In real world environments the state is almost never completely known. Exploration is often expensive. The application of planning in these environments is consequently more difficult and less robust. In this paper we present an approach for predicting new information about a partially-known state. The state is translated into a partially-known multigraph, which can then be extended using machine-learning techniques. We demonstrate the effectiveness of our approach, showing that it enhances the scalability of our planners, and leads to less time spent on sensing actions.

While recent advances in offline reasoning techniques and online execution strategies have made planning under uncertainty more robust, the application of plans in partially-known environments is still a difficult and important topic. In this paper we present an approach for predicting new information about a partially-known initial state, represented as a multi- graph utilizing Maximum-Margin Multi-Valued Regression. We evaluate this approach in four different domains, demonstrating high recall and accuracy.

Grasping in an uncertain environment is a topic of great interest in robotics. In this paper we focus on the challenge of object handover capable of coping with a wide range of different and unspecified objects. Handover is the action of object passing an object from one agent to another. In this work handover is performed from human to robot. We present a robust method that relies only on the force information from the wrist and does not use any vision and tactile information from the fingers. By analyzing readings from a wrist force sensor, models of tactile response for receiving and releasing an object were identified and tested during validation experiments.

Nonprehensile manipulation can play a significant role in complex robotic scenarios, especially for maneuvering non-graspable objects. A big challenge is to construct a robust skill for pushing highly-diverse objects. We present a strategy for pushing unknown objects that differ widely in their properties. For this purpose we introduce the concept of a pushing corridor for cluttered environments that leaves the robot sufficient space for corrective motions. We propose a reactive manipulation skill for pushing objects along this collision-free corridor. The motion of the robot is generated and adapted on the fly based on the observed reactions of the object in response to pushing actions. Our results show that the robot is able to successfully push objects in various complex scenarios.

Nonprehensile manipulation such as pushing can play a significant role in complex scenarios. Objects may have diverse, even anisotropic properties under pushing in different environments. This increases the complexity of the pushing problem. We propose an approach to adapting dynamic movement primitives (DMPs) based on the observed object-motion behaviour and experienced forces. We also investigate an alternative optimal control based technique that enable dexterous and adaptive manipulation using pushability and manipulability of objects.

. This paper deals with the problem of learning unknown edges with attributes in a partially-given multigraph. The method is an extension of Maximum Margin Multi-Valued Regression (M 3 VM) to the case where those edges are characterized by different attributes. It is applied on a large-scale problem where an agent tries to learn unknown object-object relations by exploiting known such relations. The method can handle not only binary relations but also complex, structured relations such as text, images, collections of labels, categories, etc., which can be represented by kernels. We compare the performance with a specialized, state-of-the-art matrix completion method.

Inspired by biological systems, complex object manipulation can benefit from using the environment to stabilize the involved objects. We show that this holds as well for robotic manipulation by evaluating how the environment can be used to optimize a screwing task. We compared the same manipulation with and without using the environment. We were able to improve the success rate as well as to minimize the stress for the robots joints by stabilizing the object with pressing it against a table and by using the robots impedance mode to reduce the applied forces.

This paper deals with measuring and reduction of odometry errors of a differential-drive mobile robot. For this purpose an algorithm for reducing systematic odometry errors caused by uncertainty of an effective wheelbase and unequal wheel diameters is proposed. The algorithm calculates three calibration coefficients using an optimization based on Levenberg-Marquardt algorithm. These coefficients compensate the effects of unknown values of wheel radius and unknown value of the wheelbase. The mobile robot location during a motion is determined based on visual system measurements. For evaluating the proposed algorithm three different tests are considered: straight line experiment, turn in place experiments, unidirectional square path experiment. Simulation and experimental results demonstrate the effectiveness of the proposed algorithm in reducing the systematic odometry errors.

The paper deals with design, construction and implementation of bilateral control system using fuzzy regulation. The aim of paper is implementation of a system which can be used as a didactic tool for better understanding of bilateral control concepts, as well as a base for further work. The system is composed of mechanical model of haptic paddle with one degree of freedom and virtual model of haptic paddle and its environment, interconnected using acquisition card. Virtual environments control is created using fuzzy logic. After implementing control algorithm and connecting two models - the mechanical and virtual, the functionality of system was confirmed by experimental data.

The paper deals with one of frequently encountered tasks in process industry - water level control. Proportional Integral Derivative (PID) control is often used for this purpose. Since control parameters of PID controller are fixed and tank system is inherently nonlinear, PID controller should not be used on wider level range. Therefore, this paper analyzes the effectiveness of water level control using fuzzy controller. The fuzzy controller is implemented based on mathematical model of tank and using MATLAB. The controller is implemented on Friendly ARM - embedded computer. Arduino board is used as an acquisition board for collecting sensor data from tank system Festo Didactic DD 3100 and as a PWM signal generator for water pump control. Experimental results confirm that the fuzzy control system has good adaptability in comparison with PID and provided satisfying results.

The paper deals with one of frequently encountered tasks in process industry - water level control. Proportional Integral Derivative (PID) control is often used for this purpose. Since control parameters of PID controller are fixed and tank system is inherently nonlinear, PID controller should not be used on wider level range. Therefore, this paper analyzes the effectiveness of water level control using fuzzy controller. The fuzzy controller is implemented based on mathematical model of tank and using MATLAB. The controller is implemented on Friendly ARM - embedded computer. Arduino board is used as an acquisition board for collecting sensor data from tank system Festo Didactic DD 3100 and as a PWM signal generator for water pump control. Experimental results confirm that the fuzzy control system has good adaptability in comparison with PID and provided satisfying results.

This paper deals with a simple procedure for map building as well as planning mobile robot trajectories. This approach divides the above problem into three stages: first, a design and construction of simple differential wheeled robot with ZigBee wireless communication; second, development of applications for mobile robot and PC which provide 2D map building of robot's environment using IR sensors; and third, development of applications for mobile robot and PC which provide trajectory planning and robot motion control for created maps. PC applications are implemented using Matlab GUI tool. Finally, this paper also shows developed wireless communication protocol for data exchange between mobile robot and PC. The robustness and effectiveness of the proposed system are demonstrated through experimental results.