arXiv subjects

Nan Li

Publications and source records attributed to Nan Li.

At least 253 records · Page 14Linked to original sources

arXiv2022

Development of a deep learning platform for optimising sheet stamping geometries subject to manufacturing constraints

The latest sheet stamping processes enable efficient manufacturing of complex shape structural components that have high stiffness to weight ratios, but these processes can introduce defects. To assist component design for stamping processes, this paper presents a novel deep-learning-based platform for optimising 3D component geometries. The platform adopts a non-parametric modelling approach that is capable of optimising arbitrary geometries from multiple geometric parameterisation schema. This approach features the interaction of two neural networks: 1) a geometry generator and 2) a manufacturing performance evaluator. The generator predicts continuous 3D signed distance fields (SDFs) for geometries of different classes, and each SDF is conditioned on a latent vector. The zero-level-set of each SDF implicitly represents a generated geometry. Novel training strategies for the generator are introduced and include a new loss function which is tailored for sheet stamping applications. These strategies enable the differentiable generation of high quality, large scale component geometries with tight local features for the first time. The evaluator maps a 2D projection of these generated geometries to their post-stamping physical (e.g., strain) distributions. Manufacturing constraints are imposed based on these distributions and are used to formulate a novel objective function for optimisation. A new gradient-based optimisation technique is employed to iteratively update the latent vectors, and therefore geometries, to minimise this objective function and thus meet the manufacturing constraints. Case studies based on optimising box geometries subject to a sheet thinning constraint for a hot stamping process are presented and discussed. The results show that expressive geometric changes are achievable, and that these changes are driven by stamping performance.

cs.LG↗

arXiv2021

Eco-Coasting Strategies Using Road Grade Preview: Evaluation and Online Implementation Based on Mixed Integer Model Predictive Control

Coasting has been widely used in the eco-driving guidelines to reduce fuel consumption by profiting from kinetic energy. However, the comprehensive comparison between different coasting strategies and online performance of the eco-coasting strategy using road grade preview are still unclear because of the oversimplification and the integer variable in the optimal control problems. Herein, two different coasting strategies (fuel cut-off and engine start/stop) are proposed to reveal the potential benefit of eco-coasting using the road grade preview. Engine drag torque and energy cost used for engine restart are considered in the modeling to give a fair evaluation of the offline and online performance. The offline performance of these two coasting methods is evaluated through dynamic programming (DP) under various driving scenarios with different slope profiles. Offline simulation shows that the engine start/stop method outperforms the fuel cut-off method in terms of fuel consumption and travel time by getting rid of the engine drag torque. Then, online performance of these two coasting methods is evaluated using Mixed Integer Model Predictive Control (MIMPC). A novel operational constraint on the minimum off steps is added in the MIMPC formulation to avoid frequent switch of the integer variables which represent the fuel cut-off and the engine start/stop mechanism. Simulation results show that, for both fuel cut-off and engine start/stop coasting methods, the MPC controller reduces fuel consumption to a level comparable to DP without sacrificing the travel time.

eess.SY↗

arXiv2021

Continuous Optimization-Based Drift Counteraction Optimal Control: A Spacecraft Attitude Control Case Study

This paper presents a continuous optimization approach to DCOC and its application to spacecraft high-precision attitude control. The approach computes a control input sequence that maximizes the time-before-exit by solving a nonlinear programming problem with an exponentially weighted cost function and purely continuous variables. Based on results from sensitivity analysis and exact penalty method, we prove the optimality guarantee of our approach. The practical application of our approach is demonstrated through a spacecraft high-precision attitude control example. A nominal case with three functional reaction wheels (RWs) and an underactuated case with only two functional RWs were considered. Simulation results illustrate the effectiveness of our approach as a contingency method for extending spacecraft's effective mission time in the case of RW failures.

math.OC↗

arXiv2021

Interaction-Aware Trajectory Prediction and Planning for Autonomous Vehicles in Forced Merge Scenarios

