arXiv subjects

Qi Sun

Publications and source records attributed to Qi Sun.

At least 181 records · Page 10Linked to original sources

arXiv2020

Centralized Coordination of Connected Vehicles at Intersections using Graphical Mixed Integer Optimization

This paper proposes a centralized multi-vehicle coordination scheme serving unsignalized intersections. The whole process consists of three stages: a) target velocity optimization: formulate the collision-free vehicle coordination as a Mixed Integer Linear Programming (MILP) problem, with each incoming lane representing an independent variable; b) dynamic vehicle selection: build a directed graph with result of the optimization, and reserve only some of the vehicle nodes to coordinate by applying a subset extraction algorithm; c) synchronous velocity profile planning: bridge the gap between current speed and optimal velocity in a synchronous manner. The problem size is essentially bounded by number of lanes instead of vehicles. Thus the optimization process is realtime with guaranteed solution quality. Simulation has verified efficiency and real-time performance of the scheme.

eess.SY↗

arXiv2020

Deep Multi Depth Panoramas for View Synthesis

We propose a learning-based approach for novel view synthesis for multi-camera 360$^{\circ}$ panorama capture rigs. Previous work constructs RGBD panoramas from such data, allowing for view synthesis with small amounts of translation, but cannot handle the disocclusions and view-dependent effects that are caused by large translations. To address this issue, we present a novel scene representation - Multi Depth Panorama (MDP) - that consists of multiple RGBD$α$ panoramas that represent both scene geometry and appearance. We demonstrate a deep neural network-based method to reconstruct MDPs from multi-camera 360$^{\circ}$ images. MDPs are more compact than previous 3D scene representations and enable high-quality, efficient new view rendering. We demonstrate this via experiments on both synthetic and real data and comparisons with previous state-of-the-art methods spanning both learning-based approaches and classical RGBD-based methods.

cs.CV↗

arXiv2020

Mixed Reinforcement Learning with Additive Stochastic Uncertainty

Reinforcement learning (RL) methods often rely on massive exploration data to search optimal policies, and suffer from poor sampling efficiency. This paper presents a mixed reinforcement learning (mixed RL) algorithm by simultaneously using dual representations of environmental dynamics to search the optimal policy with the purpose of improving both learning accuracy and training speed. The dual representations indicate the environmental model and the state-action data: the former can accelerate the learning process of RL, while its inherent model uncertainty generally leads to worse policy accuracy than the latter, which comes from direct measurements of states and actions. In the framework design of the mixed RL, the compensation of the additive stochastic model uncertainty is embedded inside the policy iteration RL framework by using explored state-action data via iterative Bayesian estimator (IBE). The optimal policy is then computed in an iterative way by alternating between policy evaluation (PEV) and policy improvement (PIM). The convergence of the mixed RL is proved using the Bellman's principle of optimality, and the recursive stability of the generated policy is proved via the Lyapunov's direct method. The effectiveness of the mixed RL is demonstrated by a typical optimal control problem of stochastic non-affine nonlinear systems (i.e., double lane change task with an automated vehicle).

eess.SY↗

arXiv2020

Integral-Type Event-Triggered Model Predictive Control of Nonlinear Systems with Additive Disturbance

This paper studies integral-type event-triggered model predictive control (MPC) of continuous-time nonlinear systems. An integral-type event-triggered mechanism is proposed by incorporating the integral of errors between the actual and predicted state sequences, leading to reduced average sampling frequency. Besides, a new and improved robustness constraint is introduced to handle the additive disturbance, rendering the MPC problem with a potentially enlarged initial feasible region. Furthermore, the feasibility of the designed MPC and the stability of the closed-loop system are rigorously investigated. Several sufficient conditions to guarantee these properties are established, which is related to factors such as the prediction horizon, the disturbance bound, the triggering level, and the contraction rate for the robustness constraint. The effectiveness of the proposed algorithm is illustrated by numerical examples and comparisons.

