An assistive system for communication for people with disabilities

Abstract

This invention discloses an assistive system for communication among people with disabilities, relating to the field of intelligent communication technology. The system includes: a signal capture and cloud upload unit, a signal filtering and verification unit, a data fusion and storage unit, a personal profile database construction unit, a personalized communication assistance database establishment unit, a mapping rule table formulation unit, and a mapping rule display unit. This invention captures communication signals from people with disabilities using sensor technology, and after cloud filtering and verification, ensures the accuracy and effectiveness of the communication signals. This greatly enhances the accessibility of communication between people with disabilities and others, enabling them to express themselves more smoothly, reducing communication barriers, and promoting barrier-free information exchange.

Description

Technical Field

[0001] This invention relates to intelligent communication technology, and more specifically, to an assistive system for communication for people with disabilities. Background Technology

[0002] For deaf, visually impaired, and physically disabled individuals, communication is not only a bridge for them to perceive the world and express their emotions and thoughts, but also a crucial means of integrating into society and realizing their personal value. However, the public generally lacks understanding and awareness of the unique communication methods of people with disabilities. This invisible language barrier builds a high wall of communication, hindering their progress in daily interactions, social integration, and personal growth, reducing their quality of life and sense of psychological security, and greatly limiting their scope of activities and the realization of their potential.

[0003] Therefore, developing an intelligent assistive communication system that can help people with disabilities express themselves freely and reduce obstacles in their daily lives is particularly urgent and significant. With the rapid development of technology, cutting-edge technologies such as wearable devices, AI projection, high-precision intelligent recognition, and advanced voice control are gradually being integrated into daily life and work, demonstrating strong application potential and transformative power. The integration and innovation of these technologies lay a solid technical foundation for building an efficient, user-friendly, and highly adaptable communication assistance system for people with disabilities.

[0004] However, due to the difficulty in synchronizing data between different sensor devices, the technical implementation relies on powerful computing capabilities and high-end equipment, leading to increased costs. At the same time, complex functional designs may increase the user's operational burden, while insufficient recognition accuracy may directly affect the communication effect.

[0005] There are currently no effective solutions to the problems in the relevant technologies. Summary of the Invention

[0006] In view of the problems in related technologies, the present invention proposes an assistive system for communication for people with disabilities, so as to overcome the above-mentioned technical problems existing in the existing related technologies.

[0007] Therefore, the specific technical solution adopted by the present invention is as follows:

[0008] An assistive system for communication for people with disabilities, comprising: a signal capture and cloud upload unit, a signal filtering and verification unit, a data fusion and storage unit, a personal profile database construction unit, a personalized communication assistance database establishment unit, a mapping rule table formulation unit, and a mapping rule display unit;

[0009] The signal capture and cloud upload unit is used to capture the communication signals of people with disabilities using sensor technology and upload the communication signals to the cloud server.

[0010] The signal filtering and verification unit is used to filter and verify communication signals and obtain communication data based on a cloud server.

[0011] The data fusion and storage unit is used to fuse and process communication data, and store the fused communication data to obtain a communication database;

[0012] The personal profile database construction unit is used to collect personal information of people with disabilities and build a personal profile database;

[0013] The Personalized Communication Assistance Database Establishment Unit is used to build a personalized communication model based on personal information in the personal profile database, using machine learning algorithms to generate personalized communication data and establish a personalized communication assistance database.

[0014] The mapping rule table formulation unit is used to determine the mapping relationship between the communication database and the personalized communication assistance database and to establish the mapping rule table;

[0015] The mapping rule display unit is used to send the mapping rule table to a preset communication assistance terminal and display the corresponding communication assistance information.

[0016] Furthermore, the communication signals include: gesture signals, eye movement signals, sound signals, tactile signals, and brain-computer interface signals.

[0017] Furthermore, based on cloud servers, communication signals are filtered and verified to obtain communication data, including:

[0018] The cloud server receives communication signals through the network interface and encodes the communication signals into a unified data format.