Merging is, in general, a challenging task for both human drivers and autonomous vehicles, especially in dense traffic, because the merging vehicle typically needs to interact with other vehicles to identify or create a gap and safely merge into. In this paper, we consider the problem of autonomous vehicle control for forced merge scenarios. We propose a novel game-theoretic controller, called the Leader-Follower Game Controller (LFGC), in which the interactions between the autonomous ego vehicle and other vehicles with a priori uncertain driving intentions is modeled as a partially observable leader-follower game. The LFGC estimates the other vehicles' intentions online based on observed trajectories, and then predicts their future trajectories and plans the ego vehicle's own trajectory using Model Predictive Control (MPC) to simultaneously achieve probabilistically guaranteed safety and merging objectives. To verify the performance of LFGC, we test it in simulations and with the NGSIM data, where the LFGC demonstrates a high success rate of 97.5% in merging.

eess.SY↗

arXiv2021

Primordial black holes from the perturbations in the inflaton potential in peak theory

The primordial black hole (PBH) is an effective candidate for dark matter. In this work, the PBH abundance $f$ is calculated in peak theory, with one or two perturbations in the inflaton potential. We construct an antisymmetric perturbation that can create a perfect plateau in the inflaton potential, leading inflation to the ultra-slow-roll stage. During this stage, the power spectrum of primordial curvature perturbation is remarkably enhanced on small scales, generating abundant PBHs. The PBH abundance $f\sim 0.1$ can be achieved in one or two typical mass windows at $10^{-17}M_\odot$, $10^{-13}M_\odot$, and $30M_\odot$, without spoiling the nearly scale-invariant power spectrum on large scales. For comparison, $f$ is calculated in two approximate methods of peak theory (with different spectral moments) and also in the Press--Schechter theory. It is found that the Press--Schechter theory systematically underestimates $f$ by two or three orders of magnitude compared with peak theory.

astro-ph.CO↗

arXiv2021

Energy-Efficient Autonomous Driving Using Cognitive Driver Behavioral Models and Reinforcement Learning

Autonomous driving technologies are expected to not only improve mobility and road safety but also bring energy efficiency benefits. In the foreseeable future, autonomous vehicles (AVs) will operate on roads shared with human-driven vehicles. To maintain safety and liveness while simultaneously minimizing energy consumption, the AV planning and decision-making process should account for interactions between the autonomous ego vehicle and surrounding human-driven vehicles. In this chapter, we describe a framework for developing energy-efficient autonomous driving policies on shared roads by exploiting human-driver behavior modeling based on cognitive hierarchy theory and reinforcement learning.

cs.RO↗

arXiv2021

Hawking--Page phase transitions of the black holes in a cavity

The Hawking--Page phase transitions of the Schwarzschild and charged black holes are investigated in an extended phase space, in which the black holes are enclosed in a spherical cavity of radius $r_B$ in asymptotically flat space. An effective thermodynamic volume $V=4πr_B^3/3$ is introduced for the black hole, and an effective pressure $p$ is defined as the conjugate variable of $V$. The phase transition temperature $T_{\rm HP}$ and the Gibbs free energy $G$ are systematically studied in a grand canonical ensemble with fixed electric potential $Φ$, and $T_{\rm HP}$ is found to increase with $p$ and decrease with $Φ$. If the phase transition occurs, $Φ$ must have an upper bound, such that the cavity radius is always larger than the black hole horizon radius. These phase transition behaviors are further compared to those in the anti-de Sitter space, and the remarkable similarities and notable differences are also discussed in depth. Our work reveals the relationship of the thermodynamic properties of black holes and their specific boundary conditions in different extended phase spaces.

gr-qc↗

arXiv2021

Safe Learning Reference Governor: Theory and Application to Fuel Truck Rollover Avoidance

This paper proposes a learning reference governor (LRG) approach to enforce state and control constraints in systems for which an accurate model is unavailable, and this approach enables the reference governor to gradually improve command tracking performance through learning while enforcing the constraints during learning and after learning is completed. The learning can be performed either on a black-box type model of the system or directly on the hardware. After introducing the LRG algorithm and outlining its theoretical properties, this paper investigates LRG application to fuel truck (tank truck) rollover avoidance. Through simulations based on a fuel truck model that accounts for liquid fuel sloshing effects, we show that the proposed LRG can effectively protect fuel trucks from rollover accidents under various operating conditions.

