Curiosity driven reinforcement learning for motion planning on humanoids

Most previous work on artificial curiosity (AC) and intrinsic motivation focuses on basic concepts and theory. Experimental results are generally limited to toy scenarios, such as navigation in a simulated maze, or control of a simple mechanical system with one or two degrees of freedom. To study AC in a more realistic setting, we embody a curious agent in the complex iCub humanoid robot. Our novel reinforcement learning (RL) framework consists of a state-of-the-art, low-level, reactive control layer, which controls the iCub while respecting constraints, and a high-level curious agent, which explores the iCub's state-action space through information gain maximization, learning a world model from experience, controlling the actual iCub hardware in real-time. To the best of our knowledge, this is the first ever embodied, curious agent for real-time motion planning on a humanoid. We demonstrate that it can learn compact Markov models to represent large regions of the iCub's configuration space, and that the iCub explores intelligently, showing interest in its physical constraints as well as in objects it finds in its environment.

System design for a small autonomous helicopter

The detailed system design of a small experimental autonomous helicopter is described. The system requires no ground-to-helicopter communications and hence all automation hardware is on-board the helicopter. All elements of the system are described including the control computer, the flight computer (the helicopter-to-control-computer interface), the sensors and the software. A number of critical implementation issues are also discussed.

3D reconfigurable autopilot flight system for both general aviation and unmanned aerial systems

This research project investigated and designed a modular architecture for a 3D Reconfigurable Autopilot Flight System that could be used to control actuators in both manned and unmanned aircraft. The system is based on a CAN Bus interface and allows seamless control of different types of actuators. During the course of the research the differences and similarities of autopilots for fixed-wing general aviation aircraft and unmanned aircraft were analysed focusing on the actuator interfaces. This project suggests that software and hardware used in commercial-of-the-shelf avionics could be used in manned and unmanned aviation.

Design and control of MIRA: A lightweight climbing robot for ship inspection

The inspection of marine vessels is currently performed manually. Inspectors use tools (e.g. cameras and devices for non-destructive testing) to detect damaged areas, cracks, and corrosion in large cargo holds, tanks, and other parts of a ship. Due to the size and complex geometry of most ships, ship inspection is time-consuming and expensive. The EU-funded project INCASS develops concepts for a marine inspection robotic assistant system to improve and automate ship inspections. In this paper, we introduce our magnetic wall–climbing robot: Marine Inspection Robotic Assistant (MIRA). This semiautonomous lightweight system is able to climb a vessels steel frame to deliver on-line visual inspection data. In addition, we describe the design of the robot and its building subsystems as well as its hardware and software components.

Approximate nearest-neighbour search with inverted signature slice lists

In this paper we present an original approach for finding approximate nearest neighbours in collections of locality-sensitive hashes. The paper demonstrates that this approach makes high-performance nearest-neighbour searching feasible on Web-scale collections and commodity hardware with minimal degradation in search quality.

MATLAB/Simulink simulation and C2000 code generation model for a grid-interactive 3-phase 4-wire inverter

This work is a MATLAB/Simulink model of a controller for a three-phase, four-wire, grid-interactive inverter. The model provides capacity for simulating the performance of power electroinic hardware, as well as code generation for an embedded controller. The implemented hardware topology is a three-leg bridge with a neutral connection to the centre-tap of the DC bus. An LQR-based current controller and MAF-based phase detector are implemented. The model is configured for code generation for a Texas Instruments TMS320F28335 Digital Signal Processor (DSP).

India’s indigenization of military aircraft design and manufacturing: Towards a fifth-generation fighter

India’s desire to transform itself into an international military power has brought about a rapid shift in its approach to procuring military hardware. The indigenization of India’s military manufacturing capacity forms an integral part of the strategic objectives of Indian military services, with its realization being a function of significant government investment in strategic technologies. This has a number of ramifications. An indigenous Indian military capacity, particularly in the field of aviation, forms a key part of India’s ambition of achieving regional air superiority, or even supremacy, and being capable of power projection. This is particularly in response to China’s increasing presence in South Asian airspace. A burgeoning Indian military manufacturing machine based on a comparative advantage in skilled technicians and lower-cost labour, together with strategic collaboration with foreign military hardware manufacturers, may also lead to neighbouring countries looking to India as a source of competitively priced military hardware. In short, this chapter seeks to analyse the rationale behind India’s attempt to become militarily self-sufficient in the field of aviation, discuss the technical, economic and political context in which it is achieving this transformation, and assess the potential outlook of success for India’s drive to achieve self-sufficiency in the arena of military aviation. This chapter will do so by using the case of India’s attempt to develop a fifth-generation fighter aircraft.

