The brain-computer: origin of the idea and progress in its realization

Area of Science:

• Robotics and Artificial Intelligence
• Cognitive Science and Neuroscience

Background:

• Current autonomous systems often rely on complex pre-programmed behaviors.
• Developing adaptable and efficient learning algorithms for real-world robotic applications remains a challenge.

Purpose of the Study:

• To develop a physical analogue of an organism using a brain-inspired architecture for autonomous learning.
• To demonstrate a novel learning algorithm that prioritizes acquired behaviors for enhanced efficiency.

Main Methods:

• Implementation of a scaled-down model car with a brain-inspired memory-based architecture.
• Integration of an output-driven learning algorithm connecting image processing with behavior control.
• Development of a real-world device for purposeful movement and mental map creation in an artificial environment.

Main Results:

• The system demonstrated rapid learning efficiency through the prioritization of acquired behaviors.
• The model car successfully learned to drive by connecting visual input with motor control.
• A mental map was generated from random actions, providing a cognitive structure for autonomous behavior.

Conclusions:

• Prioritizing learned behaviors is a key factor in accelerating the efficiency of autonomous systems.
• Brain-inspired architectures and output-driven algorithms offer a viable pathway for creating adaptive robots.
• The developed system provides a foundation for more sophisticated cognitive robotics and artificial intelligence.