One central property of cognitive systems is the ability to learn and to improve continually. We present a robot control language that combines programming and learning in order to make learning executable in the normal robot program. The language constructs of our learning language RoLL rely on the concept of hierarchical hybrid automata to enable a declarative, explicit specification of learn...

Fuzzy Q-learning is extending of Q-learning algorithm that uses fuzzy inference system to enable Qlearning holding continuous action and state. This learning has been implemented in various robot learning application like obstacle avoidance and target searching. However, most of them have not been realized in embedded robot. This paper presents implementation of fuzzy Q-learning for obstacle av...

In view of the balance control problem of self-balance robot, the reinforcement learning mechanism based on the advanced updating is proposed as a self-balance bionic learning algorithm of the robot. This algorithm can choose optimal behavior above the baseline with the certain probability by using baseline in the advanced updating and unifying the probability curiosity mechanism in reinforceme...

An intelligent robot agent based on domain ontology, machine learning mechanism, and Fuzzy Markup Language (FML) for students and robot co-learning is presented in this paper. The machine-human co-learning model is established to help various students learn the mathematical concepts based on their learning ability and performance. Meanwhile, the robot acts as a teacher’s assistant to co-learn w...

While considerable progress has been made in recent years toward the development of multi-robot teams, much work remains to be done before these teams are used widely in real-world applications. Two particular needs toward this end are the development of mechanisms that enable robot teams to generate cooperative behaviors on their own, and the development of techniques that allow these teams to...

An appearance-based robot self-localization problem is considered in the machine learning framework. The appearance space is composed of all possible images, which can be captured by a robot’s visual system under all robot localizations. Using recent manifold learning and deep learning techniques, we propose a new geometrically motivated solution based on training data consisting of a finite se...

We applied Reinforcement Learning on a real robot arm, in order to control its movements. The arm is actuated by two pneumatic artificial muscles, that expose a highly non-linear behavior. To enable a significant speed-up of the learning process, an empirical simulation is constructed, based on real robot observations. Furthermore, we introduce a learning strategy to facilitate learning. Using ...

This paper presents an approach for learning complex tasks on real robots, like walking or kicking in a humanoid soccer robot, profiting at most from the possibility to run simulations of a virtual model of the robot. This approach avoids to damage the real robot in the time consuming trials needed to learn a correct behavior and avoids to overfit the virtual robot model. The basic idea is to r...

S3 MIT learning biped. This movie begins with the powered robot imitating passive walking down a 0.9 degree slope, from three camera angles. Then it shows the robot learning to walk on flat terrain with foam protective pads. The controller kicks the robot into random initial conditions between learning trials. After a few minutes, the robot is walking well in place, so we command it to walk in ...