检索结果-内蒙古大学图书馆

Intent-Guided Trajectory Prediction for E-Scooter Riders and Bicyclists

学校读者我要写书评

暂无评论

Intent-Guided Trajectory Prediction for E-Scooter Riders and...

International Conference on Intelligent Transportation

作者： Zhengming Zhang Zhengming Ding Yaobin Chen Stanley Chien Lingxi Li Rini Sheroy Joshua Domeyer Renran Tian School of Industrial Engineering Purdue University USA Department of Computer Science Tulane University USA School of Electrical and Computer Engineering Purdue University USA Toyota Collaborative Safety Research Center (CSRC) USA Department of Industrial and Systems Engineering North Carolina State University USA

ISBN: (数字)9798331505929

ISBN: (纸本)9798331505936

The rapidly increasing popularity of micromobility devices, such as e-scooters and bicycles, in urban environments underscores the critical need for advanced trajectory prediction models to enhance road safety and facilitate seamless integration into intelligent transportation systems. This study introduces a novel approach to predicting the movements of bicyclists and e-scooter riders by leveraging a unique dataset that combines bird's-eye and egocentric views, collected from diverse urban settings across the United States. Employing state-of-the-art deep learning techniques, our model significantly outperforms traditional linear and polynomial regression models in terms of Average Displacement Error (ADE) and Final Displacement Error (FDE), demonstrating superior accuracy in forecasting the trajectories of vulnerable road users. Furthermore, our approach incorporates cognitive annotations to predict crossing intentions, enriching the model's predictive capabilities. The findings highlight the potential of our method to contribute to the development of proactive safety measures and collision avoidance systems, ultimately fostering safer urban mobility landscapes for all road users.

关键词： Deep learning Accuracy Urban areas Predictive models Traffic control Birds Trajectory Forecasting Vehicle dynamics Micromobility

CDDIP: Constrained Diffusion-Driven Deep Image Prior for Seismic Data Reconstruction

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Goyes-Peñafiel, Paul Kamilov, Ulugbek S. Arguello, Henry Department of Systems Engineering Universidad Industrial de Santander Bucaramanga680002 Colombia Department of Computer Science and Engineering Department of Electrical and Systems Engineering Washington University in St. Louis St. LouisMO63130 United States

Seismic data frequently exhibits missing traces, substantially affecting subsequent seismic processing and interpretation. Deep learning-based approaches have demonstrated significant advancements in reconstructing irregularly missing seismic data through supervised and unsupervised methods. Nonetheless, substantial challenges remain, such as generalization capacity and computation time cost during the inference. Our work introduces a reconstruction method that uses a pre-trained generative diffusion model for image synthesis and incorporates Deep Image Prior to enforce data consistency when reconstructing missing traces in seismic data. The proposed method has demonstrated strong robustness and high reconstruction capability of post-stack and pre-stack data with different levels of structural complexity, even in field and synthetic scenarios where test data were outside the training domain. This indicates that our method can handle the high geological variability of different exploration targets. Additionally, compared to other state-of-the-art seismic reconstruction methods using diffusion models. During inference, our approach reduces the number of sampling timesteps by up to 4x. Our implementation is available at https://***/PAULGOYES/*** © 2024, CC BY.

关键词： Seismic response

Data-Driven Human-Robot Interaction Without Velocity Measurement Using Off-Policy Reinforcement Learning

学校读者我要写书评

暂无评论

IEEE/CAA Journal of Automatica Sinica 2022年第1期9卷 47-63页

作者： Yongliang Yang Zihao Ding Rui Wang Hamidreza Modares Donald C.Wunsch Key Laboratory of Knowledge Automation for Industrial Processes Ministry of Education School of Automation&Electrical EngineeringUniversity of Science and Technology BeijingBeijing 100083China School of Automation Beijng Institute TechnologyBeijing 100081China State Key Laboratory of Management and Control for Complex Systems Institute of AutomationChinese Academy of SciencesBeijing 100190China Mechanical Engineering Department Michigan State UniversityEast LansingMI 48824 USA Department of Electrical&Computer Engineering Missouri University of Science&TechnologyRollaMO 65401 USA

In this paper,we present a novel data-driven design method for the human-robot interaction(HRI)system,where a given task is achieved by cooperation between the human and the *** presented HRI controller design is a two-level control design approach consisting of a task-oriented performance optimization design and a plant-oriented impedance controller *** task-oriented design minimizes the human effort and guarantees the perfect task tracking in the outer-loop,while the plant-oriented achieves the desired impedance from the human to the robot manipulator end-effector in the ***-driven reinforcement learning techniques are used for performance optimization in the outer-loop to assign the optimal impedance *** the inner-loop,a velocity-free filter is designed to avoid the requirement of end-effector velocity *** this basis,an adaptive controller is designed to achieve the desired impedance of the robot manipulator in the task *** simulation and experiment of a robot manipulator are conducted to verify the efficacy of the presented HRI design framework.