[0019] The encoded communication signals are inspected according to preset inspection rules to check for omissions and abnormalities;

[0020] If any omissions are found, the cloud server sends instructions to the sensor device to collect the missed communication signals; if any abnormalities are found, the abnormal communication signals are removed.

[0021] Based on the inspection results, obtain communication data.

[0022] Furthermore, the data fusion and storage unit includes: a time registration module, an initial fusion sequence acquisition module, a final fusion sequence acquisition module, and a communication database establishment module;

[0023] The time registration module is used to extract time series data from communication data and perform spatiotemporal registration on the time series data.

[0024] The initial fusion sequence acquisition module is used to apply a weighted fusion algorithm to the time series data after spatiotemporal registration to obtain the initial fusion sequence of various communication data at different times.

[0025] The final fusion sequence acquisition module is used to input the initial fusion sequence into the pre-built communication data fusion model to obtain the final fusion sequence of various types of communication data at different times;

[0026] The communication database establishment module is used to use the final fusion sequence as the final communication data, store the final communication data, and establish a communication database.

[0027] Furthermore, a weighted fusion algorithm is applied to the spatiotemporally registered time-series data to obtain the initial fusion sequences of various communication data at different times, including:

[0028] S11. Based on the time series data after spatiotemporal registration, extract the communication data at each time point, and obtain the average value by calculating the maximum and minimum values ​​of the communication data at each time point;

[0029] S12. Add the average value as a reference value to the spatiotemporal series data at any time, and compare the communication data with the reference value;

[0030] S13. Calculate the expected value of the communication data that is greater than the reference value and less than the reference value respectively, obtain the median expected value, and use the median expected value as the new reference value;

[0031] S14. Repeat steps S12 to S13 until the difference between two consecutive reference values ​​is less than a preset threshold, and the true value is obtained.

[0032] S15. Merge the real values ​​at each time point to obtain the initial fusion sequence.

[0033] Furthermore, the formula for calculating the expected median is:

[0034]

[0035] In the formula, K represents the expected median value;

[0036] E(N) represents the mathematical expectation of communication data that is greater than the reference value;

[0037] E(M) represents the mathematical expectation of communication data that is less than the reference value.

[0038] Furthermore, the initial fusion sequence is input into a pre-built communication data fusion model to obtain the final fusion sequences of various types of communication data at different times, including:

[0039] S21. Based on the initial fusion sequence, set the fading factor of the communication data, and initialize the state variables, noise model and covariance matrix;

[0040] S22. Input the initial fusion sequence into the pre-constructed communication data fusion model, combine the state transition equation of the communication data fusion model with the set fading factor, predict the state value at the current moment, and adjust the covariance matrix.

[0041] S23. Compare the communication data with the predicted state values, calculate the prediction error, and correct the state values ​​based on the Kalman gain to further adjust the covariance matrix.

[0042] S24. Repeat steps S22 to S23 to obtain the final fusion sequence at different times. Furthermore, the formula for calculating the state value is:

[0043] Di = ADi-1 + Bivi + pi

[0044] In the formula, D i This represents the state value of the i-th communication data.

[0045] A represents the state transition equation from the (i-1)th communication data to the ith communication data;

[0046] D i-1 This represents the status value of the (i-1)th communication data.

[0047] B i Represents the covariance matrix of the i-th communication data;

[0048] v i This represents the fading factor of the i-th communication data;

[0049] p i Let represent the noise model of the i-th communication data.

[0050] Furthermore, based on personal information in the personal profile database, machine learning algorithms are used to construct personalized communication models, generate personalized communication data, and establish a personalized communication support database, including:

[0051] The personal information in the personal profile database is preprocessed, and feature data is extracted from the preprocessed personal information to establish a feature dataset;

[0052] By combining communication support needs with machine learning algorithms, a personalized communication model can be constructed.

[0053] The personalized communication model is trained and evaluated using the feature dataset to obtain the evaluated personalized communication model.

[0054] Input personal information into the evaluated personalized communication model and output personalized communication data;

[0055] By integrating personalized communication data with personal information, a personalized communication support database can be established.

[0056] Furthermore, the feature data includes: personal feature data, historical communication records, text content features, behavioral pattern features, and preference and interest features.

