Abstract-To date, there has been limited work in applying Deep Belief Networks (DBNs) for acoustic modeling in LVCSR tasks, with past work using standard speech features. However, a typical LVCSR system makes use of both feature and modelspace speaker adaptation and discriminative training. This paper explores the performance of DBNs in a state-of-the-art LVCSR system, showing improvements over Multi-Layer Perceptrons (MLPs) and GMM/HMMs across a variety of features on an English Broadcast News task. In addition, we provide a recipe for data parallelization of DBN training, showing that data parallelization can provide linear speed-up in the number of machines, without impacting WER.

No abstract

The article presents some aspects of a complex control system of a teleoperated military mobile robot Tactical Robotic System (TAROS) related to virtual reality and assistance to a human operator in general. Described is especially the unique and innovative system of virtual operator station which uses the HMD Oculus Rift to put the operator into a virtual space containing visual feedback from the robot and camera images, including stereovision. The virtual operator station serves as a cost-effective and portable replacement of what otherwise would be a large room with expensive equipment. Mentioned is also another system that helps the operator with remote manipulation tasks – the anti-collision system preventing damage done to the mechanical parts of the robot by incautious movements of the manipulator arm.

This paper mentions some problems related to utilization of a head-mounted display (HMD) for remote control of mobile robots by a human operator and also presents a possible solution. Considered is specifically the new HMD device called Oculus Rift, which is a very interesting device because of its great parameters and low price. The device is described in the beginning, together with some of the specific principles of the Oculus 3D display. Then follows the design of a new graphical user interface for teleoperation, with main focus on visualization of stereoscopic images from robot cameras. Demonstrated is also a way how to display additional data and information to the operator. The overall aim is to create a comfortable and highly effective interface suitable both for exploration and manipulation tasks in mobile robotics.

Abstract. Combine harvesters have a very important task in grain crop harvesting. Combine harvesters must fulfil lot of technological requirements regarding the field conditions, crop type, as well as the quality performance and optimal costs. The problem is that it is the main harvesting technology of cereals in the world and it has seasonal character of work, which has to be done in a relatively short period of time with adequate costs. The aim of this paper is an evaluation of the economic parameters of combine harvesters in different working conditions. The study is based on exact cost analysis of combine harvesters of New Holland brand by different ages and different concepts of threshing. Data were collected during all working seasons of the combine. In average, there is a data set from 10 seasons. In evaluation there were 9 machines NH CR 9080 and another 9 machines NH CX 8080. The working parameters evaluated are fuel consumption and operational costs. Performance of the machines was measured as harvested per hectare per day and per whole season. Costs are calculated as fixed and variable. The maintenance and service costs, and fuel consumption per hectare harvested are deeply analyzed. The result shows that the effect of the cost analysis depends on the terms of annual performance of the machine. There is 2.1 l per hectare difference in the fuel consumption between tangential and axial threshing systems. A special result of this study is evidence of the spare part costs for every machine and setting up the individual coefficient of repairs.