Behavioral decision-making of mobile robots simulating the memory consolidation mechanism of human brain

The ability of robots to make efficient and effective behavior decisions is a critical prerequisite for completing various tasks in unknown and complex environments. When operating in continuously changing environments, robots often struggle to achieve satisfactory decision-making outcomes due to a lack of prior knowledge about the environment. Investigating how to endow robots with human-like learning and decision-making capabilities to make sound decisions in changing environments remains a significant topic in robotics research. This paper models the processes of memory replay and memory consolidation by mimicking the memory conversion mechanism of the human brain, from the cellular level (synaptic consolidation) and system level (system consolidation), and proposes a human-like brain memory consolidation model based on a motivated developmental network. By leveraging different neurotransmitters’ neuromodulatory effects and considering the dependency of memory conversion on the neocortex during the memory transformation process, the model simulates the conversion of short-term memory to long-term memory and the reconsolidation of long-term memory, enhancing memory plasticity and improving the robot’s learning and decision-making capabilities in dynamic environments. Physical experiments and comparative tests have verified the feasibility and superiority of the proposed model.