[0057] The beneficial effects of this invention are as follows:

[0058] 1. This invention captures the communication signals of people with disabilities through sensor technology and filters and verifies them in the cloud to ensure the accuracy and effectiveness of the communication signals. This greatly enhances the barrier-free communication between people with disabilities and others, enabling them to express themselves more smoothly, reducing communication barriers, and promoting barrier-free information exchange.

[0059] 2. This invention constructs a communication database by integrating communication data, which not only lays a solid foundation for subsequent in-depth data analysis, but also greatly accelerates the processing and response speed of communication information and improves overall communication efficiency. In addition, through machine learning algorithms, feature data can be mined from the personal profile database to construct a personalized communication model that meets the unique needs of each person with disabilities, generating highly customized communication assistance data, realizing personalized customization of communication services, thereby improving communication efficiency.

[0060] 3. This invention ensures effective mapping between the communication database and the personalized communication assistance database through the collaborative work of the mapping rule table formulation unit and the mapping rule display unit, and displays the mapping results in real time on the preset communication assistance terminal, providing intuitive and easy-to-understand communication assistance information for people with disabilities, and further improving the communication experience. Attached Figure Description

[0061] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0062] Figure 1 This is a schematic diagram of an assistive system for communication for people with disabilities according to an embodiment of the present invention.

[0063] In the picture:

[0064] 1. Signal capture and cloud upload unit; 2. Signal filtering and verification unit; 3. Data fusion and storage unit; 4. Personal profile database construction unit; 5. Personalized communication assistance database establishment unit; 6. Mapping rule table formulation unit; 7. Mapping rule display unit. Detailed Implementation

[0065] To further illustrate the various embodiments, the present invention provides accompanying drawings, which are part of the disclosure of the present invention. These drawings are mainly used to illustrate the embodiments and can be used in conjunction with the relevant descriptions in the specification to explain the operating principles of the embodiments. With reference to these drawings, those skilled in the art should be able to understand other possible implementation methods and the advantages of the present invention.

[0066] According to an embodiment of the present invention, an assistive system for communication by people with disabilities is provided.

[0067] The present invention will now be further described in conjunction with the accompanying drawings and specific embodiments, such as... Figure 1 As shown, according to an embodiment of the present invention, an assistive system for communication for people with disabilities includes: a signal capture and cloud upload unit 1, a signal filtering and verification unit 2, a data fusion and storage unit 3, a personal profile database construction unit 4, a personalized communication assistance database establishment unit 5, a mapping rule table formulation unit 6, and a mapping rule display unit 7.

[0068] The signal capture and cloud upload unit 1 is used to capture the communication signals of people with disabilities using sensor technology and upload the communication signals to the cloud server.

[0069] Specifically, communication signals include: gesture signals, eye movement signals, sound signals, tactile signals, and brain-computer interface signals.

[0070] It should be further noted that sensor technology is a device or apparatus capable of sensing a measured object (such as gestures, eye movements, sound, touch, and brain activity) and converting it into a usable output signal (such as an electrical signal), including:

[0071] Gesture recognition sensors capture hand gestures through cameras and use image processing technology to identify specific gestures. They can be integrated into wearable devices or smart home systems to capture gesture signals from deaf and mute individuals.

[0072] Eye-tracking sensors capture the movement trajectory of the eyeballs using infrared light or cameras, thereby analyzing the gaze direction and eye movement information of visually impaired individuals. They have wide applications in fields such as assisted reading and interface navigation.

[0073] Voice recognition sensors (for non-traditional sound signals) may be used more often by deaf people to capture the sound signals they emit through specific devices (such as speech synthesizers) or to identify sounds in the environment for alerts or navigation.

[0074] Tactile sensors can sense tactile information such as pressure and vibration of objects. For people with limb disabilities, these sensors can be integrated into prostheses, wheelchairs or other assistive devices to capture and respond to tactile signals.