eess.SY↗

arXiv2021

Strong lens modelling: comparing and combining Bayesian neural networks and parametric profile fitting

The vast quantity of strong galaxy-galaxy gravitational lenses expected by future large-scale surveys necessitates the development of automated methods to efficiently model their mass profiles. For this purpose, we train an approximate Bayesian convolutional neural network (CNN) to predict mass profile parameters and associated uncertainties, and compare its accuracy to that of conventional parametric modelling for a range of increasingly complex lensing systems. These include standard smooth parametric density profiles, hydrodynamical EAGLE galaxies and the inclusion of foreground mass structures, combined with parametric sources and sources extracted from the Hubble Ultra Deep Field. In addition, we also present a method for combining the CNN with traditional parametric density profile fitting in an automated fashion, where the CNN provides initial priors on the latter's parameters. On average, the CNN achieved errors 19 $\pm$ 22 per cent lower than the traditional method's blind modelling. The combination method instead achieved 27 $\pm$ 11 per cent lower errors over the blind modelling, reduced further to 37 $\pm$ 11 per cent when the priors also incorporated the CNN-predicted uncertainties, with errors also 17 $\pm$ 21 per cent lower than the CNN by itself. While the CNN is undoubtedly the fastest modelling method, the combination of the two increases the speed of conventional fitting alone by factors of 1.73 and 1.19 with and without CNN-predicted uncertainties, respectively. This, combined with greatly improved accuracy, highlights the benefits one can obtain through combining neural networks with conventional techniques in order to achieve an efficient automated modelling approach.

astro-ph.IM↗

arXiv2021

Cloud-Assisted Nonlinear Model Predictive Control for Finite-Duration Tasks

Cloud computing creates new possibilities for control applications by offering powerful computation and storage capabilities. In this paper, we propose a novel cloud-assisted model predictive control (MPC) framework in which we systematically fuse a cloud MPC that uses a high-fidelity nonlinear model but is subject to communication delays with a local MPC that exploits simplified dynamics (due to limited computation) but has timely feedback. Unlike traditional cloud-based control that treats the cloud as powerful, remote, and sole controller in a networked-system control setting, the proposed framework aims at seamlessly integrating the two controllers for enhanced performance. In particular, we formalize the fusion problem for finite-duration tasks by explicitly considering model mismatches and errors due to request-response communication delays. We analyze stability-like properties of the proposed cloud-assisted MPC framework and establish approaches to robustly handling constraints within this framework in spite of plant-model mismatch and disturbances. A fusion scheme is then developed to enhance control performance while satisfying stability-like conditions, the efficacy of which is demonstrated with multiple simulation examples, including an automotive control example to show its industrial application potentials.

eess.SY↗

arXiv2021

PyAutoLens: Open-Source Strong Gravitational Lensing

Strong gravitational lensing, which can make a background source galaxy appears multiple times due to its light rays being deflected by the mass of one or more foreground lens galaxies, provides astronomers with a powerful tool to study dark matter, cosmology and the most distant Universe. PyAutoLens is an open-source Python 3.6+ package for strong gravitational lensing, with core features including fully automated strong lens modeling of galaxies and galaxy clusters, support for direct imaging and interferometer datasets and comprehensive tools for simulating samples of strong lenses. The API allows users to perform ray-tracing by using analytic light and mass profiles to build strong lens systems. Accompanying PyAutoLens is the autolens workspace (see https://github.com/Jammy2211/autolens_workspace), which includes example scripts, lens datasets and the HowToLens lectures in Jupyter notebook format which introduce non experts to strong lensing using PyAutoLens. Readers can try PyAutoLens right now by going to the introduction Jupyter notebook on Binder (see https://mybinder.org/v2/gh/Jammy2211/autolens_workspace/master) or checkout the readthedocs (see https://pyautolens.readthedocs.io/en/latest/) for a complete overview of PyAutoLens's features.

astro-ph.IM↗

arXiv2021

A Novel Framework Integrating AI Model and Enzymological Experiments Promotes Identification of SARS-CoV-2 3CL Protease Inhibitors and Activity-based Probe

