arXiv subjects

Yanan Liu

Publications and source records attributed to Yanan Liu.

At least 37 records · Page 2Linked to original sources

arXiv2023

Fault-tolerant $H^\infty$ control for optical parametric oscillators with pumping fluctuations

Optical Parametric Oscillators (OPOs) have wide applications in quantum optics for generating squeezed states and developing advanced technologies. When the phase or/and the amplitude of the pumping field for an OPO have fluctuations due to fault signals, time-varying uncertainties will be introduced in the dynamic parameters of the system. In this paper, we investigate how to design a fault-tolerant $H^\infty$ controller for an OPO with a disturbance input and time-varying uncertainties, which can achieve the required $H^\infty$ performance of the quantum system. We apply robust $H^\infty$ control theory to a quantum system, and design a passive controller and an active controller based on the solutions to two Riccati equations. The passive controller has a simple structure and is easy to be implemented by using only passive optical components, while the active quantum controller may achieve improved performance. The control performance of the proposed two controllers and one controller that was designed without consideration of system uncertainties is compared by numerical simulations in a specific OPO, and the results show that the designed controllers work effectively for fluctuations in both the phase and amplitude of the pumping field.

quant-ph↗

arXiv2023

Greedy versus Map-based Optimized Adaptive Algorithms for random-telegraph-noise mitigation by spectator qubits

In a scenario where data-storage qubits are kept in isolation as far as possible, with minimal measurements and controls, noise mitigation can still be done using additional noise probes, with corrections applied only when needed. Motivated by the case of solid-state qubits, we consider dephasing noise arising from a two-state fluctuator, described by random telegraph process, and a noise probe which is also a qubit, a so-called spectator qubit (SQ). We construct the theoretical model assuming projective measurements on the SQ, and derive the performance of different measurement and control strategies in the regime where the noise mitigation works well. We start with the Greedy algorithm; that is, the strategy that always maximizes the data qubit coherence in the immediate future. We show numerically that this algorithm works very well, and find that its adaptive strategy can be well approximated by a simpler algorithm with just a few parameters. Based on this, and an analytical construction using Bayesian maps, we design a one-parameter ($Θ$) family of algorithms. In the asymptotic regime of high noise-sensitivity of the SQ, we show analytically that this $Θ$-family of algorithms reduces the data qubit decoherence rate by a divisor scaling as the square of this sensitivity. Setting $Θ$ equal to its optimal value, $Θ^\star$, yields the Map-based Optimized Adaptive Algorithm for Asymptotic Regime (MOAAAR). We show, analytically and numerically, that MOAAAR outperforms the Greedy algorithm, especially in the regime of high noise sensitivity of SQ.

quant-ph↗

arXiv2022

Egocentric Hand-object Interaction Detection

In this paper, we propose a method to jointly determine the status of hand-object interaction. This is crucial for egocentric human activity understanding and interaction. From a computer vision perspective, we believe that determining whether a hand is interacting with an object depends on whether there is an interactive hand pose and whether the hand is touching the object. Thus, we extract the hand pose, hand-object masks to jointly determine the interaction status. In order to solve the problem of hand pose estimation due to in-hand object occlusion, we use a multi-cam system to capture hand pose data from multiple perspectives. We evaluate and compare our method with the most recent work from Shan et al. \cite{Shan20} on selected images from EPIC-KITCHENS \cite{damen2018scaling} dataset and achieve $89\%$ accuracy on HOI (hand-object interaction) detection which is comparative to Shan's ($92\%$). However, for real-time performance, our method can run over $\textbf{30}$ FPS which is much more efficient than Shan's ($\textbf{1}\sim\textbf{2}$ FPS). A demo can be found from https://www.youtube.com/watch?v=XVj3zBuynmQ

cs.CV↗

arXiv2022

Searching for outliers in the Chandra Source Catalog

Astronomers are increasingly faced with a deluge of information, and finding worthwhile targets of study in the sea of data can be difficult. Outlier identification studies are a method that can be used to focus investigations by presenting a smaller set of sources that could prove interesting because they do not follow the trends of the underlying population. We apply a principal component analysis (PCA) and an unsupervised random forest algorithm (uRF) to sources from the Chandra Source Catalog v.2 (CSC2). We present 119 high-significance sources that appear in all repeated applications of our outlier identification algorithm (OIA). We analyse the characteristics of our outlier sources and cross-match them with the SIMBAD data base. Our outliers contain several sources that were previously identified as having unusual or interesting features by studies. This OIA leads to the identification of interesting targets that could motivate more detailed study.