[0075] Brain-computer interface (BCI) sensors are a technology that directly connects brain signals to external devices. They capture the brain's electrical activity (such as brainwaves) through electrodes or implanted sensors and convert it into recognizable signals. This technology has great potential in assisting communication and controlling external devices, especially for people with disabilities who have lost traditional communication abilities due to nerve damage.

[0076] Signal filtering and verification unit 2 is used to filter and verify communication signals and obtain communication data based on a cloud server.

[0077] Specifically, based on cloud servers, communication signals are filtered and verified to obtain communication data, including:

[0078] The cloud server receives communication signals through the network interface and encodes the communication signals into a unified data format.

[0079] The encoded communication signals are inspected according to preset inspection rules to check for omissions and abnormalities;

[0080] If any omissions are found, the cloud server sends instructions to the sensor device to collect the missed communication signals; if any abnormalities are found, the abnormal communication signals are removed.

[0081] Based on the inspection results, obtain communication data.

[0082] The data fusion and storage unit 3 is used to fuse and process communication data, and store the fused communication data to obtain a communication database.

[0083] Specifically, the data fusion and storage unit 3 includes: a time registration module, an initial fusion sequence acquisition module, a final fusion sequence acquisition module, and a communication database establishment module;

[0084] The time registration module is used to extract time series data from communication data and perform spatiotemporal registration on the time series data.

[0085] It should be further explained that extracting time series data from communication data and performing spatiotemporal registration of the time series data includes:

[0086] Based on the timestamps in the communication data, extract the time series information for each piece of communication data to ensure the correct time sequence. The communication data can be grouped or aggregated according to different time intervals (such as seconds, minutes, hours) to form a unified time series format. For data with different time intervals, downsampling (if the communication data is too dense) or interpolation techniques (if the communication data is sparse) can be used to ensure the continuity and consistency of the time series. If the communication data contains user geographic location data (such as GPS information, IP address, etc.), this spatial information needs to be extracted from the communication records and converted into standard latitude and longitude coordinates to facilitate subsequent spatiotemporal registration processing.

[0087] Based on the time series and spatial data of communication data, a spatiotemporal registration model is established. The goal of this model is to align data from different times and spaces, ensuring that communication data from different points in time can be compared and analyzed on a unified time and spatial dimension. Time alignment involves aligning time series data from different sources based on timestamp information. Common methods include alignment based on fixed time intervals and event-based time alignment. Spatial alignment uses spatial coordinates to correct the location of communication data from different locations. If spatial data has errors or offsets, registration algorithms (such as the ICP algorithm) can be used for location matching and correction. Spatiotemporal fusion merges the time and spatial aligned data to generate a unified spatiotemporal sequence that reflects communication activities occurring at different times and locations, facilitating subsequent spatiotemporal analysis.

[0088] The initial fusion sequence acquisition module is used to apply a weighted fusion algorithm to the time series data after spatiotemporal registration to obtain the initial fusion sequence of various communication data at different times.

[0089] Specifically, a weighted fusion algorithm is applied to the spatiotemporally registered time-series data to obtain the initial fusion sequences of various communication data at different times, including:

[0090] S11. Based on the time series data after spatiotemporal registration, extract the communication data at each time point, and obtain the average value by calculating the maximum and minimum values ​​of the communication data at each time point.

[0091] S12. Add the average value as a reference value to the spatiotemporal sequence data at any time, and compare the communication data with the reference value.

[0092] S13. Calculate the expected value of the communication data that is greater than the reference value and less than the reference value respectively, obtain the median of the expected value, and use the median of the expected value as the new reference value.

[0093] Specifically, the formula for calculating the expected median is:

[0094]

[0095] In the formula, K represents the expected median value;

[0096] E(N) represents the mathematical expectation of communication data that is greater than the reference value;

[0097] E(M) represents the mathematical expectation of communication data that is less than the reference value.

[0098] S14. Repeat steps S12 to S13 until the difference between two consecutive reference values ​​is less than a preset threshold, and the true value is obtained.

[0099] S15. Merge the real values ​​at each time point to obtain the initial fusion sequence.

[0100] The final fusion sequence acquisition module is used to input the initial fusion sequence into a pre-built communication data fusion model to obtain the final fusion sequence of various types of communication data at different times.

