A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2017; you can also visit the original URL.
The file type is application/pdf.
Filters
Iterative learning control for performance optimisation
2016
2016 American Control Conference (ACC)
Preserved Fulltext
Iterative learning control (ILC) is a popular design methodology to achieve high performance trajectory tracking of systems operating in a repetitive manner. ...
This paper further extends the applicability of ILC by showing that ILC can solve a more general problem of optimising some system performance for which trajectory tracking is just a special case. ...
CONCLUSION Iterative learning control is a popular control design method for high performance trajectory tracking of systems operating in a repetitive manner. ...
doi:10.1109/acc.2016.7525233
dblp:conf/amcc/Chu16
fatcat:c3slfsqbcvhl5i3t4zf5trdkfm
An extremum seeking approach to sampled-data iterative learning control of continuous-time nonlinear systems**This research was supported in part by the Institute for Mathematics and its Applications with funds provided by the National Science Foundation and the Australian Research Council
2016
IFAC-PapersOnLine
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
Iterative learning control (ILC) of continuous-time nonlinear plants with periodic sampled-data inputs is considered via an extremum seeking approach. ...
Abstract: Iterative learning control (ILC) of continuous-time nonlinear plants with periodic sampled-data inputs is considered via an extremum seeking approach. ...
Iterative learning control for reference tracking Iterative learning control (ILC) is a recursive learning based algorithm for solving the reference tracking problem (Moore, 1993; Moore et al., 1992; ...
doi:10.1016/j.ifacol.2016.10.292
fatcat:c3xfpgzi7zf63p2r4d57753f2i
Research on Parameter Optimization of Dynamic Priority Scheduling Algorithm Based on Improved Reinforcement Learning
2020
IET Generation, Transmission & Distribution
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Aiming to solve this problem, a dynamic priority scheduling algorithm based on improved reinforcement learning (RL) is proposed for parameter optimisation. ...
A scheduling algorithm optimised by RL can be better applied to industrial control and power system resource scheduling, which not only improves control efficiency but reduces scheduling costs. ...
The number of learning iterations of the improved Qlearning algorithm is reduced from 7867 for basic Q-learning to 116 iterations. ...
doi:10.1049/iet-gtd.2019.1468
fatcat:mlpojwx3yzeo3cqh4ajopfqfh4
Control as Hybrid Inference
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
CHI thus provides a principled framework for harnessing the sample efficiency of model-based planning while retaining the asymptotic performance of model-free policy optimisation. ...
Here, we unify these approaches by casting model-free policy optimisation as amortised variational inference, and model-based planning as iterative variational inference, within a 'control as hybrid inference ...
Leveraging these insights, we propose control as hybrid inference (CHI), a framework for combining amortised and iterative inference in the context of control. ...
arXiv:2007.05838v1
fatcat:5cjuh5sdt5futfxwqxhktpq67u
Adaptive Model Predictive Control by Learning Classifiers
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Other Versions
This is then integrated into a model predictive path integral control framework yielding robust controllers for a variety of challenging robotics tasks. ...
Stochastic model predictive control has been a successful and robust control framework for many robotics tasks where the system dynamics model is slightly inaccurate or in the presence of environment disturbances ...
This controller learning capability can be achieved with data-driven approaches for MPC optimisation (Görges, 2017) . ...
arXiv:2203.06783v2
fatcat:iwatvipiavcedlrihad2kchrhq
Fingerprint Policy Optimisation for Robust Reinforcement Learning
[article]
2019
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Note that this fulltext copy is not of the "primary" version of this work. The version it corresponds to is:
2018 pre-print arXiv:1805.10662v2 fatcat:ian5dppytjhpjcp6wncie7lnhmOther Versions
The central idea is to use Bayesian optimisation (BO) to actively select the distribution of the environment variable that maximises the improvement generated by each iteration of the policy gradient method ...
Our experiments show that FPO can efficiently learn policies that are robust to significant rare events, which are unlikely to be observable under random sampling, but are key to learning good policies ...
Acknowledgements We would like to thank Binxin Ru for sharing the code for FITBO, and Yarin Gal for the helpful discussions. ...
arXiv:1805.10662v3
fatcat:iyyzn3o4uja3bkk6tzhve2p2ke
Testing the Limits of SMILES-based De Novo Molecular Generation with Curriculum and Deep Reinforcement Learning
[article]
2022
bioRxiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
To help overcome these issues, we propose a new curriculum learning-inspired, recurrent Iterative Optimisation Procedure that enables the optimisation of generated molecules for seen and unseen molecular ...
Deep reinforcement learning methods have been shown to be potentially powerful tools for de novo design. ...
Iterative optimisation procedure for diversity control For de novo design tools to be effective, it should be possible to control the specificity of the generated molecules. ...
doi:10.1101/2022.07.15.500218
fatcat:rul72ktb5zdppeqcy7qtoornwi
Follow the Gradient: Crossing the Reality Gap using Differentiable Physics (RealityGrad)
[article]
2021
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
We demonstrate RealitGrad on a dynamic control task for a serial link robot manipulator and present results that show its efficiency and ability to quickly improve not just the robot's performance in real ...
