篇名 | The Robot Cognitive Developmental Algorithm Based on Cerebellum-Basal Ganglia-Cerebral Cortex Circuits |
---|---|
卷期 | 29:4 |
作者 | Hongge Ren 、 Rui Yin 、 Tao Shi 、 Fujin Li 、 Yingfan Xiang |
頁次 | 001-011 |
關鍵字 | cognitive development 、 intrinsic motivation 、 learning automata 、 sensory motor system 、 two-wheeled robot 、 EI 、 MEDLINE 、 Scopus |
出刊日期 | 201808 |
DOI | 10.3966/199115992018082904001 |
Aiming at continuous two-wheeled robot behavior learning problem, we simulated the human psychological cognitive mechanism and sensory motor phenomena of brain nerve, and proposed a robot cognitive developmental algorithm based on cerebellum-basal ganglia-cerebral cortex circuits. Based on the theory of biological sensory motor system, the algorithm takes the learning automata as the frame. The mapping of the robot state to the behavior is realized in the cerebellum by supervised learning style. Then, we use evaluation module of probability selection based on intrinsic motivation principle for selecting the action of basal ganglia. Finally, the cerebral cortex receives the nerve signal and transfers to the basal nucleus and the cerebellum, forming the complete sensory motor feedback loop. The proposed algorithm was applied to the system of two-wheeled robot, and the experiments were carried out. Simulation experimental results show that robot in unknown environment, through independent learning development, gradually master motion balance control skill, reflecting the effectiveness and robustness of the algorithm. Compared with the classical learning automata algorithm, highlight the superiority of the algorithm.