Learning Flexible Decision: Making and Motor Control in Robots for Human-Robot Interaction (A02)

在我们的研究提案中，我们将系统级计算神经科学的创新概念与机器人相结合。我们将开发一个皮质-基底节相互作用的神经计算模型，用于决策和运动控制，并结合小脑的两个模型，一个用于开环控制，另一个用于预测运动动作的感觉后果。我们将在模拟和带有人造皮肤的人形iCub机器人上测试这些模型，以实现和抓取需要朝着特定目标执行运动的任务。在协作工作中，我们将把这种方法扩展到人与机器人的交互，其中对象将从人类传递到机器人。

In our research proposal we integrate innovative concepts of systems level computational neuroscience with robotics. We will develop a neuro-computational model of cortex-basal ganglia interactions for decision making and motor control combined with two models of the cerebellum, one for open-loop control and one for the predictions of the sensory consequences of motor actions. We will test the models in simulation and on the humanoid iCub robot with artificial skin on reaching and grasping tasks where movements shall be executed towards particular goals. In collaborative work, we will extend this approach towards human-robot interaction where an object will be passed from a human to the robot.