Agent-Oriented Control of Robotic Systems Based on Artificial Intelligence: Utilising Cutting-Edge Neuroscientific Insights for Verification, Validation, and Optimisation

Abstract

Using artificial intelligence and the latest developments in neuroscience, the authors investigate the challenges in improving the efficiency of agent-oriented control in robotic systems. The goal is to develop a theoretical concept taking into account the existential-objective, simulation-cognitive, and neurobehavioural levels of control in agent-oriented systems. This will lay the groundwork for improving the hardware and software of such agents. The article employs methods of system-analytical and comparative analysis, neuro-oriented modelling of control concepts, and formal verification approaches. The research results in the creation of an original theoretical foundation for Neuro-Agentic Verification Control (NAVC), which combines digital twin simulation, synaptic neural networks, and cognitive-ontological principles. The authors paid significant attention to the integration of formal interfaces and natural language, and other cognitive systems. Finally, the authors provide formal proofs to demonstrate the validity of the key components of the proposed architecture. The article’s international significance is determined by the relevance of addressing challenges related to transparency, safety, and reliability of autonomous robotic systems in critical domains, including transportation, industry, and defence.