Voice control system for people with disabilities.
Voice control system for people with disabilities.
Activity Start Date
Advantages and Innovations:
Adaptive technology of processing of speech signals based on a unique method, empirical mode decomposition (EMD). The adaptability of the method is that the basis functions used in the processing (decomposition), are extracted directly from the original speech signal and allow to analyze only the characteristic features.
The use of EDM in the handling will provide:
- high adaptability to short-term changes in speech signals;
- detection of hidden modulations and areas of concentration of energy;
- the ability to analyze both global and local properties of speech signals;
- use a smaller computational resources.
In turn, these above mentioned advantages of handling with the use of EMD will increase the consumer properties of the voice control system "smart house":
- recognition accuracy;
- allowable noise level ensuring accurate recognition;
- personalization (speaker-independent/dependent voice control).
Conducted research work aimed at confirming declared by authors of the concept and identification of disadvantages of the use of adaptive method EMD in the processing of speech signals.
Date of creation:
Description of the offer:
There is a need to create systems of social adaptation, i.e. providing people with disabilities with specialized tools that will help them to solve problems in the areas of health, welfare, transport, social protection and education. For this reason, it is important to develop systems for people with disabilities, intended for management and technical purposes ("smart house") with voice commands, in the capacity of which can be any of acoustic signals, is able to reproduce by people with disabilities (distorted speech, certain sound in the form of bellowing, wheezing, etc.).
In existing voice control systems for processing and analysis of speech signals, as a rule, there are two main approaches: time-domain analysis and frequency-domain analysis. But both approaches are not adaptive - processing is carried out on the data sample a sufficiently large duration, short and local changes does not significantly contribute to the analysis of speech signal. From this it follows that for the development of effective methods, algorithms and means of processing speech signals in voice control, it is necessary to use adaptive mathematical methods. For this reason proposed to use the method of empirical mode decomposition.
Department of the Russian university has developed a voice control system "smart house", based on adaptive technology of processing of speech signals and is used as a management tool highly intelligent complex "smart house" for people with disabilities.
The superiority of this approach is the high adaptability. Components of the speech signal, resulting from the decomposition helps to identify the unique properties of user voice for voice control system.
The proposed adaptive empirical mode decomposition method, which ensures competitiveness in comparison with other methods of speech processing, meets the high adaptability to short-term changes in speech signals, detection of hidden modulations and areas of concentration of energy, possibility of the analysis both global and local properties of speech signals, the use of less computational resources.
The closest competitors (technologies of speech processing) are: off-the shelf software component for speech recognition, open application programming interface. The main difference between the proposed method from the above mentioned competitors is the use of adaptive empirical mode decomposition in the algorithms and the nodes in the speech processing. Which improves the competitiveness of consumer properties of the proposed voice control system for people with disabilities by improving performance; improving recognition accuracy of voice commands; full working capacity without access to the internet.
The authors are looking for partners working in this area for scientific cooperation. In the framework of research cooperation agreement the authors plan to conduct further joint research in this direction to achieve the desired result – an approbation of this technique on people with disabilities for further implementation of its development in their life: to improve quality of life and adaptation in society (the use of this technology as applied in pathological deviations in the work of the respiratory system and extends the functionality of the prediction of diseases of the respiratory organs of speech).
07005004 Education and educational products and materials
Partner's sphere of activity:
Type: equal partner, a research organization, universities, institutes, research institutes
Area of activity: scientific research, development.
Partner's role: conducting joint research on the development and use of voice control system "smart house" for people with disabilities.
Валидирован в EEN
A department of a Russian university, specializing in speech technologies, has developed a voice control system that is used as a management tool highly intelligent complex "smart house" for people with disabilities. The authors are looking for partners working in this area for cooperation in the framework of research cooperation agreement.