检索结果-内蒙古大学图书馆

D3GU: Multi-target Active Domain Adaptation via Enhancing Domain Alignment

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Zhang, Lin Xu, Linghan Motamed, Saman Chakraborty, Shayok De la Torre, Fernando Robotics Institute Carnegie Mellon University United States Department of Computer Science Florida State University United States

Unsupervised domain adaptation (UDA) for image classification has made remarkable progress in transferring classification knowledge from a labeled source domain to an unlabeled target domain, thanks to effective domain alignment techniques. Recently, in order to further improve performance on a target domain, many Single-Target Active Domain Adaptation (ST-ADA) methods have been proposed to identify and annotate the salient and exemplar target samples. However, it requires one model to be trained and deployed for each target domain and the domain label associated with each test sample. This largely restricts its application in the ubiquitous scenarios with multiple target domains. Therefore, we propose a Multi-Target Active Domain Adaptation (MT-ADA) framework for image classification, named D3GU, to simultaneously align different domains and actively select samples from them for annotation. This is the first research effort in this field to our best knowledge. D3GU applies Decomposed Domain Discrimination (D3) during training to achieve both source-target and target-target domain alignments. Then during active sampling, a Gradient Utility (GU) score is designed to weight every unlabeled target image by its contribution towards classification and domain alignment tasks, and is further combined with KMeans clustering to form GU-KMeans for diverse image sampling. Extensive experiments on three benchmark datasets, Office31, OfficeHome, and DomainNet, have been conducted to validate consistently superior performance of D3GU for MT-ADA 1 © 2024, CC BY.

关键词： Image classification

Towards Contact-Aided Motion Planning for Tendon-Driven Continuum Robots

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Rao, Priyanka Salzman, Oren Burgner-Kahrs, Jessica Continuum Robotics Laboratory Department of Mathematical & Computational Sciences University of Toronto ON Canada Department of Computer Science Technion Haifa3200003 Israel

Tendon-driven continuum robots (TDCRs), with their flexible backbones, offer the advantage of being used for navigating complex, cluttered environments. However, to do so, they typically require multiple segments, often leading to complex actuation and control challenges. To this end, we propose a novel approach to navigate cluttered spaces effectively for a single-segment long TDCR which is the simplest topology from a mechanical point of view. Our key insight is that by leveraging contact with the environment we can achieve multiple curvatures without mechanical alterations to the robot. Specifically, we propose a search-based motion planner for a single-segment TDCR. This planner, guided by a specially designed heuristic, discretizes the configuration space and employs a best-first search. The heuristic, crucial for efficient navigation, provides an effective cost-to-go estimation while respecting the kinematic constraints of the TDCR and environmental interactions. We empirically demonstrate the efficiency of our planner—testing over 525 queries in environments with both convex and non-convex obstacles, our planner is demonstrated to have a success rate of about 80% while baselines were not able to obtain a success rate higher than 30%. The difference is attributed to our novel heuristic which is shown to significantly reduce the required search space. © 2024, CC BY-SA.

关键词： Motion planning

Crop and Weeds Classification in Aerial Imagery of Sesame Crop Fields Using a Patch-Based Deep Learning Model-Ensembling Method 2

学校读者我要写书评

暂无评论

Crop and Weeds Classification in Aerial Imagery of Sesame Cr...

2nd International Conference on Digital Futures and Transformative Technologies, ICoDT2 2022

作者： Moazzam, S. Imran Khan, Umar S. Nawaz, Tahir Qureshi, Waqar S. National University of Sciences and Technology National Centre of Robotics and Automation Department of Mechatronics Engineering Islamabad Pakistan Technological University Dublin Department of Computer Science Dublin Ireland

ISBN: (纸本)9781665498197

Sesame (Sesamum indicum L.) is an important commercial and food crop, and its yields is limited by many insects, pests, diseases, and weeds. Autonomous aerial agrochemicals spray application on sesame fields using a drone aims to save crop from these yield limiting factors and in addition agrochemicals application quantity and site could be controlled, and human health is expected to be protected. For accurate and selective spray application, autonomous systems would need some parameters to distinguish between crop, weed and background. In this research an aerial sesame field dataset has been collected with the focus to classify patch areas of sesame and weeds present in the field. Dataset was captured using Agrocam. We have developed a patch image-based classification approach along with a novel SesameWeedNet convolutional neural network (CNN) inspired by the layer's configuration of VGG networks and depth-wise convolutions of the MobileNet. The small model contains 6 convolutional layers, and it runs faster and accurately on small patch images. Our approach breaks 1920×1080-pixel images into smaller patch images of size 45×45 pixels. After that, these small patch images are fed to a relatively small CNN for training, validation, and finally for classification. Patch based model ensemble and dataset grouping are two major parts in our methodology. Our system recommends the dataset grouping according to vegetation present in the images to enhance classification results. We have achieved accuracy up to 96.7% with our proposed method. We have tested our system under sunlight variation, in wet and dry soil conditions and at different growth stages. To the best of our knowledge, no attempt has been made to classify and treat crop and weeds in sesame fields at the post-emergence stage previously. In this research we have made the contribution of aerial sesame-weed dataset and a complete deep learning-based approach to classify weeds in sesame fields under variable lighti