astro-ph.HE↗

arXiv2022

On-Sensor Binarized Fully Convolutional Neural Network with A Pixel Processor Array

This work presents a method to implement fully convolutional neural networks (FCNs) on Pixel Processor Array (PPA) sensors, and demonstrates coarse segmentation and object localisation tasks. We design and train binarized FCN for both binary weights and activations using batchnorm, group convolution, and learnable threshold for binarization, producing networks small enough to be embedded on the focal plane of the PPA, with limited local memory resources, and using parallel elementary add/subtract, shifting, and bit operations only. We demonstrate the first implementation of an FCN on a PPA device, performing three convolution layers entirely in the pixel-level processors. We use this architecture to demonstrate inference generating heat maps for object segmentation and localisation at over 280 FPS using the SCAMP-5 PPA vision chip.

cs.CV↗

arXiv2021

Multi-Scale Feature Fusion: Learning Better Semantic Segmentation for Road Pothole Detection

This paper presents a novel pothole detection approach based on single-modal semantic segmentation. It first extracts visual features from input images using a convolutional neural network. A channel attention module then reweighs the channel features to enhance the consistency of different feature maps. Subsequently, we employ an atrous spatial pyramid pooling module (comprising of atrous convolutions in series, with progressive rates of dilation) to integrate the spatial context information. This helps better distinguish between potholes and undamaged road areas. Finally, the feature maps in the adjacent layers are fused using our proposed multi-scale feature fusion module. This further reduces the semantic gap between different feature channel layers. Extensive experiments were carried out on the Pothole-600 dataset to demonstrate the effectiveness of our proposed method. The quantitative comparisons suggest that our method achieves the state-of-the-art (SoTA) performance on both RGB images and transformed disparity images, outperforming three SoTA single-modal semantic segmentation networks.

cs.CV↗

arXiv2021

Fully-simulated Integration of Scamp5d Vision System and Robot Simulator

This paper proposed a fully-simulated environment by integrating an on-sensor visual computing device, SCAMP, and CoppeliaSim robot simulator via interface and remote API. Within this platform, a mobile robot obstacle avoidance and target navigation with pre-set barriers is exploited with on-sensor visual computing, where images are captured in a robot simulator and processed by an on-sensor processing server after being transferred. We made our developed platform and associated algorithms for mobile robot navigation available online.

cs.RO↗

arXiv2021

Signatures of elastoviscous buckling in the dilute rheology of stiff polymers

As a stiff polymer tumbles in shear flow, it experiences compressive viscous forces that can cause it to buckle and undergo a sequence of morphological transitions with increasing flow strength. We use numerical simulations to uncover the effects of these transitions on the steady shear rheology of a dilute suspension of stiff polymers. Our results agree with classic scalings for Brownian rods in relatively weak flows but depart from them above the buckling threshold. Signatures of elastoviscous buckling include enhanced shear thinning and an increase in the magnitude of normal stress differences. We discuss our findings in the light of past work on rigid Brownian rods and non-Brownian elastic fibers and highlight the subtle role of thermal fluctuations in triggering instabilities.

physics.flu-dyn↗

arXiv2021

Direct Servo Control from In-Sensor CNN Inference with A Pixel Processor Array

This work demonstrates direct visual sensory-motor control using high-speed CNN inference via a SCAMP-5 Pixel Processor Array (PPA). We demonstrate how PPAs are able to efficiently bridge the gap between perception and action. A binary Convolutional Neural Network (CNN) is used for a classic rock, paper, scissors classification problem at over 8000 FPS. Control instructions are directly sent to a servo motor from the PPA according to the CNN's classification result without any other intermediate hardware.

cs.CV↗

arXiv2021

Bringing A Robot Simulator to the SCAMP Vision System

This work develops and demonstrates the integration of the SCAMP-5d vision system into the CoppeliaSim robot simulator, creating a semi-simulated environment. By configuring a camera in the simulator and setting up communication with the SCAMP python host through remote API, sensor images from the simulator can be transferred to the SCAMP vision sensor, where on-sensor image processing such as CNN inference can be performed. SCAMP output is then fed back into CoppeliaSim. This proposed platform integration enables rapid prototyping validations of SCAMP algorithms for robotic systems. We demonstrate a car localisation and tracking task using this proposed semi-simulated platform, with a CNN inference on SCAMP to command the motion of a robot. We made this platform available online.