GPU accelerated algorithms for computing matrix function vector products with applications to exponential integrators and fractional diffusion

The efficient computation of matrix function vector products has become an important area of research in recent times, driven in particular by two important applications: the numerical solution of fractional partial differential equations and the integration of large systems of ordinary differential equations. In this work we consider a problem that combines these two applications, in the form of a numerical solution algorithm for fractional reaction diffusion equations that after spatial discretisation, is advanced in time using the exponential Euler method. We focus on the efficient implementation of the algorithm on Graphics Processing Units (GPU), as we wish to make use of the increased computational power available with this hardware. We compute the matrix function vector products using the contour integration method in [N. Hale, N. Higham, and L. Trefethen. Computing Aα, log(A), and related matrix functions by contour integrals. SIAM J. Numer. Anal., 46(5):2505–2523, 2008]. Multiple levels of preconditioning are applied to reduce the GPU memory footprint and to further accelerate convergence. We also derive an error bound for the convergence of the contour integral method that allows us to pre-determine the appropriate number of quadrature points. Results are presented that demonstrate the effectiveness of the method for large two-dimensional problems, showing a speedup of more than an order of magnitude compared to a CPU-only implementation.

A LIN inspired optical bus for signal isolation in multilevel or modular power electronic converters

Proposed in this paper is a low-cost, half-duplex optical communication bus for control signal isolation in modular or multilevel power electronic converters. The concept is inspired by the Local Interconnect Network (LIN) serial network protocol as used in the automotive industry. The proposed communications bus utilises readily available optical transceivers and is suitable for use with low-cost microcontrollers for distributed control of multilevel converters. As a signal isolation concept, the proposed optical bus enables very high cell count modular multilevel cascaded converters (MMCCs) for high-bandwidth, high-voltage and high-power applications. Prototype hardware is developed and the optical bus concept is validated experimentally in a 33-level MMCC converter operating at 120 Vrms and 60 Hz.

Efectividad clínica de la intervención terapéutica sobre la mano con realidad virtual en sujetos hemipléjicos: revisión sistemática

Objetivo Revisar de forma sistemática y realizar una evaluación critica de las diferentes publicaciones existentes sobre el uso de la realidad virtual como elemento rehabilitador en la mano. Estrategias de búsqueda Se ha realizado una búsqueda en la base de datos Medline completada con búsquedas en Google Schoolar. Se ha limitado la búsqueda a 10 años de extensión, idiomas: inglés y español. Selección de estudios: se ha centrado sobre aquellos artículos relacionados con la realidad virtual y la rehabilitación funcional de la mano en el accidente cerebrovascular. Síntesis de resultados Debido a la falta de homogeneidad, no ha sido posible evaluar metodológicamente mediante instrumento Caspe. Se detectaron 200 artículos de los cuales 12 cumplen los criterios de selección de los mismos. Dichos artículos han sido agrupados en relación con diferentes criterios como son: tipo de estudio, sistema de evaluación utilizado, periféricos usados o resultados obtenidos entre otros. Conclusiones No se ha obtenido una homogeneidad conjunta sobre los artículos, por ello se debe continuar realizando avances en este ámbito. A pesar de ello, los entornos de realidad virtual pueden ser un instrumento válido para la recuperación funcional de la mano en los accidentes cerebrovasculares, no siendo aptos como elemento único rehabilitador. Abstract in English Objective To make a systematic review and critical assessment of the different existing publications on the use of virtual reality as an element in rehabilitation of the hand. Search Method A search was made in the Medline database completed with searches in Google Schoolar. The search was limited to a period of ten years, and English and Spanish languages. Study selection The study was focused on those articles related to virtual reality and the functional rehabilitation of the hand in stroke. Synthesis of results It was not possible to perform a methodological evaluation using the Caspe instrument due to lack of homogeneity. Two hundred 200 articles were found. Of these, 12 met the criteria for their selection. These articles were grouped in relation to different criteria such as: type of study, assessment system used, hardware used or outcomes results obtained, among others. Conclusions Joint homogeneity was not obtained for the articles. Therefore, progress must still be made in this area. In spite of this, virtual reality environments can be a valid tool for functional recovery of the hand in strokes but they are not adequate for use as a single rehabilitative treatment element.

