Dr Kai Yuan

In traditional robot control, algorithms are programmed based on a specific task, which means that if the robot encounters a new scenario that it is not programmed for, it has difficulties to handle novel situations. This greatly reduces the real world applications. To improve the robustness and generality of robot programmes, and thus the capabilities of robots, my research explores ways to combine classical optimal control methods with gradient-free methods from Machine Learning, such as Deep Reinforcement Learning and Bayesian Optimisation. A hierarchical control framework allows the combination of classical control on the lower levels to control the actuators of the robot achieving stability and balance control and Machine Learning on the higher-levels of control for decision making and planning. Using Machine Learning for the decision making and planning enables the robot to behave more intelligent, dynamically and animal-like, while the control algorithms for the lower levels maintain the robustness and stability properties of classical control.