The identification of protein-ligand interaction plays a key role in biochemical research and drug discovery. Although deep learning has recently shown great promise in discovering new drugs, there remains a gap between deep learning-based and experimental approaches. Here we propose a novel framework, named AIMEE, integrating AI Model and Enzymology Experiments, to identify inhibitors against 3CL protease of SARS-CoV-2, which has taken a significant toll on people across the globe. From a bioactive chemical library, we have conducted two rounds of experiments and identified six novel inhibitors with a hit rate of 29.41%, and four of them showed an IC50 value less than 3 μM. Moreover, we explored the interpretability of the central model in AIMEE, mapping the deep learning extracted features to domain knowledge of chemical properties. Based on this knowledge, a commercially available compound was selected and proven to be an activity-based probe of 3CLpro. This work highlights the great potential of combining deep learning models and biochemical experiments for intelligent iteration and expanding the boundaries of drug discovery.

q-bio.MN↗

arXiv2021

Safe Reinforcement Learning Using Robust Action Governor

Reinforcement Learning (RL) is essentially a trial-and-error learning procedure which may cause unsafe behavior during the exploration-and-exploitation process. This hinders the application of RL to real-world control problems, especially to those for safety-critical systems. In this paper, we introduce a framework for safe RL that is based on integration of a RL algorithm with an add-on safety supervision module, called the Robust Action Governor (RAG), which exploits set-theoretic techniques and online optimization to manage safety-related requirements during learning. We illustrate this proposed safe RL framework through an application to automotive adaptive cruise control.

cs.LG↗

arXiv2021

Rapid feasibility assessment of components formed through hot stamping: A deep learning approach

The novel non-isothermal Hot Forming and cold die Quenching (HFQ) process can enable the cost-effective production of complex shaped, high strength aluminium alloy panel components. However, the unfamiliarity of designing for the new process prevents its widescale adoption in industrial settings. Recent research efforts focus on the development of advanced material models for finite element simulations, used to assess the feasibility of new component designs for the HFQ process. However, FE simulations take place late in design processes, require forming process expertise and are unsuitable for early-stage design explorations. To address these limitations, this study presents a novel application of a Convolutional Neural Network (CNN) based surrogate as a means of rapid manufacturing feasibility assessment for components to be formed using the HFQ process. A diverse dataset containing variations in component geometry, blank shapes, and processing parameters, together with corresponding physical fields is generated and used to train the model. The results show that near indistinguishable full field predictions are obtained in real time from the model when compared with HFQ simulations. This technique provides an invaluable tool to aid component design and decision making at the onset of a design process for complex-shaped components formed under HFQ conditions.

cs.CE↗

arXiv2021

Implications of a possible TeV break in the cosmic-ray electron and positron flux

A TeV spectral break in the total flux of cosmic-ray electrons and positrons (CREs) at which the spectral power index softens from $\sim 3$ to $\sim 4$ has been observed by H.E.S.S. and recently confirmed by DAMPE with a high significance of $6.6~σ$. Such an observation is apparently inconsistent with the data from other experiments such as Fermi-LAT, AMS-02 and CALET. We perform a global analysis to the latest CRE data including Fermi-LAT, AMS-02, CALET, DAMPE and H.E.S.S. with energy scale uncertainties taken into account to improve the consistency between the data sets. The fit result strongly favors the existence of the break at $\sim 1$ TeV with an even higher statistical significance of $13.3~σ$. In view of the tentative CRE break, we revisit a number of models of nearby sources, such as a single generic Pulsar Wind Nebula (PWN), known multiple PWNe from the ATNF catalog, and their combinations with either an additional Dark Matter (DM) component or a Supernova Remnant (SNR). We show that the CRE break at $\sim 1$ TeV, together with the known CR positron excess points towards the possibility that the nearby sources should be highly charge asymmetric. Among the models under consideration, the one with a PWN plus SNR is most favored by the current data. The favoured distance and age of the PWN and SNR sources are both within $0.6$ kpc and around $10^{5}$ yr respectively. Possible candidate sources include PSR J0954-5430, Vela and Monogem ring, etc. We find that for the models under consideration, the additional DM component is either unnecessary, or predicts too much photons in tension with the H.E.S.S. data of $γ$-rays from the direction of the Galactic Center. We also show that the current measurement of the anisotropies in the arrival direction of the CRE can be useful in determining the property of the sources.