math.OC↗

arXiv2020

DiffTaichi: Differentiable Programming for Physical Simulation

We present DiffTaichi, a new differentiable programming language tailored for building high-performance differentiable physical simulators. Based on an imperative programming language, DiffTaichi generates gradients of simulation steps using source code transformations that preserve arithmetic intensity and parallelism. A light-weight tape is used to record the whole simulation program structure and replay the gradient kernels in a reversed order, for end-to-end backpropagation. We demonstrate the performance and productivity of our language in gradient-based learning and optimization tasks on 10 different physical simulators. For example, a differentiable elastic object simulator written in our language is 4.2x shorter than the hand-engineered CUDA version yet runs as fast, and is 188x faster than the TensorFlow implementation. Using our differentiable programs, neural network controllers are typically optimized within only tens of iterations.

cs.LG↗

arXiv2020

Hierarchical Reinforcement Learning for Self-Driving Decision-Making without Reliance on Labeled Driving Data

Decision making for self-driving cars is usually tackled by manually encoding rules from drivers' behaviors or imitating drivers' manipulation using supervised learning techniques. Both of them rely on mass driving data to cover all possible driving scenarios. This paper presents a hierarchical reinforcement learning method for decision making of self-driving cars, which does not depend on a large amount of labeled driving data. This method comprehensively considers both high-level maneuver selection and low-level motion control in both lateral and longitudinal directions. We firstly decompose the driving tasks into three maneuvers, including driving in lane, right lane change and left lane change, and learn the sub-policy for each maneuver. Then, a master policy is learned to choose the maneuver policy to be executed in the current state. All policies including master policy and maneuver policies are represented by fully-connected neural networks and trained by using asynchronous parallel reinforcement learners (APRL), which builds a mapping from the sensory outputs to driving decisions. Different state spaces and reward functions are designed for each maneuver. We apply this method to a highway driving scenario, which demonstrates that it can realize smooth and safe decision making for self-driving cars.

eess.SY↗

arXiv2019

Eccentricity Effects on Blur and Depth Perception

Foveation and focus cue are the two most discussed topics on vision in designing near-eye displays. Foveation reduces rendering load by omitting spatial details in the content that the peripheral vision cannot appreciate; Providing richer focal cue can resolve vergence-accommodation conflict thereby lessening visual discomfort in using near-eye displays. We performed two psychophysical experiments to investigate the relationship between foveation and focus cue. The first study measured blur discrimination sensitivity as a function of visual eccentricity, where we found discrimination thresholds significantly lower than previously reported. The second study measured depth discrimination threshold where we found a clear dependency on visual eccentricity. We discuss the results from the two studies and suggest further investigation.

cs.GR↗

arXiv2019

OpenMPL: An Open Source Layout Decomposer

Multiple patterning lithography has been widely adopted in advanced technology nodes of VLSI manufacturing. As a key step in the design flow, multiple patterning layout decomposition (MPLD) is critical to design closure. Due to the NP-hardness of the general decomposition problem, various efficient algorithms have been proposed with high quality solutions. However, with increasingly complicated design flow and peripheral processing steps, developing a high-quality layout decomposer becomes more and more difficult, slowing down the further advancement in this field. This paper presents OpenMPL [1], an open-source layout decomposition framework, with well-separated peripheral processing and the core solving steps. We demonstrate the flexibility of the framework with efficient implementations of various state-of-the-art algorithms, which enable us to reproduce most of the recent results on widely-recognized benchmarks. We believe OpenMPL can pave the road for developing layout decomposition engines and stimulate further researches on this problem.

cs.OH↗

arXiv2019

Interactions between non-resonant rf fields and atoms with strong spin-exchange collisions

We study the interactions between oscillating non-resonant rf fields and atoms with strong spin-exchange collisions in the presence of a weak dc magnetic field. We find that the atomic Larmor precession frequency shows a new functional form to the rf field parameters when the spin-exchange collision rate is tuned. In the weak rf field amplitude regime, a strong modification of atomic Larmor frequency appears when the spin-exchange rate is comparable to the rf field frequency. This new effect has been neglected before due to its narrow observation window. We compare the experimental results with density matrix calculations, and explain the data by an underdamped oscillator model. When the rf field amplitude is large, there is a minimum atomic gyromagnetic ratio point due to the rf photon dressing, and we find that strong spin-exchange interactions modify the position of such a point.

