as claimed in claim 20, further including comparing for each phrase the calculated time difference to a threshold and assigning a minimum weighting value to phrases that have a calculated time difference at least exceeding the threshold.
22. A method as claimed in any one of claims 18 to 21, further including calculating for each phrase in the group of potential candidate phrases a distance between either the current location of the communication aid or a user selected location and the location data associated with the phrase wherein the weighting value is based upon the calculated distance.
23. A method as claimed in claim 22, wherein the weighting value is proportional to the inverse of the calculated distance.
24. A method claimed as claimed in claim 23, further including comparing for each phrase the calculated distance to a threshold and assigning a minimum weighting value to phrases that have a calculated distance at least exceeding the threshold.
25. A method as claimed in any one of claims 18 to 24, further including receiving a user input of one or more words and determining the group of potential candidate phrases by identifying potential candidate phrases each with a non-zero weighting value based upon the presence of the user inputted one or more words in the phrase.
26. A method as claimed in claim 25, wherein, for a user input of a plurality of words in a user inputted order, further including determining a weighting value for each phrase in the group of potential candidate phrases based upon the user inputted words appearing in the phrase in the same order as the user inputted order.
27. A method as claimed in either of claims 25 or 26, wherein the communication aid further includes a reference source and the method further includes determining the number of occasions the user inputted one or more words appears in the reference source.
28. A method as claimed in claim 27, further including determining for each phrase in the group of potential candidate phrases the number of occasions the user inputted one or more words appears in each phrase and calculating a weighting value for each phrase corresponding to the ratio of the number of occasions the user inputted one or more words appears in the reference source to the number of occasions the user inputted one or more words appears in the phrase.
29. A method as claimed in any one of claims 18 to 28, further including adding new time data and / or location data to a phrase in the phrase library each time the phrase is selected by the user.
30. A method as claimed in any one of claims 18 to 29, further including storing new phrases inputted by the user in association with one or both of time data and location data.
31. A method as claimed in claim 14 further comprising receiving user inputs from a user selection input device.
32. A method as claimed in claim 31, wherein the user selection input device comprises one or more of: an eye tracking device, a contact sensor, a touch screen display, a mouse, a joystick and a tube incorporating a pressure sensor.