Reflex control systems of technical devices

A university in Kyiv (Ukraine) offers a technology for building reflex control systems of technical devices which provides a reflection of users’ interaction with a technical device in the memory of the intellectual systems and the development adequate reaction on the users influences (like living beings). The technology reduces the cost of developing systems that respond to a person's voice, visual image, etc. The university is looking for partners interested in the creation of "smart" devices.

POD Reference

TOUA20140702001

Summary

A university in Kyiv (Ukraine) offers a technology for building reflex control systems of technical devices which provides a reflection of users’ interaction with a technical device in the memory of the intellectual systems and the development adequate reaction on the users influences (like living beings). The technology reduces the cost of developing systems that respond to a person's voice, visual image, etc. The university is looking for partners interested in the creation of "smart" devices.

Description

Conventional voice control system is based on the analysis of natural language texts, which requires the creation of significant in volume dictionaries and linguistic processor. That requires a significant investment of time and money. The proposed technology - a new idea, based on the results obtained in the theory of non-force interaction. The necessary reaction of any technical device is determined through non-force interactions on it (including human voice).

From this and realization of the idea: using the mathematical apparatus of non-force interaction to calculate an impact on the possible reactions and to choose the reaction, which has the maximum of non-force impact. In systems, that are created by the proposed technology have no dictionaries, text input is not analyzed.

They contain sets of input actions (sounds) and the value of the impact of these sets on the possible reactions.

Technology involves three phases:

- Accumulation of information about the adequate reactions (mostly carried out through observation and reflection in the reflex system actions and reactions to them);

- Calculation patterns of response to the combination of effects;

- Production reactions.

Prototypes were developed, which respond to the voice of the person (systems of TV control by the voice, the prediction of results of the football matches, etc). In the systems of TV control by the voice, man utters the phrase that contains information about the wishing "reaction" and performs this action by the TV remote. For example: "Turn on, please, channel ICTV». After training, the system itself starts to react to the commands that contain other non-informative components, due to the fact, that the impact of informative components to the desired reaction is much stronger.

As experimental studies showed that the probability of correct response is almost equal to 1.

For football matches kept statistics of previous meetings. Reaction - the result of next game. Just like with TV voice control system, an appropriate reaction is produced (predicted result). The probability of correctly predicting the results of matches is about 0.6.

The software according to the proposed technology could be used in the existing technical devices, in the expert and recognition systems. Particularly interesting application - the ability to create dentures which could reflexively react to the human nervous system signals.

Advantages and Innovations

The main advantages of the technology:

- The low cost of development,

- Short term of implementation of such projects,

- The high efficiency of the products.

The systems created by the technology are simpler, cheaper, don’t limit the natural language, easily trained; well able to allocate from the talk spurt the useful information and vice versa, don’t respond to appeals which do not contain any commands, can search useful information in a stream of free speech, are tolerant of errors in the speech of the speaker; allow a reprogram of system behavior by voice, perform the speaker identification; are adapted to the style and habits of the speaker.

Stage of Development

Available for demonstration

IPR status

Patent(s) applied for but not yet granted

Profile Origin

National R&D programme

Deadline Date

7/16/2015 12:00:00 AM

Keywords

Technology Keywords

001002003 Artificial Intelligence (AI)
001002022 Smart Appliances

Market Keywords

002007016 Artificial intelligence related software
002007020 AI programming aids

Partner Sought

Type and Role of Partner Sought

Type of partner sought: Investors, research and production companies in intelligent software and technical systems.

Specific area of activity of the partner: creation of "smart" devices in medicine, engineering, instrumentation, engineering, aircraft manufacturing, artificial intelligence systems.

Tasks to be performed: partners would be interested in collaboration regarding creation the intellectual reflexively reacting fillings for technical devices.

Type and Size of Partner Sought

R&D Institution
SME 11-50
University

Type of Partnership Considered

Research cooperation agreement
Technical cooperation agreement