physics.atom-ph↗

arXiv2019

Nonlocal Neural Networks, Nonlocal Diffusion and Nonlocal Modeling

Nonlocal neural networks have been proposed and shown to be effective in several computer vision tasks, where the nonlocal operations can directly capture long-range dependencies in the feature space. In this paper, we study the nature of diffusion and damping effect of nonlocal networks by doing spectrum analysis on the weight matrices of the well-trained networks, and then propose a new formulation of the nonlocal block. The new block not only learns the nonlocal interactions but also has stable dynamics, thus allowing deeper nonlocal structures. Moreover, we interpret our formulation from the general nonlocal modeling perspective, where we make connections between the proposed nonlocal network and other nonlocal models, such as nonlocal diffusion process and Markov jump process.

cs.LG↗

arXiv2018

Stochastic Training of Residual Networks: a Differential Equation Viewpoint

During the last few years, significant attention has been paid to the stochastic training of artificial neural networks, which is known as an effective regularization approach that helps improve the generalization capability of trained models. In this work, the method of modified equations is applied to show that the residual network and its variants with noise injection can be regarded as weak approximations of stochastic differential equations. Such observations enable us to bridge the stochastic training processes with the optimal control of backward Kolmogorov's equations. This not only offers a novel perspective on the effects of regularization from the loss landscape viewpoint but also sheds light on the design of more reliable and efficient stochastic training strategies. As an example, we propose a new way to utilize Bernoulli dropout within the plain residual network architecture and conduct experiments on a real-world image classification task to substantiate our theoretical findings.

cs.LG↗

arXiv2018

Deep Reinforcement Learning for Resource Management in Network Slicing

Network slicing is born as an emerging business to operators, by allowing them to sell the customized slices to various tenants at different prices. In order to provide better-performing and cost-efficient services, network slicing involves challenging technical issues and urgently looks forward to intelligent innovations to make the resource management consistent with users' activities per slice. In that regard, deep reinforcement learning (DRL), which focuses on how to interact with the environment by trying alternative actions and reinforcing the tendency actions producing more rewarding consequences, is assumed to be a promising solution. In this paper, after briefly reviewing the fundamental concepts of DRL, we investigate the application of DRL in solving some typical resource management for network slicing scenarios, which include radio resource slicing and priority-based core network slicing, and demonstrate the advantage of DRL over several competing schemes through extensive simulations. Finally, we also discuss the possible challenges to apply DRL in network slicing from a general perspective.

cs.NI↗

arXiv2018

Deep Learning-based Intelligent Dual Connectivity for Mobility Management in Dense Network

Ultra-dense network deployment has been proposed as a key technique for achieving capacity goals in the fifth-generation (5G) mobile communication system. However, the deployment of smaller cells inevitably leads to more frequent handovers, thus making mobility management more challenging and reducing the capacity gains offered by the dense network deployment. In order to fully reap the gains for mobile users in such a network environment, we propose an intelligent dual connectivity mechanism for mobility management through deep learning-based mobility prediction. We first use LSTM (Long Short Term Memory) algorithm, one of deep learning algorithms, to learn every user equipment's (UE's) mobility pattern from its historical trajectories and predict its movement trends in the future. Based on the corresponding prediction results, the network will judge whether a handover is required for the UE. For the handover case, a dual connection will be established for the related UE. Thus, the UE can get the radio signal from two base stations in the handover process. Simulation results verify that the proposed intelligent dual connectivity mechanism can significantly improve the quality of service of mobile users in the handover process while guaranteeing the network energy efficiency.

cs.NI↗

arXiv2017

Mapping electronic states of dual-parallel and symmetric zigzag grain boundaries of graphene on highly oriented pyrolytic graphite

