Robots have become an integral part of everyday life. They are used in industry, they solve service tasks, participate in special operations acting autonomously as well as within complex solutions.
Robots have technical vision, understand voice commands and are capable of acting autonomously according to the predefined algorithm.
Robots become really autonomous due to logical artificial intelligence, which makes them capable of building algorithm autonomously in a dynamic mode according to input data and assigned tasks.
Intellectualization is the last factor that constrains development of autonomous robots.
A universal software intellectualization platform, "ROBORAZUM", has been developed by the specialists of the research institute "Mivar". The use of the platform provides robotic suites with artificial intelligence, allows them to operate really autonomously.
"ROBORAZUM" is implemented in the form of an embedded software platform with flexible capabilities to adapt for the control of any robotic suites and systems. To operate efficiently, the platform does not require powerful computational resources.
Own technological developments and approaches of the team "Mivar" have been used in the platform:
- Mivar knowledge representation model
- Mivar-based method of rapid synthesis of algorithms for solving random tasks
- Comprehensive approach to designing self-organizing control systems
If a robot can be controlled remotely, ROBORAZUM will make it autonomous!
- Logical kernel (RAZUMATOR)
- Autonomous control subsystems
The key component of the platform "ROBORAZUM" is the logical kernel "RAZUMATOR" which is responsible for logic and automated execution algorithm construction, allows robots to make decisions autonomously and efficiently solve tasks without human involvement.
"RAZUMATOR" is implemented in the form of an independent software component. This approach allows us to integrate the platform as a logical subsystem into any existing control system.
Advantages of the solution
- Minimum system requirements
- Flexible adaptation to any type of devices
- Easy integration into existing control system
Key characteristics and functional capabilities
- Autonomous adaptive control of robots of various design and purpose without any need of extra programming
- Application of semantic processing, when solving tasks of planning, communication and perception
- Rapid intelligent synthesis of new solution algorithms for assigned tasks
- Comprehensive spatial perception (technical vision, spatial analysis of sound stage, etc.)
- Natural language user interface
- Single integrated knowledge base for operation of all intelligent subsystems
- Basic knowledge allowing a robot to communicate with a human, self-train and solve simple tasks
- Self-training based on the analysis of methodological materials, visual examples, own experience
- Convenient specialization of a robot through additional training according to its applied purpose