[0101] Specifically, the initial fusion sequence is input into a pre-built communication data fusion model to obtain the final fusion sequences of various types of communication data at different times, including:

[0102] S21. Based on the initial fusion sequence, set the fading factor of the communication data, and initialize the state variables, noise model and covariance matrix.

[0103] S22. Input the initial fusion sequence into the pre-constructed communication data fusion model, combine the state transition equation of the communication data fusion model with the set fading factor, predict the state value at the current moment, and adjust the covariance matrix.

[0104] Specifically, the formula for calculating the state value is:

[0105] Di = ADi-1 + Bivi + p i

[0106] In the formula, D i This represents the state value of the i-th communication data.

[0107] A represents the state transition equation from the (i-1)th communication data to the ith communication data;

[0108] D i-1 This represents the status value of the (i-1)th communication data.

[0109] B i Represents the covariance matrix of the i-th communication data;

[0110] v i This represents the fading factor of the i-th communication data;

[0111] p i Let represent the noise model of the i-th communication data.

[0112] S23. Compare the communication data with the predicted state values, calculate the prediction error, and correct the state values ​​according to the Kalman gain, further adjusting the covariance matrix.

[0113] S24. Repeat steps S22 to S23 to obtain the final fusion sequence at different times.

[0114] The communication database establishment module is used to use the final fusion sequence as the final communication data, store the final communication data, and establish a communication database.

[0115] Unit 4, the Personal Profile Database Construction Unit, is used to collect personal information of people with disabilities and construct a personal profile database.

[0116] It should be further explained that personal information includes identification information (name, gender, date of birth, contact information, ID card number or equivalent unique identification code, etc.), health status information (disability category: such as visual impairment, hearing impairment, speech impairment, physical disability, intellectual disability, mental disability, etc.; disability level: the disability level assessed according to standards, such as level one, level two, level three, level four, etc.; medical history; treatment status, etc.), communication ability assessment (the assessment results of the person with disability's current communication ability, including the effectiveness, proficiency and limitations of the communication methods used, such as gestures, eye movements, touch, voice or brain-computer interfaces), personal preferences and habits (the person with disability's personal preferences in communication, such as preferred communication methods, commonly used words or phrases, interest in specific topics, etc., as well as their daily habits and living environment, which helps to provide communication assistance that is more tailored to individual needs).

[0117] Unit 5, which establishes a personalized communication support database, is used to build a personalized communication model based on personal information in the personal profile database, generate personalized communication data, and establish a personalized communication support database.

[0118] Specifically, based on personal information in a personal profile database, machine learning algorithms are used to build personalized communication models, generate personalized communication data, and establish a personalized communication support database, including:

[0119] The personal information in the personal profile database is preprocessed, and feature data is extracted from the preprocessed personal information to establish a feature dataset.

[0120] Specifically, the feature data includes: personal feature data, historical communication records, text content features, behavioral pattern features, and preference and interest features.

[0121] By combining communication support needs with machine learning algorithms, a personalized communication model can be built.

[0122] The personalized communication model is trained and evaluated using a feature dataset to obtain the evaluated personalized communication model.

[0123] Personal information is input into the evaluated personalized communication model, which outputs personalized communication data.

[0124] By integrating personalized communication data with personal information, a personalized communication support database can be established.

[0125] The mapping rule table formulation unit 6 is used to determine the mapping relationship between the communication database and the personalized communication assistance database and to establish the mapping rule table.

[0126] It should be further explained that determining the mapping relationship between the communication database and the personalized communication support database, and establishing the mapping rule table, includes:

[0127] Identify which data elements in the communication database (such as signal type, keywords, sentiment tags, etc.) may be associated with information in the personalized communication assistance database; assess the importance and relevance of these associations, and determine which factors are most critical for establishing mapping relationships; based on the analysis results, design a series of mapping rules that define how to map data in the communication database to corresponding entries or models in the personalized communication assistance database. Mapping rules include conditional judgments (e.g., "If the signal type is a gesture and the gesture matches X, then call the personalized communication model Y") and data transformations (e.g., converting raw signal data into a format recognizable by the model); organize the designed mapping rules into a table, i.e., a mapping rule table, which should clearly list the input data items in the communication database, the mapping logic, and the output targets in the personalized communication assistance database.