关键词： Crops

CoCap: Coordinated motion Capture for multi-actor scenes in outdoor environments

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Rauniyar, Aditya Corah, Micah Scherer, Sebastian Robotics Institute School of Computer Science Carnegie Mellon University PittsburghPA United States Department of Computer Science The Colorado School of Mines GoldenCO United States

Motion capture has become increasingly important, not only in computer animation but also in emerging fields like the virtual reality, bioinformatics, and humanoid training. Capturing outdoor environments offers extended horizon scenes but introduces challenges with occlusions and obstacles. Recent approaches using multi-drone systems to capture multiple actor scenes often fail to account for multi-view consistency and reasoning across cameras in cluttered environments. Coordinated motion Capture (CoCap), inspired by Conflict-Based Search (CBS), addresses this issue by coordinating view planning to ensure multi-view reasoning during conflicts. In scenarios with high occlusions and obstacles, where the likelihood of inter-robot collisions increases, CoCap demonstrates performance that approaches the ideal outcomes of unconstrained planning, outperforming existing sequential planning methods. Additionally, CoCap offers a single-robot view search approach for real-time applications in dense environments. Copyright © 2024, The Authors. All rights reserved.

关键词： Robot programming

PFAS Incineration: Known Unknowns, Reaction Simulation, and ML to Predict Product Properties

学校读者我要写书评

暂无评论

PFAS Incineration: Known Unknowns, Reaction Simulation, and ...

2023 IEEE MIT Undergraduate Research Technology Conference, URTC 2023

作者： Soiffer, Duncan Pyorre, Tuomas Kenney, David Teixeira, Andrew Worcester Polytechnic Institute Department of Computer Science and Mathematical Sciences WorcesterMA United States Worcester Polytechnic Institute Department of Robotics Engineering WorcesterMA United States Worcester Polytechnic Institute Department of Chemical Engineering WorcesterMA United States

ISBN: (纸本)9798350308600

Per- and polyfluoroalkyl substances (PFAS) are a group of chemicals used for their impressive stability, a property which has also led them to be found polluting virtually every sector of the environment-including our own bodies. Presently, our ability to destroy PFAS is limited largely to incineration, whose products and side effects are poorly understood. Research on PFAS incineration is constrained by high experimental costs and difficulty in accurately measuring reaction products. This presents an opportunity for computer simulation and machine learning to provide valuable assistance. In this study, three kinds of thermal PFAS decomposition are simulated. Further, to broaden physical understanding of PFAS incineration products, a recent in-house machine learning model for predicting the octanol-water partition coefficient (LogP) of chemicals is improved and applied to simulation results. A decrease in LogP is observed in all treatments, with the most dramatic effect (-2 average LogP) observed in oxidative pyrolysis. © 2023 IEEE.

关键词： Machine learning

AN EXPERIMENTAL DESIGN PERSPECTIVE ON MODEL-BASED REINFORCEMENT LEARNING 10

学校读者我要写书评

暂无评论

AN EXPERIMENTAL DESIGN PERSPECTIVE ON MODEL-BASED REINFORCEM...

10th International Conference on Learning Representations, ICLR 2022

作者： Mehta, Viraj Paria, Biswajit Schneider, Jeff Ermon, Stefano Neiswanger, Willie Robotics Insitute Machine Learning Department Carnegie Mellon University PittsburghPA United States Computer Science Department Stanford University StanfordCA United States

In many practical applications of RL, it is expensive to observe state transitions from the environment. For example, in the problem of plasma control for nuclear fusion, computing the next state for a given state-action pair requires querying an expensive transition function which can lead to many hours of computer simulation or dollars of scientific research. Such expensive data collection prohibits application of standard RL algorithms which usually require a large number of observations to learn. In this work, we address the problem of efficiently learning a policy while making a minimal number of state-action queries to the transition function. In particular, we leverage ideas from Bayesian optimal experimental design to guide the selection of state-action queries for efficient learning. We propose an acquisition function that quantifies how much information a state-action pair would provide about the optimal solution to a Markov decision process. At each iteration, our algorithm maximizes this acquisition function, to choose the most informative state-action pair to be queried, thus yielding a data-efficient RL approach. We experiment with a variety of simulated continuous control problems and show that our approach learns an optimal policy with up to 5-1, 000× less data than model-based RL baselines and 103 - 105× less data than model-free RL baselines. We also provide several ablated comparisons which point to substantial improvements arising from the principled method of obtaining data. © 2022 ICLR 2022 - 10th International Conference on Learning Representationss. All rights reserved.