astro-ph.HE↗

arXiv2021

The impact of line-of-sight structures on measuring $H_0$ with strong lensing time-delays

Measurements of The Hubble-Lemaitre constant from early- and local-universe observations show a significant discrepancy. In an attempt to understand the origin of this mismatch, independent techniques to measure H0 are required. One such technique, strong lensing time delays, is set to become a leading contender amongst the myriad methods due to forthcoming large strong lens samples. It is therefore critical to understand the systematic effects inherent in this method. In this paper, we quantify the influence of additional structures along the line-of-sight by adopting realistic light cones derived from the CosmoDC2 semi-analytical extra-galactic catalogue. Using multiple lens plane ray-tracing to create a set of simulated strong lensing systems, we have investigated the impact of line-of-sight structures on time-delay measurements and in turn, on the inferred value of H0. We have also tested the reliability of existing procedures for correcting for line-of-sight effects. We find that if the integrated contribution of the line-of-sight structures is close to a uniform mass sheet, the bias in H0 can be adequately corrected by including a constant external convergence $κ_{ext}$ in the lens model. However, for realistic line-of-sight structures comprising many galaxies at different redshifts, this simple correction overestimates the bias by an amount that depends linearly on the median external convergence. We, therefore, conclude that lens modelling must incorporate multiple lens planes to account for line-of-sight structures for accurate and precise inference of H0.

astro-ph.CO↗

arXiv2021

The high energy window of probing dark matter with cosmic-ray antideuterium and antihelium

Cosmic-ray (CR) anti-nuclei are often considered as important observables for dark matter (DM) indirect detections at low kinetic energies below GeV per nucleon. Since the primary CR fluxes drop quickly towards high energies, the secondary anti-nuclei in CR are expected to be significantly suppressed in high energy regions ($\gtrsim 100$ GeV per nucleon). If DM particles are heavy, the annihilation productions of DM can be highly boosted, thus the fluxes of anti-nuclei produced by DM annihilations may exceed the secondary background at high energies, which opens a high energy window for DM indirect detections. We investigate the possibility of detecting heavy DM particles which annihilate into high energy anti-nuclei. We use Monte-Carlo generators $\texttt{PYTHIA}$, $\texttt{EPOS-LHC}$ and $\texttt{DPMJET}$ and the coalescence model to simulate the production of anti-nuclei, and constrain the DM annihilation cross sections by using the AMS-02 and HAWC antiproton data and the HESS galactic center gamma-ray data. We find that the conclusion depends on the choice of DM density profiles. For the "Cored" type profile with a DM particle mass $\gtrsim 10$ TeV, the contributions from DM annihilations can exceed the secondary background in high energy regions, which opens the high energy window. While for the "Cuspy" type profile, the excess disappears.

hep-ph↗

arXiv2021

Optimization of graded filleted lattice structures subject to yield and buckling constraints

To reduce the stress concentration and ensure the structural safety for lattice structure designs, in this paper, a new optimization framework is developed for the optimal design of graded lattice structures, innovatively integrating fillet designs as well as yield and elastic buckling constraints. Both strut and fillet radii are defined as design variables. Homogenization method is employed to characterize the effective elastic constants and yield stresses of the lattice metamaterials. Metamaterial models are developed to represent the relationships between the metamaterial effective properties and lattice geometric variables. A yield constraint, based on the modified Hills yield criterion, is developed as a function of relative strut radii and fillet parameters. An elastic buckling constraint, based on the Euler buckling formula and the Johnson formula, is developed as a function of relative strut radii. Both yield and buckling constraints are integrated into an optimization problem formulation; a new optimization framework is proposed and a case study of minimizing the compliance of a Messerschmitt-Bolkow-Blohm beam is conducted. The yield and buckling constraints guarantee the safety of the optimized beams composed of BCC and PC lattices. Reductions in compliance and stress concentration are achieved by the optimized MBB beams.

cond-mat.mes-hall↗