Memristor-Based Emotion–Circadian Coupled Memory Regulation Circuit and Its Application in Intelligent Industrial Robots

Memory homeostasis in the human brain's memory system is interactively influenced by circadian rhythms via neuroplasticity regulation and emotional states. In the current research field, there are still few hardware implementations that can couple rhythm-emotional dynamics similar to those in the brain. A memristor-based brain-inspired memory regulation circuit is proposed in this article. A synaptic weight dynamic balance model is constructed. The circadian rhythm mechanism is simulated under the circuit. In a waking state, short-term memory (STM) pathways should be established first. Memory integration is strengthened during sleep. The state-dependent regulation of neural plasticity is achieved. A 2-D emotional valence–arousal module and a cortisol concentration synaptic decay model are integrated into the circuit to simulate the mechanisms of long-term memory and STM suppression and decreased emotional stability caused by sleep deprivation. A time varying forgetting rate circuit combined with learning intensity monitoring is also designed. The memory regulation characteristics under different emotional states are monitored, such as forgetting compensation in negative high arousal states and memory weakening in low arousal states. Finally, the circuit is applied to intelligent industrial robots. The circuit provides a hardware implementation approach for the memory–emotion coupling mechanism in brain-inspired intelligent systems.