[0128] The mapping rule display unit 7 is used to send the mapping rule table to a preset communication assistance terminal and display the corresponding communication assistance information.

[0129] It should be further explained that the mapping rule table data is packaged into a suitable transmission format, such as JSON, XML, or binary files, to ensure data integrity and readability. Appropriate data transmission methods, such as HTTP requests, WebSocket connections, or FTP transfers, are selected based on the network environment and terminal characteristics. The packaged mapping rule table data is then sent to the preset communication assistance terminal via the selected transmission method; this can be done automatically by the server-side program. When the communication assistance terminal receives communication signals (such as gestures, eye movements, and sounds) from a person with disabilities, the application can parse these signals in real time. Based on the parsed communication signals, the application queries the corresponding mapping rule in the locally stored mapping rule table. Based on the queried mapping rule and information from the personalized communication assistance database, corresponding communication assistance information is generated and displayed visually on the communication assistance terminal screen, such as through text prompts, images, animations, or voice output.

[0130] In summary, by utilizing the technical solutions described above, sensor technology captures the communication signals of individuals with disabilities, which are then filtered and verified in the cloud, ensuring the accuracy and effectiveness of these signals. This significantly enhances the accessibility of communication between individuals with disabilities and others, enabling them to express themselves more smoothly, reducing communication barriers, and promoting barrier-free information exchange. By integrating communication data, a communication database is constructed, laying a solid foundation for subsequent in-depth data analysis and greatly accelerating the processing and response speed of communication information, thus improving overall communication efficiency. Furthermore, through machine learning algorithms, feature data can be extracted from the personal profile database to construct personalized communication models that meet the unique needs of each individual with disabilities, generating highly customized communication assistance data. This achieves personalized customization of communication services, thereby improving communication efficiency. The collaborative work of the mapping rule table formulation unit and the mapping rule display unit ensures effective mapping between the communication database and the personalized communication assistance database, and the mapping results are displayed in real time on the preset communication assistance terminal, providing individuals with disabilities with intuitive and easy-to-understand communication assistance information, further enhancing the communication experience.

[0131] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. An assistive system for communication for people with disabilities, characterized in that, The system includes: a signal acquisition and cloud upload unit, a signal filtering and verification unit, a data fusion and storage unit, a personal profile database construction unit, a personalized communication assistance database establishment unit, a mapping rule table formulation unit, and a mapping rule display unit; The signal capture and cloud upload unit is used to capture the communication signals of people with disabilities using sensor technology and upload the communication signals to the cloud server. The signal filtering and verification unit is used to filter and verify communication signals and obtain communication data based on a cloud server. The data fusion and storage unit is used to fuse communication data and store the fused communication data to obtain a communication database. The personal profile database construction unit is used to collect personal information of people with disabilities and construct a personal profile database. The personalized communication assistance database establishment unit is used to construct a personalized communication model based on personal information in the personal profile database, using machine learning algorithms to generate personalized communication data and establish a personalized communication assistance database. The mapping rule table formulation unit is used to determine the mapping relationship between the communication database and the personalized communication assistance database, and to establish a mapping rule table; The mapping rule display unit is used to send the mapping rule table to a preset communication assistance terminal and display the corresponding communication assistance information.

2. The assistive system for communication for people with disabilities according to claim 1, characterized in that, The communication signals include: gesture signals, eye movement signals, sound signals, tactile signals, and brain-computer interface signals.

3. The assistive system for communication for people with disabilities according to claim 1, characterized in that, The process of filtering and verifying communication signals and acquiring communication data based on a cloud server includes: The cloud server receives communication signals through the network interface and encodes the communication signals into a unified data format. The encoded communication signals are inspected according to preset inspection rules to check for omissions and abnormalities; If any omissions are found, the cloud server sends instructions to the sensor device to collect the missed communication signals; if any abnormalities are found, the abnormal communication signals are removed. Based on the inspection results, obtain communication data.