cs.RO↗

arXiv2020

Loop-box: Multi-Agent Direct SLAM Triggered by Single Loop Closure for Large-Scale Mapping

In this paper, we present a multi-agent framework for real-time large-scale 3D reconstruction applications. In SLAM, researchers usually build and update a 3D map after applying non-linear pose graph optimization techniques. Moreover, many multi-agent systems are prevalently using odometry information from additional sensors. These methods generally involve intensive computer vision algorithms and are tightly coupled with various sensors. We develop a generic method for the keychallenging scenarios in multi-agent 3D mapping based on different camera systems. The proposed framework performs actively in terms of localizing each agent after the first loop closure between them. It is shown that the proposed system only uses monocular cameras to yield real-time multi-agent large-scale localization and 3D global mapping. Based on the initial matching, our system can calculate the optimal scale difference between multiple 3D maps and then estimate an accurate relative pose transformation for large-scale global mapping.

cs.RO↗

arXiv2020

Agile Reactive Navigation for A Non-Holonomic Mobile Robot Using A Pixel Processor Array

This paper presents an agile reactive navigation strategy for driving a non-holonomic ground vehicle around a preset course of gates in a cluttered environment using a low-cost processor array sensor. This enables machine vision tasks to be performed directly upon the sensor's image plane, rather than using a separate general-purpose computer. We demonstrate a small ground vehicle running through or avoiding multiple gates at high speed using minimal computational resources. To achieve this, target tracking algorithms are developed for the Pixel Processing Array and captured images are then processed directly on the vision sensor acquiring target information for controlling the ground vehicle. The algorithm can run at up to 2000 fps outdoors and 200fps at indoor illumination levels. Conducting image processing at the sensor level avoids the bottleneck of image transfer encountered in conventional sensors. The real-time performance of on-board image processing and robustness is validated through experiments. Experimental results demonstrate that the algorithm's ability to enable a ground vehicle to navigate at an average speed of 2.20 m/s for passing through multiple gates and 3.88 m/s for a 'slalom' task in an environment featuring significant visual clutter.

cs.RO↗

arXiv2020

Optimised hyperbolic microchannels for the mechanical characterisation of bio-particles

The transport of bio-particles in viscous flows exhibits a rich variety of dynamical behaviour, such as morphological transitions, complex orientation dynamics or deformations. Characterising such complex behaviour under well controlled flows is key to understanding the microscopic mechanical properties of biological particles as well as the rheological properties of their suspensions. While generating regions of simple shear flow in microfluidic devices is relatively straightforward, generating straining flows in which the strain rate is maintained constant for a sufficiently long time to observe the objects' morphologic evolution is far from trivial. In this work, we propose an innovative approach based on optimised design of microfluidic converging-diverging channels coupled with a microscope-based tracking method to characterise the dynamic behaviour of individual bio-particles under homogeneous straining flow. The tracking algorithm, combining a motorised stage and microscopy imaging system controlled by external signals, allows us to follow individual bio-particles transported over long-distances with high-quality images. We demonstrate experimentally the ability of the numerically optimised microchannels to provide linear velocity streamwise gradients along the centreline of the device, allowing for extended consecutive regions of homogeneous elongation and compression. We selected three test cases (DNA, actin filaments and protein aggregates) to highlight the ability of our approach for investigating the dynamics of objects with a wide range of sizes, characteristics and behaviours of relevance in the biological world.

cond-mat.soft↗

arXiv2020

Fault-tolerant Coherent H-infinity Control for Linear Quantum Systems

Robustness and reliability are two key requirements for developing practical quantum control systems. The purpose of this paper is to design a coherent feedback controller for a class of linear quantum systems suffering from Markovian jumping faults so that the closed-loop quantum system has both fault tolerance and H-infinity disturbance attenuation performance. This paper first extends the physical realization conditions from the time-invariant case to the time-varying case for linear stochastic quantum systems. By relating the fault tolerant H-infinity control problem to the dissipation properties and the solutions of Riccati differential equations, an H-infinity controller for the quantum system is then designed by solving a set of linear matrix inequalities (LMIs). In particular, an algorithm is employed to introduce additional noises and to construct the corresponding input matrices to ensure the physical realizability of the quantum controller. For real applications of the developed fault-tolerant control strategy, we present a linear quantum system example from quantum optics, where the amplitude of the pumping field randomly jumps among different values. It is demonstrated that a quantum H-infinity controller can be designed and implemented using some basic optical components to achieve the desired control goal.

