Hybrid model based hierarchical reinforcement learning for contact rich manipulation task

Sammanfattning: Contact-rich manipulation tasks forms a crucial application in industrial, medical and household settings, requiring strong interaction with a complex environment. In order to efficiently engage in such tasks with human-like agility, it is crucial to search for a method which can effectively handle such contact-rich scenarios. In this work, contact-rich tasks are approached from the perspective of a hybrid dynamical system. A novel hierarchical reinforcement learning is developed: model-based option critic which extensively utilises the structure of the hybrid dynamical model of the contact-rich tasks. The proposed method outperforms the state of the art method PPO and also the previous work of hierarchical reinforcement learning: option-critic, in terms of ability to adapt to uncertainty/changes in the contact-rich tasks.