关键词： Markov processes

Characterizing the Complexity of Social Robot Navigation Scenarios

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Stratton, Andrew Hauser, Kris Mavrogiannis, Christoforos Department of Robotics University of Michigan Ann Arbor United States Department of Computer Science University of Illinois at Urbana-Champaign United States

Social robot navigation algorithms are often demonstrated in overly simplified scenarios, prohibiting the extraction of practical insights about their relevance to real-world domains. Our key insight is that an understanding of the inherent complexity of a social robot navigation scenario could help characterize the limitations of existing navigation algorithms and provide actionable directions for improvement. Through an exploration of recent literature, we identify a series of factors contributing to the complexity of a scenario, disambiguating between contextual and robot-related ones. We then conduct a simulation study investigating how manipulations of contextual factors impact the performance of a variety of navigation algorithms. We find that dense and narrow environments correlate most strongly with performance drops, while the heterogeneity of agent policies and directionality of interactions have a less pronounced effect. Our findings motivate a shift towards developing and testing algorithms under higher-complexity settings. Copyright © 2024, The Authors. All rights reserved.

关键词： Social robots

Versatile Navigation under Partial Observability via Value-guided Diffusion Policy

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Zhang, Gengyu Tang, Hao Yan, Yan Department of Computer Science Illinois Institute of Technology United States Robotics Institute Carnegie Mellon University United States

Route planning for navigation under partial observability plays a crucial role in modern robotics and autonomous driving. Existing route planning approaches can be categorized into two main classes: traditional autoregressive and diffusion-based methods. The former often fails due to its myopic nature, while the latter either assumes full observability or struggles to adapt to unfamiliar scenarios, due to strong couplings with behavior cloning from experts. To address these deficiencies, we propose a versatile diffusion-based approach for both 2D and 3D route planning under partial observability. Specifically, our value-guided diffusion policy first generates plans to predict actions across various timesteps, providing ample foresight to the planning. It then employs a differentiable planner with state estimations to derive a value function, directing the agent's exploration and goal-seeking behaviors without seeking experts while explicitly addressing partial observability. During inference, our policy is further enhanced by a best-plan-selection strategy, substantially boosting the planning success rate. Moreover, we propose projecting point clouds, derived from RGB-D inputs, onto 2D grid-based bird-eye-view maps via semantic segmentation, generalizing to 3D environments. This simple yet effective adaption enables zero-shot transfer from 2D-trained policy to 3D, cutting across the laborious training for 3D policy, and thus certifying our versatility. Experimental results demonstrate our superior performance, particularly in navigating situations beyond expert demonstrations, surpassing state-of-the-art autoregressive and diffusion-based baselines for both 2D and 3D scenarios. © 2024, CC BY-NC-SA.

关键词： Observability

An Energy-Efficient Mobile Agent-Based Data Aggregation Scheme for Wireless Body Area Networks

学校读者我要写书评

暂无评论

computers, Materials & Continua 2022年第3期70卷 5929-5948页

作者： Gulzar Mehmood Muhammad Zahid Khan Muhammad Fayaz Mohammad Faisal Haseeb Ur Rahman Jeonghwan Gwak Department of Computer Science&Information Technology University of MalakandChakdara Dir Lower23021Pakistan Department of Computer Science University of Central AsiaNarynKyrgyzstan Department of Software Korea National University of TransportationChungju27469Korea Department of Biomedical Engineering Korea National University of TransportationChungju27469Korea Department of AI Robotics Engineering Korea National University of TransportationChungju27469Korea Department of IT and Energy Convergence(BK21 FOUR) Korea National University of TransportationChungju27469Korea

Due to the advancement in wireless technology and miniaturization,Wireless Body Area Networks(WBANs)have gained enormous popularity,having various applications,especially in the healthcare *** are intrinsically resource-constrained;therefore,they have specific design and development *** such highly desirable requirement is an energy-efficient and reliable Data Aggregation(DA)mechanism for *** efficient and reliableDAmay ultimately push the network to operate without much human intervention and further extend the network *** conventional client-serverDAparadigm becomes unsuitable and inefficient for WBANs when a large amount of data is generated in the ***,in most of the healthcare applications(patient’s critical conditions),it is highly important and required to send data as soon as possible;therefore,reliable data aggregation in WBANs is of great *** tackle the shortcomings of the client-serverDAparadigm,theMobile Agent-Basedmechanismproved to be a more workable *** aMobile Agent-Based mechanism,a taskspecific mobile agent(code)traverses to the intended sources to gather *** agents travel on a predefined path called itinerary;however,planning a suitable and reliable itinerary for a mobile agent is also a challenging issue *** paper presents a new Mobile Agent-Based DA scheme for WBANs,which is energy-efficient and ***,in the proposed scheme,the network is divided into clusters,and cluster-heads are ***,a mobile agent is generated from the base station to collect the required data from cluster *** the case,if any fault occurs in the existing itinerary,an alternate itinerary is planned in real-time without compromising the network *** our simulation-based validation,we have found that the proposed system delivers significantly improved fault-tolerance and reliability with energy-efficiency and extended network lifetime in WBANs.

关键词： WBANs mobile-agent data aggregation reliability