In the past decade, numerous machine learning algorithms have been shown to successfully learn optimal policies control real robotic systems. However, it is common encounter failing behaviors as loop progresses. Specifically, in robot applications where undesired but not catastrophic, many struggle with leveraging data obtained from failures. This usually caused by (i) failed experiment ending ...

A novel robot learning algorithm called Predictive Sequence Learning (PSL) is presented and evaluated. PSL is a model-free prediction algorithm inspired by the dynamic temporal difference algorithm S-Learning. While S-Learning has previously been applied as a reinforcement learning algorithm for robots, PSL is here applied to a Learning from Demonstration problem. The proposed algorithm is eval...

This paper presents a series of experiments with an autonomous mobile robot to acquire a model of the robot's interaction with a speciic environment. In these experiments, an artiicial neural network model of the environment is acquired, using data obtained from the robot interacting with the real world. To investigate the feasibility of this approach we have modelled fundamental sensory percep...

During interaction with learning robots, students often experience difficulties understanding the robot intent and its practical realization. To address this challenge, we propose a connected environment that integrates the robot, its digital twin and virtual sensors. The environment supports digital experiments that enable the physical robot to determine optimal policy for performing a manipul...

In this paper we describe an unsupervised robot learning method which is based on the robot learning a mapping between sensors and trajectory velocities. This enables the robot to acquire object avoidance, wall following and goal seeking behaviors simultaneously without incurring the credit assignment problem. To improve the robot’s perception and behaviors we provide the robot with 7 FAM matri...

This paper presents an attempt on incremental robot learning from demonstration. Based on previously learnt knowledge about a task in simpler situations, a robot learns to fulfill the same task properly in a more complicated situation through analyzing comparative demonstrations and extracting new knowledge, especially the constraints that the task in the new situation imposes on the robot’s be...

as a robot could be stable statically standing on three or more legs, a six legged walking robot can be highly flexible in movements and perform different missions without dealing with serious kinematic and dynamic problems. an experimental six legged walking robot with 18 degrees of freedom is studied and built in this paper. the kinematic and gait analysis formulations are demonstrated by an e...

This paper presents an efficient reinforcement learning method, called the PQ-learning, for intelligent behavior acquisition by an autonomous robot. This method uses a special action value propagation technique, named the spatial propagation and temporal propagation, to achieve fast learning convergence in large state spaces. Compared with the approaches in literature, the proposed method offer...