quant-ph↗

arXiv2019

Flexible filaments buckle into helicoidal shapes in strong compressional flows

The occurrence of coiled or helical morphologies is common in nature, from plant roots to DNA packaging into viral capsids, as well as in applications such as oil drilling processes. In many examples, chiral structures result from the buckling of a straight fiber either with intrinsic twist or to which end moments have been applied in addition to compression forces. Here, we elucidate a generic way to form regular helicoidal shapes from achiral straight filaments transported in viscous flows with free ends. Through a combination of experiments using fluorescently labeled actin filaments in microfluidic divergent flows and of two distinct sets of numerical simulations, we demonstrate the robustness of helix formation. A nonlinear stability analysis is performed and explains the emergence of such chiral structures from the nonlinear interaction of perpendicular planar buckling modes, an effect that solely requires a strong compressional flow, independent of the exact nature of the fiber or type of flow field. The fundamental mechanism for the uncovered morphological transition and characterization of the emerging conformations advance our understanding of several biological and industrial processes and can also be exploited for the controlled microfabrication of chiral objects.

physics.flu-dyn↗

arXiv2019

Optical deep learning nano-profilometry

Determining the dimensions of nanostructures is critical to ensuring the maximum performance of many geometry-sensitive nanoscale functional devices. However, accurate metrology at the nanoscale is difficult using optics-based methods due to the diffraction limit. In this article, we propose an optical nano-profilometry framework with convolutional neural networks, which can retrieve deep sub-wavelength geometrical profiles of nanostructures from their optical images or scattering spectra. The generality, efficiency, and accuracy of the proposed framework are validated by performing two different measurements on three distinct nanostructures. We believe this work may catalyze more explorations of optics-based nano-metrology with deep learning.

physics.app-ph↗

arXiv2019

Representation Learning for Classical Planning from Partially Observed Traces

Specifying a complete domain model is time-consuming, which has been a bottleneck of AI planning technique application in many real-world scenarios. Most classical domain-model learning approaches output a domain model in the form of the declarative planning language, such as STRIPS or PDDL, and solve new planning instances by invoking an existing planner. However, planning in such a representation is sensitive to the accuracy of the learned domain model which probably cannot be used to solve real planning problems. In this paper, to represent domain models in a vectorization representation way, we propose a novel framework based on graph neural network (GNN) integrating model-free learning and model-based planning, called LP-GNN. By embedding propositions and actions in a graph, the latent relationship between them is explored to form a domain-specific heuristics. We evaluate our approach on five classical planning domains, comparing with the classical domain-model learner ARMS. The experimental results show that the domain models learned by our approach are much more effective on solving real planning problems.

cs.AI↗

arXiv2018

Morphological transitions of elastic filaments in shear flow

The morphological dynamics, instabilities and transitions of elastic filaments in viscous flows underlie a wealth of biophysical processes from flagellar propulsion to intracellular streaming, and are also key to deciphering the rheological behavior of many complex fluids and soft materials. Here, we combine experiments and computational modeling to elucidate the dynamical regimes and morphological transitions of elastic Brownian filaments in a simple shear flow. Actin filaments are employed as an experimental model system and their conformations are investigated through fluorescence microscopy in microfluidic channels. Simulations matching the experimental conditions are also performed using inextensible Euler-Bernoulli beam theory and non-local slender-body hydrodynamics in the presence of thermal fluctuations, and agree quantitatively with observations. We demonstrate that filament dynamics in this system is primarily governed by a dimensionless elasto-viscous number comparing viscous drag forces to elastic bending forces, with thermal fluctuations only playing a secondary role. While short and rigid filaments perform quasi-periodic tumbling motions, a buckling instability arises above a critical flow strength. A second transition to strongly-deformed shapes occurs at a yet larger value of the elasto-viscous number and is characterized by the appearance of localized high-curvature bends that propagate along the filaments in apparent "snaking" motions. A theoretical model for the so far unexplored onset of snaking accurately predicts the transition and explains the observed dynamics. For the first time, we present a complete characterization of filament morphologies and transitions as a function of elasto-viscous number and scaled persistence length and demonstrate excellent agreement between theory, experiments and simulations.

physics.flu-dyn↗