We propose a novel iterative approach for crossing the reality gap that utilises live robot rollouts and differentiable physics. ...
[16] applied Iterative Linear Quadratic Control, on the linearized dynamics of a cartpole system for direct trajectory optimisation. ...
arXiv:2109.04674v1
fatcat:wh2pivygvbcj3e5ihwkkedyt2y
Bayesian functional optimisation with shape prior
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
Other Versions
We demonstrate the effectiveness of our approach for short polymer fibre design and optimising learning rate schedules for deep networks. ...
We develop a novel Bayesian optimisation framework for such functional optimisation of expensive black-box processes. ...
We apply our algorithm for learning rate schedule optimisation and found that an optimised learning rate schedule can even make SGD to perform better than both the method proposed by (Vien, Zimmermann ...
arXiv:1809.07260v2
fatcat:yreg3cb75bcwxhe77cmecy7a64
Learning control for best dynamic performance
2006
Journal of Biomechanics
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
In our study, a conceptual framework for efficient control synthesis in voluntary, ballistic-like movements is proposed. ...
A very important question is how that large-scale neuro-musculo-skeletal systems are efficiently controlled in dynamic motion tasks. ...
We propose a control learning approach that optimises the performance within a minimum number of test movements. It has the following main steps: 1. ...
doi:10.1016/s0021-9290(06)83276-8
fatcat:paikjmnbrnhwfnm4gczlzlhapm
FEREBUS: a high-performance modern Gaussian process regression engine
2023
Digital Discovery
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
FEREBUS is a highly optimised Gaussian process regression (GPR) engine, which provides both model and optimiser flexibility to produce tailored models designed for domain specific applications. ...
B. acknowledges the MRC National Productivity Investment Fund (NPIF) for the award of a PhD studentship. ...
For both UAPSO and IAPSO, a learning automaton is used to select the control parameters. ...
doi:10.1039/d2dd00082b
fatcat:yeeyhlvhird6bdz36qkq6xvwp4
Optimisation of Structured Neural Controller Based on Continuous-Time Policy Gradient
[article]
2022
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2022; you can also visit the original URL.
The file type is application/pdf.
Other Versions
performance provided by machine learning techniques. ...
Such a hybrid paradigm for fixed-structure control synthesis is particularly useful for optimising adaptive nonlinear controllers to achieve improved performance in online operation, an area where the ...
Specifically, this study addressed application of the CTPG method for performance optimisation of structured control systems prescribing a wellknown structure for the actual control input variable while ...
arXiv:2201.06262v4
fatcat:z5avwt3g4zhrrowe4vnynakkga
A Data-Driven Approach to Iterative Learning Control via Convex Optimization
2020
IET Control Theory & Applications
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2021; you can also visit the original URL.
The file type is application/pdf.
A new data-driven iterative learning control methodology is presented which uses the frequency response data of a system in order to avoid the problem of unmodelled dynamics associated with low-order parametric ...
A convex optimisation problem is formulated to design the learning filters such that the convergence criterion is minimised. ...
Iterative learning control (ILC) (which was first proposed in [18] ) seeks to address the tracking problem by implementing a learning algorithm where the desired performance is achieved by incorporating ...
doi:10.1049/iet-cta.2018.6446
fatcat:w4pqi5nncncr7bnx4itvprirz4
Optimisation of hand posture stimulation using an electrode array and iterative learning control
2013
Journal of Automatic Control
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2017; you can also visit the original URL.
The file type is application/pdf.
Nonlinear optimisation-based search algorithms have been developed for the precise stimulation of muscles in the wrist and hand, to enable stroke patients to attain predefined gestures. ...
Initial performance results from unimpaired subjects show the successful reproduction of six reference hand postures using the system. ...
The approach taken utilises iterative learning control (ILC), a control technique that was developed for industrial systems that repetitively complete the same movement. ...
doi:10.2298/jac1301001e
fatcat:rlfny4vtuzaybhzc3e3ncrs2qq
Amortized variance reduction for doubly stochastic objectives
[article]
2020
arXiv
pre-print
Preserved Fulltext
A copy of this work was available on the public web and has been preserved in the Wayback Machine. The capture dates from 2020; you can also visit the original URL.
The file type is application/pdf.
We propose a new approach in which we use a recognition network to cheaply approximate the optimal control variate for each mini-batch, with no additional model gradient computations. ...
We illustrate the properties of this proposal and test its performance on logistic regression and deep Gaussian processes. ...
For each period, we iteratively sample a gradient value then perform an optimisation step on the recognition network. ...
arXiv:2003.04125v1
fatcat:wqqkktmx7rapvilro4socxpywu
« Previous
Showing results 1 — 15 out of 48,130 results