4. The assistive system for communication for people with disabilities according to claim 1, characterized in that, The data fusion and storage unit includes: a time registration module, an initial fusion sequence acquisition module, a final fusion sequence acquisition module, and a communication database establishment module; The time registration module is used to extract time series data from communication data and perform spatiotemporal registration on the time series data; The initial fusion sequence acquisition module is used to apply a weighted fusion algorithm to the time series data after spatiotemporal registration to obtain the initial fusion sequence of various communication data at different times. The final fusion sequence acquisition module is used to input the initial fusion sequence into a pre-built communication data fusion model to obtain the final fusion sequence of various types of communication data at different times; The communication database establishment module is used to use the final fusion sequence as the final communication data, store the final communication data, and establish a communication database.

5. The assistive system for communication for people with disabilities according to claim 4, characterized in that, The weighted fusion algorithm applied to the spatiotemporally registered time series data to obtain the initial fusion sequences of various communication data at different times includes: S11. Based on the time series data after spatiotemporal registration, extract the communication data at each time point, and obtain the average value by calculating the maximum and minimum values ​​of the communication data at each time point; S12. Add the average value as a reference value to the spatiotemporal series data at any time, and compare the communication data with the reference value; S13. Calculate the expected value of the communication data that is greater than the reference value and less than the reference value respectively, obtain the median expected value, and use the median expected value as the new reference value; S14. Repeat steps S12 to S13 until the difference between two consecutive reference values ​​is less than a preset threshold, and the true value is obtained. S15. Merge the real values ​​at each time point to obtain the initial fusion sequence.

6. The assistive system for communication for people with disabilities according to claim 5, characterized in that, The formula for calculating the expected median is: In the formula, K represents the expected median value; E(N) represents the mathematical expectation of communication data that is greater than the reference value; E(M) represents the mathematical expectation of communication data that is less than the reference value.

7. The assistive system for communication for people with disabilities according to claim 4, characterized in that, The process of inputting the initial fusion sequence into a pre-built communication data fusion model to obtain the final fusion sequences of various types of communication data at different times includes: S21. Based on the initial fusion sequence, set the fading factor of the communication data, and initialize the state variables, noise model and covariance matrix; S22. Input the initial fusion sequence into the pre-constructed communication data fusion model, combine the state transition equation of the communication data fusion model with the set fading factor, predict the state value at the current moment, and adjust the covariance matrix. S23. Compare the communication data with the predicted state values, calculate the prediction error, and correct the state values ​​based on the Kalman gain to further adjust the covariance matrix. S24. Repeat steps S22 to S23 to obtain the final fusion sequence at different times.

8. The assistive system for communication for people with disabilities according to claim 7, characterized in that, The formula for calculating the state value is: Di=ADi-1+Bivi+ p the In the formula, D i This represents the state value of the i-th communication data. A represents the state transition equation from the (i-1)th communication data to the ith communication data; D i-1 This represents the status value of the (i-1)th communication data. B i Represents the covariance matrix of the i-th communication data; v i This represents the fading factor of the i-th communication data; p i Let represent the noise model of the i-th communication data.

9. An assistive system for communication for people with disabilities according to claim 1, characterized in that, The process of building a personalized communication model based on personal information in a personal profile database, using machine learning algorithms to generate personalized communication data, and establishing a personalized communication support database includes: The personal information in the personal profile database is preprocessed, and feature data is extracted from the preprocessed personal information to establish a feature dataset; By combining communication support needs with machine learning algorithms, a personalized communication model can be constructed. The personalized communication model is trained and evaluated using the feature dataset to obtain the evaluated personalized communication model. Input personal information into the evaluated personalized communication model and output personalized communication data; By integrating personalized communication data with personal information, a personalized communication support database can be established.

10. An assistive system for communication for people with disabilities according to claim 9, characterized in that, The feature data includes: personal feature data, historical communication records, text content features, behavioral pattern features, and preference and interest features.

Images