No abstract

Abstract-Gaussian processes are gaining increasing popularity among the control community, in particular for the modelling of discrete time state space systems. However, it has not been clear how to incorporate model information, in the form of known state relationships, when using a Gaussian process as a predictive model. An obvious example of known prior information is position and velocity related states. Incorporation of such information would be beneficial both computationally and for faster dynamics learning. This paper introduces a method of achieving this, yielding faster dynamics learning and a reduction in computational effort from O Dn 2 to O (D − F )n 2 in the prediction stage for a system with D states, F known state relationships and n observations. The effectiveness of the method is demonstrated through its inclusion in the PILCO learning algorithm with application to the swing-up and balance of a torque-limited pendulum and the balancing of a robotic unicycle in simulation.

must be robust to changes in the local mapped environment. However, for many scenarios typical of UAV operations, it is still desirable to implement waypoint-driven guidance trajectories off-line to ensure optimality of the chosen flight profile. This is particularly true of lightweight micro-UAVs where optimal path planning is essential to maximise operational effectiveness against the low range and endurance induced by weight constraints on the powerplant.Path planning for UAVs, both on-line and off-line, has been an active research area for well over the past decade or more (3)(4)(5) . Many approaches have been proposed using techniques developed from a variety of different engineering and computing disciplines, however all techniques can be classified as belonging to one of two types: either a globally optimal or locally adaptive process. Locally adaptive algorithms are very popular with the mobile robotics community forming an integral part of the wider activity of selflocalisation and mapping (SLAM) (6) . Many such techniques are based upon graph theoretic algorithms originally developed within the artificial intelligence community such as Dijkstra, A*, D* etc. (7 8) and more recent graph or search-tree construction algorithms based on random searches such as the Rapidly-exploring Random Tree (RRT) algorithm (9) . However, tree construction and search can be a computationally intensive process if the desired path trajectories are of high dimension, for example when position, rate, attitude, orientation, time and other kinematic constraints are required. The resulting computational constraints mean that finding the optimal path by such methods is due more to luck than design.Global, optimal waypoint-driven path-planning techniques are ABSTRACTOperating micro-UAVs autonomously in complex urban areas requires that the guidance algorithms on-board are robust to changes in the operating environment. Limited endurance capability demands an optimal guidance algorithm, which will change as the environment does. All optimal path-planning routines are computationally intensive, with processor load a function of the environmental complexity. This paper presents a new algorithm, the reactive route selection algorithm, for storing a bank of optimal trajectories computed off-line and blending between these optimal trajectories as the operating environment changes. An example is presented using a mixed-integer linear program to generate the optimal trajectories.

Abstract-The contribution described in this paper is an algorithm for learning nonlinear, reference tracking, control policies given no prior knowledge of the dynamical system and limited interaction with the system through the learning process. Concepts from the field of reinforcement learning, Bayesian statistics and classical control have been brought together in the formulation of this algorithm which can be viewed as a form of indirect self tuning regulator. On the task of reference tracking using a simulated inverted pendulum it was shown to yield generally improved performance on the best controller derived from the standard linear quadratic method using only 30 s of total interaction with the system. Finally, the algorithm was shown to work on the simulated double pendulum proving its ability to solve nontrivial control tasks.

In the ancient city of Kyoto, contemporary artisans and designers are using heritage techniques and traditional clothing aesthetics to reinvent wafuku (Japanese clothing, including kimono) for modern life. Japan beyond the Kimono explores these shifts, highlighting developments in the Kyoto fashion industry such as its integration of digital weaving and printing techniques and the influence of social media on fashion distribution systems. Through case studies of designers, artisans, and retailers, Jenny Hall provides a comprehensive picture of the reasons behind the production and consumption of these rejuvenated fashion goods. She argues that conceptualisations of Japanese tradition include innovation and change, which is vital to understanding how Japanese cultural heritage is both sustained and evolving. Essential reading for students and scholars of fashion, anthropology, and Japanese studies, Jenny Hall’s sensory ethnography is the first of its kind, describing the lived experiences of people in the Kyoto textiles industry, explaining the renewal of traditional techniques and styles, and placing them both within contexts such as transnational ‘craftscapes’ and fast or slow fashion systems.