关键词： Adaptive impedance control data-driven method human-robot interaction(HRI) reinforcement learning velocity-free

Machine Learning Techniques in Additive Manufacturing: A Review

学校读者我要写书评

暂无评论

Machine Learning Techniques in Additive Manufacturing: A Rev...

Artificial Intelligence, Blockchain, and Internet of Things (AIBThings), IEEE International Conference on

作者： Ahmed Alsalem Mohamed Zohdy Department of Industrial and Systems Engineering Oakland University Rochester MI USA Department of Electrical and Computer Engineering Oakland University Rochester MI USA

Additive manufacturing (AM) is a 3D printing process widely used in industries such as automotive, healthcare, aerospace, and consumer goods. It creates different 3D objects by adding different layers over one another. AM manufacturing processes are capable enough to produce complex parts, consisting of intricate designs and geometries, characterized by well-established grain structure and properties without expensive tooling in reduced lead time and at low cost. It is a multi-physical process, including various parameters that impact the final product's quality. Moreover, Machine Learning (ML) techniques are suitable approaches aimed at understanding and predicting the complex processes of AM. In this study, an overview of AM and the challenges in the AM is presented, as well as the ML approaches to address the challenges in AM processes are also discussed. The recently introduced ML techniques in AM processes are evaluated explicitly. To be more precise, limited training data, material choice, and standardization regarding data pose major challenges that need to be addressed when applying ML techniques to AM methods.

关键词：

Performance of NPG in Countable State-Space Average-Cost RL

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Murthy, Yashaswini Grosof, Isaac Maguluri, Siva Theja Srikant, R. Electrical and Computer Engineering University of Illinois Urbana-Champaign UrbanaIL61801 United States Industrial and Systems Engineering Georgia Institute of Technology AtlantaGA30332 United States

We consider policy optimization methods in reinforcement learning settings where the state space is arbitrarily large, or even countably infinite. The motivation arises from control problems in communication networks, matching markets, and other queueing systems. Specifically, we consider the popular Natural Policy Gradient (NPG) algorithm, which has been studied in the past only under the assumption that the cost is bounded and the state space is finite, neither of which holds for the aforementioned control problems. Assuming a Lyapunov drift condition, which is naturally satisfied in some cases and can be satisfied in other cases at a small cost in performance, we design a state-dependent step-size rule which dramatically improves the performance of NPG for our intended applications. In addition to experimentally verifying the performance improvement, we also theoretically show that the iteration complexity of NPG can be made independent of the size of the state space. The key analytical tool we use is the connection between NPG stepsizes and the solution to Poisson’s equation. In particular, we provide policy-independent bounds on the solution to Poisson’s equation, which are then used to guide the choice of NPG stepsizes. © 2024, CC BY.

关键词： State space methods

Deep nonparametric estimation of operators between infinite dimensional spaces

学校读者我要写书评

暂无评论

The Journal of Machine Learning Research

The Journal of Machine Learning Research 2024年第1期25卷 1239-1305页

作者： Hao Liu Haizhao Yang Minshuo Chen Tuo Zhao Wenjing Liao Department of Mathematics Hong Kong Baptist University Hong Kong Department of Mathematics and Department of Computer Science University of Maryland College Park Department of Electrical and Computer Engineering Princeton University School of Industrial and Systems Engineering Georgia Institute of Technology School of Mathematics Georgia Institute of Technology

Learning operators between infinitely dimensional spaces is an important learning task arising in machine learning, imaging science, mathematical modeling and simulations, etc. This paper studies the nonparametric estimation of Lipschitz operators using deep neural networks. Non-asymptotic upper bounds are derived for the generalization error of the empirical risk minimizer over a properly chosen network class. Under the assumption that the target operator exhibits a low dimensional structure, our error bounds decay as the training sample size increases, with an attractive fast rate depending on the intrinsic dimension in our estimation. Our assumptions cover most scenarios in real applications and our results give rise to fast rates by exploiting low dimensional structures of data in operator estimation. We also investigate the influence of network structures (e.g., network width, depth, and sparsity) on the generalization error of the neural network estimator and propose a general suggestion on the choice of network structures to maximize the learning efficiency quantitatively.

