Filetype
Robot skills systems are meant to reduce robot setup time for new manufacturing tasks. Yet, for dexterous, contact-rich tasks, it is often difficult to find the right skill parameters. One strategy is to learn these parameters by allowing the robot system to learn directly on the task. For a learning problem, a robot operator can typically specify the type and range of values of the parameters. Ne
We propose a generalisation of a behaviour tree and motiongenerator based robot arm policy representation for learning and solving tasks such as contact-rich tasks like peg insertion or pushing an object. We use planning to generate skill sequences needed to execute these tasks and rely on reinforcement learning to obtain parameters of the policy. We assume gaussian processes as a suitable method
We propose a pipeline to setup up robot systems for new tasks using symbolic planning, reasoning and reinforcement learning. Furthermore, we
We propose a novel multi-pose loss function to train a neural network for 6D pose estimation, using synthetic data and evaluating it on real images. Our loss is inspired by the VSD (Visible Surface Discrepancy) metric and relies on a differentiable renderer and CAD models. This novel multi-pose approach produces multiple weighted pose estimates to avoid getting stuck in local minima. Our method re
Purpose—Ischemic myocardial contracture (IMC) or ‘‘stoneheart’’ is a condition with rapid onset following circulatory death. It inhibits transplantability of hearts donated uponcirculatory death (DCD). We investigate the effectiveness of hemodynamic normalization upon withdrawal of life-sustaining therapy (WLST) in a large-animal controlled DCD model, with the hypothesis that reduction in cardiac
This work analyzes the nonlinear effects in the track and hold circuit applied in high-speed ADCs or RF sampling receiver (RX) front-ends. Non-ideal effects inside the main sampling NMOS switch are studied. Parasitic varactor and sampling on-resistance modulation effects are analyzed through frequency domain Volterra series and the EKV MOS transistor model. Polynomial curve fitting is applied show
In recent years the number of connected devices and the demand for high data-rates have been significantly increased. This enormous growth is more pronounced by the introduction of the Internet of things (IoT) in which several devices are interconnected to exchange data for various applications like smart homes and smart cities. Moreover, new applications such as eHealth, autonomous vehicles, and c
The notion of Cloud RAN is taking a prominent role in narrative for the next generation wireless infrastructure. It is also seen as a mean to industrial communication systems. In order to provide reliable wireless connectivity for industrial deployments, by conventional means, the cloud infrastructure needs to be reliable and incur little latency, which however, is contradictory to the stochastic
We address the problem of insufficient quality of robot simulators to produce precise sensor readings for joint positions, velocities and torques. Realistic simulations of sensor readings are particularly important for real time robot control laws and for data intensive Reinforcement Learning of robot movements in simulation. We systematically construct two architectures based on Long Short-Term M
Small autonomous surface vehicles (ASV) will need both teleoperation support and redundant positioning technology to comply with expected future regulations. When at sea, they are limited by a satellite communication link with low throughput. We have designed and implemented a graphical user interface (GUI) for teleoperation using a communication link with low throughput, and one positioning syste
In the context of future manufacturing lines, removing fixtures will be a fundamental step to increase the flexibility of autonomous systems in assembly and logistic operations. Vision-based 3D pose estimation is a necessity to accurately handle objects that might not be placed at fixed positions during the robot task execution. Industrial tasks bring multiple challenges for the robust pose estima
The cellular interior is a dense environment. Understanding how such an environment impacts the properties of proteins and other macromolecules, as well as how weak, non-specific interactions drive processes such as protein droplet formation through liquid-liquid phase separation, is a major challenge in biological physics. The complexity of this environment often makes experimental studies extrem
We have developed a technique to measure low-power electromagnetic fields from mm-wave devices non-intrusively by combining a metasurface, designed to absorb power and focus the radiated power in a thermally isolated region, with an infrared camera. The metasurface consists of thermally isolated elements of low mass and highly emissive material for maximal IR conversion of the incident wave. The I
Static analysis tools play a crucial role in software development by detecting bugs and vulnerabilities. However, running these tools separately from the code editing process often causes developers to switch contexts, which can reduce productivity. Previous work has shown how Reference Attribute Grammars (RAGs) can be used for declarative implementation of competitive tooling for intraprocedural
Formal models for the safety validation of autonomous vehicles have become increasingly important. To this end, we present a safety framework for longitudinal automated driving. This framework enables the calculation of minimum safe inter-vehicular distances for arbitrary ego vehicle control policies in a computationally efficient manner. We use this framework to enhance and generalize the Respons
The success in the adoption of autonomous vehicles is dependent on their ability to solve rarely occurring safety-critical corner cases. Vehicular communications (V2X) aim at improving safety and efficiency of autonomous driving by adding the capability of explicit inter-vehicular information exchange. We argue that V2X enables another important function, namely the support of ethical driving deci
Static code checkers are tools that help software engineers by automatically finding defects without executing the programs. These tools contain a set of detectors that rely on static program analyses to find common programming defects or to enforce coding guidelines.While existing code checker frameworks package a rich collection of detectors, aimed at common bug defects, the effort to adapt thes
Bayesian optimization is a widely used method for optimizing expensive black-box functions, with Expected Improvement being one of the most commonly used acquisition functions. In contrast, information-theoretic acquisition functions aim to reduce uncertainty about the function's optimum and are often considered fundamentally distinct from EI. In this work, we challenge this prevailing perspective
Recent work reported that simple Bayesian optimization methods perform well for high-dimensional real-world tasks, seemingly contradicting prior work and tribal knowledge. This paper investigates the 'why'. We identify fundamental challenges that arise in high-dimensional Bayesian optimization and explain why recent methods succeed. Our analysis shows that vanishing gradients caused by Gaussian pr