Identifying and forecasting the reverse salient in video game consoles: A performance gap ratio comparative analysis

This study uses the reverse salient methodology to contrast subsystems in video game consoles in order to discover, characterize, and forecast the most significant technology gap. We build on the current methodologies (Performance Gap and Time Gap) for measuring the magnitude of Reverse Salience, by showing the effectiveness of Performance Gap Ratio (PGR). The three subject subsystems in this analysis are the CPU Score, GPU core frequency, and video memory bandwidth. CPU Score is a metric developed for this project, which is the product of the core frequency, number of parallel cores, and instruction size. We measure the Performance Gap of each subsystem against concurrently available PC hardware on the market. Using PGR, we normalize the evolution of these technologies for comparative analysis. The results indicate that while CPU performance has historically been the Reverse Salient, video memory bandwidth has taken over as the quickest growing technology gap in the current generation. Finally, we create a technology forecasting model that shows how much the video RAM bandwidth gap will grow through 2019 should the current trend continue. This analysis can assist console developers in assigning resources to the next generation of platforms, which will ultimately result in longer hardware life cycles.

MegaVoltage Cone Beam Computed Tomography with a standard medical linear accelerator

Accurate patient positioning is vital for improved clinical outcomes for cancer treatments using radiotherapy. This project has developed Mega Voltage Cone Beam CT using a standard medical linear accelerator to allow 3D imaging of the patient position at treatment time with no additional hardware required. Providing 3D imaging functionality at no further cost allows enhanced patient position verification on older linear accelerators and in developing countries where access to new technology is limited.

Scalable and flexible large display arrays: A novel approach to the architectural enhancement of a prototype large display array

Large Display Arrays (LDAs) use Light Emitting Diodes (LEDs) in order to inform a viewing audience. A matrix of individually driven LEDs allows the area represented to display text, images and video. LDAs have undergone rapid development over the past 10 years in both the modular and semi-flexible formats. This thesis critically analyses the communication architecture and processor functionality of current LDAs and presents an alternative method, that is, Scalable Flexible Large Display Arrays (SFLDAs). SFLDAs are more adaptable to a variety of applications because of enhancements in scalability and flexibility. Scalability is the ability to configure SFLDAs from 0.8m2 to 200m2. Flexibility is increased functionality within the processors to handle changes in configuration and the use of a communication architecture that standardises two-way communication throughout the SFLDA. While common video platforms such as Digital Video Interface (DVI), Serial Digital Interface (SDI), and High Definition Multimedia Interface (HDMI) are considered as solutions for the communication architecture of SFLDAs, so too is modulation, fibre optic, capacitive coupling and Ethernet. From an analysis of these architectures, Ethernet was identified as the best solution. The use of Ethernet as the communication architecture in SFLDAs means that both hardware and software modules are capable of interfacing to the SFLDAs. The Video to Ethernet Processor Unit (VEPU), Scoreboard, Image and Control Software (SICS) and Ethernet to LED Processor Unit (ELPU) have been developed to form the key components in designing and implementing the first SFLDA. Data throughput rate and spectrophotometer tests were used to measure the effectiveness of Ethernet within the SFLDA constructs. The result of testing and analysis of these architectures showed that Ethernet satisfactorily met the requirements of SFLDAs.

A new single-phase frequency adaptive Phase-Locked Loop (PLL) for harmonic measurement

The paper presents an improved Phase-Locked Loop (PLL) for measuring the fundamental frequency and selective harmonic content of a distorted signal. This information can be used by grid interfaced devices and harmonic compensators. The single-phase structure is based on the Synchronous Reference Frame (SRF) PLL. The proposed PLL needs only a limited number of harmonic stages by incorporating Moving Average Filters (MAF) for eliminating the undesired harmonic content at each stage. The frequency dependency of MAF in effective filtering of undesired harmonics is also dealt with by a proposed method for adaptation to frequency variations of input signal. The method is suitable for high sampling rates and a wide frequency measurement range. Furthermore, an extended model of this structure is proposed which includes the response to both the frequency and phase angle variations. The proposed algorithm is simulated and verified using Hardware-in-the-Loop (HIL) testing.

Review and future options for computer modelling in the sugar industry

Computer modelling has been used extensively in some processes in the sugar industry to achieve significant gains. This paper reviews the investigations carried out over approximately the last twenty five years, including the successes but also areas where problems and delays have been encountered. In that time the capability of both hardware and software have increased dramatically. For some processes such as cane cleaning, cane billet preparation, and sugar drying, the application of computer modelling towards improved equipment design and operation has been quite limited. A particular problem has been the large number of particles and particle interactions in these…