Reinforcement learning algorithm application and multi-body system design by using MapleSim and Modelica

Advanced intelligent systems such as robots must be capable to interact with dynamic environment and adapt their behavior to it efficiently. Currently, modeling humanoid robots with sophisticated learning and cognitive capabilities is one of the most challenging issues in the field of intelligent robotics. Robots must be equipped with the ability to modify and add to its knowledge base information gained from its past failings. This might provide stable robust walking on unseen terrains as well. Moreover, a further critical stage in designing and evaluating such a sophisticated complex system is modeling and simulation. This paper describes preliminary work on designing a simple multi-body system by using MapleSim, which is a tool for multi-body modeling/simulation and reinforcement learning algorithm is applied to this multi-body system in terms of using Modelica models.