| 1. |
|
|
| 2. |
|
|
| 3. |
- Abdul Khader, Shahbaz, et al.
(författare)
-
Learning Deep Neural Policies with Stability Guarantees
-
Annan publikation (övrigt vetenskapligt/konstnärligt)abstract
- Deep reinforcement learning (DRL) has been successfully used to solve various robotic manipulation tasks. However, most of the existing works do not address the issue of control stability. This is in sharp contrast to the control theory community where the well-established norm is to prove stability whenever a control law is synthesized. What makes traditional stability analysis difficult for DRL are the uninterpretable nature of the neural network policies and unknown system dynamics. In this work, unconditional stability is obtained by deriving an interpretable deep policy structure based on the energy shaping control of Lagrangian systems. Then, stability during physical interaction with an unknown environment is established based on passivity. The result is a stability guaranteeing DRL in a model-free framework that is general enough for contact-rich manipulation tasks. With an experiment on a peg-in-hole task, we demonstrate, to the best of our knowledge, the first DRL with stability guarantee on a real robotic manipulator.
|
|
| 4. |
- Abdul Khader, Shahbaz, et al.
(författare)
-
Learning Stable Normalizing-Flow Control for Robotic Manipulation
-
Annan publikation (övrigt vetenskapligt/konstnärligt)abstract
- Reinforcement Learning (RL) of robotic manipu-lation skills, despite its impressive successes, stands to benefitfrom incorporating domain knowledge from control theory. Oneof the most important properties that is of interest is controlstability. Ideally, one would like to achieve stability guaranteeswhile staying within the framework of state-of-the-art deepRL algorithms. Such a solution does not exist in general,especially one that scales to complex manipulation tasks. Wecontribute towards closing this gap by introducing normalizing-flow control structure, that can be deployed in any latest deepRL algorithms. While stable exploration is not guaranteed,our method is designed to ultimately produce deterministiccontrollers with provable stability. In addition to demonstratingour method on challenging contact-rich manipulation tasks, wealso show that it is possible to achieve considerable explorationefficiency–reduced state space coverage and actuation efforts–without losing learning efficiency.
|
|
| 5. |
- Abdul Waheed, Malik, 1981-, et al.
(författare)
-
Generalized Architecture for a Real-time Computation of an Image Component Features on a FPGA
-
Annan publikation (övrigt vetenskapligt/konstnärligt)abstract
- This paper describes a generalized architecture for real-time component labeling and computation of image component features. Computing real-time image component features is one of the most important paradigms for modern machine vision systems. Embedded machine vision systems demand robust performance, power efficiency as well as minimum area utilization. The presented architecture can easily be extended with additional modules for parallel computation of arbitrary image component features. Hardware modules for component labeling and feature calculation run in parallel. This modularization makes the architecture suitable for design automation. Our architecture is capable of processing 390 video frames per second of size 640x480 pixels. Dynamic power consumption is 24.20mW at 86 frames per second on a Xilinx Spartran6 FPGA.
|
|
| 6. |
- Abdulrazaq, Muhammed, et al.
(författare)
-
Numerical modelling of the extensional dynamics in elastoviscoplastic fluids
-
Annan publikation (övrigt vetenskapligt/konstnärligt)abstract
- The extensional dynamics of an elasto-viscoplastic (EVP) fluid is studied by means of numerical simulations closely modelling an experimental configuration. Specifically, we track the interface between the EVP material and the Newtonian medium using an algebraic volume of fluid method (MTHINC-VOF) and employ a fully Eulerian immersed boundary method (IBM) to model the motion of the piston responsible of the extension of the material.We investigate the role of different values of the yield stress, surface tension at the interface between the EVP material and the surrounding fluid, polymer viscosity ratio, and extension rates on the necking thickness of the material, extensional viscosity, and yielding of the material. The results of the simulations reveal that when the yield stress of the EVP material is much larger than the viscous stresses, the material undergoes an elastic deformation, regardless of the selected values of extension rate, interfacial forces, and viscosity ratio. Moreover, increasing the ratio of the polymeric viscosity to the total viscosity of the system accelerates the EVP rupture due to the high stress concentration in the central part of the material sample. Specific and novel to our study, we show that interfacial forces cannot be ignored when the surface tension coefficient is such that a Capillary number based on the extensional rate is order 1. For large values of the surface tension coefficient, the EVP material fails sooner, with a clear deviation from the exponential reduction in the neck thickness.
|
|
| 7. |
- Abedin, Ahmad, et al.
(författare)
-
Growth of epitaxial SiGe alloys as etch-stop layers in germanium-on-insulator fabrication
-
Annan publikation (övrigt vetenskapligt/konstnärligt)abstract
- In this study, the application of epitaxially grown SixGe1-x films as etch stop layers in a germanium-on-insulator substrate fabrication flow is investigated. Layers with Ge contents from 15% to 70% were epitaxially grown on Si (1 0 0) using silane and germane. It was found that the Ge content in the films is independent of the growth temperature for fixed partial pressure ratios. At low growth temperatures the activation energy is found to be 1.8 eV which points to a hydrogen desorption limited growth rate mechanism. At growth temperatures of less than 500℃, the surface roughness is <1 nm. This surface roughness does not change when the films are grown on Ge substrates. Finally, a fully strained Si0.5Ge0.5 film was grown on Ge strain relaxed buffer at 450℃. This layer demonstrates etch selectivity of >400:1 towards Ge in diluted SC-1. This result enables the integration of the Si0.5Ge0.5 film as an etch stop layer for single crystalline germanium-on-insulator substrate fabrication.
|
|
| 8. |
|
|
| 9. |
|
|
| 10. |
|
|