关键词： deep neural networks nonparametric estimation operator learning generalization error analysis

Multi-Agent Performative Prediction Beyond the Insensitivity Assumption: A Case Study for Mortgage Competition

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Wang, Guanghui Acharya, Krishna Lakshmikanthan, Lokranjan Muthukumar, Vidya Ziani, Juba College of Computing Georgia Institute of Technology United States School of Industrial and Systems Engineering Georgia Institute of Technology United States School of Electrical and Computer Engineering Georgia Institute of Technology United States

Performative prediction models account for feedback loops in decision-making processes where predictions influence future data distributions. While existing work largely assumes insensitivity of data distributions to small strategy changes, this assumption usually fails in real-world competitive (i.e. multi-agent) settings. For example, in Bertrand-type competitions, a small reduction in one firm’s price can lead that firm to capture the entire demand, while all others sharply lose all of their customers. We study a representative setting of multi-agent performative prediction in which insensitivity assumptions do not hold, and investigate the convergence of natural dynamics. To do so, we focus on a specific game that we call the "Bank Game", where two lenders compete over interest rates and credit score thresholds. Consumers act similarly as to in a Bertrand Competition, with each consumer selecting the firm with the lowest interest rate that they are eligible for based on the firms’ credit thresholds. Our analysis characterizes the equilibria of this game and demonstrates that when both firms use a common and natural no-regret learning dynamic—exponential weights—with proper initialization, the dynamics always converge to stable outcomes despite the general-sum structure. Notably, our setting admits multiple stable equilibria, with convergence dependent on initial conditions. We also provide theoretical convergence results in the stochastic case when the utility matrix is not fully known, but each learner can observe sufficiently many samples of consumers at each time step to estimate it, showing robustness to slight mis-specifications. Finally, we provide experimental results that validate our theoretical findings. © 2025, CC BY.

关键词： Prediction models

Data-Driven Assessment of Operational and Harsh Environmental Stress Factors for Solar Inverters Using Field Measurements

学校读者我要写书评

暂无评论

Data-Driven Assessment of Operational and Harsh Environmenta...

IEEE General Meeting Power& Energy Society

作者： Yi Luo Liming Liu Yuxuan Yuan Zhaoyu Wang Feng Qiu Shijia Zhao Murat Yildirim Electrical and Computer Engineering Iowa State University Ames IA USA Energy Systems Division Argonne National Laboratory Lemont IL USA Industrial and Systems Engineering Wayne State University Detroit MI USA

ISBN: (数字)9798350381832

ISBN: (纸本)9798350381849

The operation of inverters in distributed photovoltaic (PV) systems within distribution systems is critical for ensuring the reliability and durability of the PV system, particularly when challenged by diverse operational and environmental stress conditions. This study delves into analyzing the inverter stress factors to assess the impact of under-sizing and varying environmental stresses on the reliability and service life of PV inverters. A new operational stress metric, equivalent under-sizing rate (EUR), is proposed to quantify the actual inverter stress from the under-sizing problem. Along with the designed weather and extreme event metrics, the proposed EUR is evaluated in a real-world PV inverter dataset. To handle the censored working time variable, the Cox proportional hazards model is adopted to conduct the data-driven stress factor analysis. The regression results reveal aside from the high temperature and large temperature variance, the inverter under-sizing problem is also a significant factor that could influence the lifetime of inverters. Meanwhile, further applications and potential future directions for the proposed work are explored.

关键词： Temperature measurement Measurement Photovoltaic systems Analytical models Temperature Inverters Hazards Reliability Stress Stress measurement

A farewell to the bias-variance tradeoff? An overview of the theory of overparameterized machine learning

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Dar, Yehuda Muthukumar, Vidya Baraniuk, Richard G. Electrical and Computer Engineering Department Rice University Schools of Electrical and Computer Engineering and Industrial and Systems Engineering Georgia Institute of Technology

The rapid recent progress in machine learning (ML) has raised a number of scientific questions that challenge the longstanding dogma of the field. One of the most important riddles is the good empirical generalization of overparameterized models. Overparameterized models are excessively complex with respect to the size of the training dataset, which results in them perfectly fitting (i.e., interpolating) the training data, which is usually noisy. Such interpolation of noisy data is traditionally associated with detrimental overfitting, and yet a wide range of interpolating models – from simple linear models to deep neural networks – have recently been observed to generalize extremely well on fresh test data. Indeed, the recently discovered double descent phenomenon has revealed that highly overparameterized models often improve over the best underparameterized model in test performance. Understanding learning in this overparameterized regime requires new theory and foundational empirical studies, even for the simplest case of the linear model. The underpinnings of this understanding have been laid in very recent analyses of overparameterized linear regression and related statistical learning tasks, which resulted in precise analytic characterizations of double descent. This paper provides a succinct overview of this emerging theory of overparameterized ML (henceforth abbreviated as TOPML) that explains these recent findings through a statistical signal processing perspective. We emphasize the unique aspects that define the TOPML research area as a subfield of modern ML theory and outline interesting open questions that remain. Copyright © 2021, The Authors. All rights reserved.

关键词： Interpolation