The grain boundaries (GBs) of a graphene surface were extensively studied because GBs with specific defect configurations result in the formation of new curved structures, which can be treated as new carbon allotropes. We studied the structures and electronic spectra of two periodic GBs in graphene on highly oriented pyrolytic graphite (HOPG) surfaces using scanning tunneling microscopy and spectroscopy (STM/S). Our results demonstrated that a GB consisting of dual parallel periodic dislocation cores of pentagonal-heptagonal (5-7) carbon rings gives rise to an enhanced localized state at 0.45 eV above the Dirac point in graphene surfaces, which is attributed to van Hove singularities (VHSs). Moreover, the energy positions of the localized states are varied between 0.40 and 0.47 eV depending on the site and eventually decayed to 0.36 eV. The variation of the energy positions is induced by two parallel GBs because of the higher electron density at the GB as a defect center and the lower electron away from the GB. Another periodic GB with a symmetric zigzag structural feature induced VHSs at -0.038 and 0.12 eV near the Fermi level. Moreover, intervalley scattering was observed on both these GBs. This means that carrier concentration and thus, the conductance around the periodic GBs can be significantly enhanced by localized density of states. This finding suggests that graphene with a properly embedded ordered GB is promising for improving the performance of graphene-based electronic devices.

cond-mat.mtrl-sci↗

arXiv2016

Application of Non-orthogonal Multiple Access in LTE and 5G Networks

As the latest member of the multiple access family, non-orthogonal multiple access (NOMA) has been recently proposed for 3GPP Long Term Evolution (LTE) and envisioned to be an essential component of 5th generation (5G) mobile networks. The key feature of NOMA is to serve multiple users at the same time/frequency/code, but with different power levels, which yields a significant spectral efficiency gain over conventional orthogonal MA. This article provides a systematic treatment of this newly emerging technology, from its combination with multiple-input multiple-output (MIMO) technologies, to cooperative NOMA, as well as the interplay between NOMA and cognitive radio. This article also reviews the state of the art in the standardization activities concerning the implementation of NOMA in LTE and 5G networks.

cs.IT↗

arXiv2016

Multilevel Monte Carlo Finite Element Method for A Stochastic Optimal Control Problem

In this paper, we consider the implementation of multi-level Monte Carlo method to a stochastic optimal control problem with log-normal coefficients and its surrogate model problem. From the perspective of two optimization problems, i.e., minimizing the accuracy using a fixed computational cost and minimizing the total computational cost to attain a given accuracy, we derive formulas to determine the optimal sample sizes for each level of multi-level Monte Carlo method. Furthermore, we put forward the multi-level Monte Carlo algorithm for our stochastic optimal control problem and some tricks to deal with the multi-level log-normal coefficients. Finally, we present the numerical results of both the elliptic SPDEs and our control problem to validate the effectiveness over the traditional Monte Carlo method.

math.OC↗

arXiv2015

Full Duplex Networking: Mission Impossible?

Mobile traffic is projected to increase 1000 times from 2010 to 2020. This poses significant challenges on the 5th generation (5G) wireless communication system design, including network structure, air interface, key transmission schemes, multiple access, and duplexing schemes. In this paper, full duplex networking issues are discussed, aiming to provide some insights on the design and possible future deployment for 5G. Particularly, the interference scenarios in full duplex are analyzed, followed by discussions on several candidate interference mitigation approaches, interference proof frame structures, transceiver structures for channel reciprocity recovery, and super full duplex base station where each sector operates in time division duplex (TDD) mode. The extension of TDD and frequency division duplex (FDD) to full duplex is also examined. It is anticipated that with future standardization and deployment of full duplex systems, TDD and FDD will be harmoniously integrated, supporting all the existing half duplex mobile phones efficiently, and leading to a substantially enhanced 5G system performance.

cs.IT↗

arXiv2015

Ultra-fast Multiple Genome Sequence Matching Using GPU

In this paper, a contrastive evaluation of massively parallel implementations of suffix tree and suffix array to accelerate genome sequence matching are proposed based on Intel Core i7 3770K quad-core and NVIDIA GeForce GTX680 GPU. Besides suffix array only held approximately 20%~30% of the space relative to suffix tree, the coalesced binary search and tile optimization make suffix array clearly outperform suffix tree using GPU. Consequently, the experimental results show that multiple genome sequence matching based on suffix array is more than 99 times speedup than that of CPU serial implementation. There is no doubt that massively parallel matching algorithm based on suffix array is an efficient approach to high-performance bioinformatics applications.

cs.DS↗