检索结果-内蒙古大学图书馆

LiDAR Simulation for Performance Evaluation of UAS Detect and Avoid

学校读者我要写书评

暂无评论

LiDAR Simulation for Performance Evaluation of UAS Detect an...

2021 International Conference on Unmanned Aircraft systems, ICUAS 2021

作者： Riordan, James Manduhu, Manduhu Black, Julie Dow, Alexander Dooly, Gerard Matalonga, Santiago Almada Research Centre School of Computing Engineering and Physical Science University of the West of Scotland Lanarkshire United Kingdom Centre for Robotics and Intelligent Systems University of Limerick Department of Electronic and Computer Engineering

ISBN: (纸本)9780738131153

The solution to mitigating risks associated with beyond Visual Line of Sight (BVLOS) operations of Unmanned Aerial System (UAS) generally focuses on the use of advanced Unmanned Traffic Management (UTM) systems. However, this solution does not take into account other uncooperative objects in the airspace. A more robust approach is to have UTM integrations coupled with onboard machine vision which is tied to automated collision avoidance systems. Future BVLOS regulations in urban situations may require robust embedded software that is capable of detecting air collision hazards in realtime at near and far ranges as uncooperative small aircraft and other unpredictable small objects with fast-changing and unscheduled trajectories pose significant hazards to UAS. This work presents the concept and initial prototyping of a Digital Twin to evaluate the capability of UAS mounted LiDAR to detect small-object air collision risks. A Digital Twin of the Port of Hamburg is augmented with typical port and harbour aerial hazards such as birds, drones, helicopters, and low flying aircraft. The use case scenarios are created in Maya and Unity, with Optix ray tracing of typical LiDAR imaging configurations used to replicate the cause and effect relationship between different LiDAR specifications and their response to small flying objects. Our results demonstrate the inhomogeneous point clouds generated at different spatial-temporal parts of the LiDAR scanning cycle and field of view. These results confirm the challenges of detecting small uncooperative objects by LiDAR. © 2021 IEEE.

关键词： Hazards

Spiking neural networks: Towards bio-inspired multimodal perception in robotics

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Oikonomou, Katerina Maria Balaska, Vasiliki Tsintotas, Konstantinos A. Mavridis, Christos N. Kansizoglou, Ioannis Gasteratos, Antonios The Department of Production and Management Engineering Laboratory of Robotics and Automation Democritus University of Thrace Xanthi Greece The Department of Division of Decision and Control Systems School of Electrical Engineering and Computer Science KTH Royal Institute of Technology Stockholm Sweden

Spiking neural networks (SNNs) have captured apparent interest over the recent years, stemming from neuroscience and reaching the field of artificial intelligence. However, due to their nature SNNs remain far behind in achieving the exceptional performance of deep neural networks (DNNs). As a result, many scholars are exploring ways to enhance SNNs by using learning techniques from DNNs. While this approach has been proven to achieve considerable improvements in SNN performance, we propose another perspective: enhancing the biological plausibility of the models to leverage the advantages of SNNs fully. Our approach aims to propose a brain-like combination of audio-visual signal processing for recognition tasks, intended to succeed in more bio-plausible human-robot interaction applications. © 2024, CC BY-NC-SA.

关键词： Adversarial machine learning

FeCAM: exploiting the heterogeneity of class distributions in exemplar-free continual learning 23

学校读者我要写书评

暂无评论

FeCAM: exploiting the heterogeneity of class distributions i...

Proceedings of the 37th International Conference on Neural Information Processing systems

作者： Dipam Goswami Yuyang Liu Bartłomiej Twardowski Joost van de Weijer Department of Computer Science Universitat Autònoma de Barcelona and Computer Vision Center Barcelona University of Chinese Academy of Sciences and State Key Laboratory of Robotics Shenyang Institute of Automation Chinese Academy of Sciences and Institutes for Robotics and Intelligent Manufacturing Chinese Academy of Sciences Department of Computer Science Universitat Autònoma de Barcelona and Computer Vision Center Barcelona and IDEAS-NCBR

Exemplar-free class-incremental learning (CIL) poses several challenges since it prohibits the rehearsal of data from previous tasks and thus suffers from catastrophic forgetting. Recent approaches to incrementally learning the classifier by freezing the feature extractor after the first task have gained much attention. In this paper, we explore prototypical networks for CIL, which generate new class prototypes using the frozen feature extractor and classify the features based on the Euclidean distance to the prototypes. In an analysis of the feature distributions of classes, we show that classification based on Euclidean metrics is successful for jointly trained features. However, when learning from non-stationary data, we observe that the Euclidean metric is suboptimal and that feature distributions are heterogeneous. To address this challenge, we revisit the anisotropic Mahalanobis distance for CIL. In addition, we empirically show that modeling the feature covariance relations is better than previous attempts at sampling features from normal distributions and training a linear classifier. Unlike existing methods, our approach generalizes to both many- and few-shot CIL settings, as well as to domain-incremental settings. Interestingly, without updating the backbone network, our method obtains state-of-the-art results on several standard continual learning benchmarks. Code is available at https://***/dipamgoswami/FeCAM.

关键词：

Learning Layer-wise Equivariances Automatically using Gradients

学校读者我要写书评

暂无评论

arXiv 2023年

作者： van der Ouderaa, Tycho F.A. Immer, Alexander van der Wilk, Mark Department of Computing Imperial College London United Kingdom Department of Computer Science ETH Zurich Switzerland Max Planck Institute for Intelligent Systems Tübingen Germany

Convolutions encode equivariance symmetries into neural networks leading to better generalisation performance. However, symmetries provide fixed hard constraints on the functions a network can represent, need to be specified in advance, and can not be adapted. Our goal is to allow flexible symmetry constraints that can automatically be learned from data using gradients. Learning symmetry and associated weight connectivity structures from scratch is difficult for two reasons. First, it requires efficient and flexible parameterisations of layer-wise equivariances. Secondly, symmetries act as constraints and are therefore not encouraged by training losses measuring data fit. To overcome these challenges, we improve parameterisations of soft equivariance and learn the amount of equivariance in layers by optimising the marginal likelihood, estimated using differentiable Laplace approximations. The objective balances data fit and model complexity enabling layer-wise symmetry discovery in deep networks. We demonstrate the ability to automatically learn layer-wise equivariances on image classification tasks, achieving equivalent or improved performance over baselines with hard-coded symmetry. Copyright © 2023, The Authors. All rights reserved.

关键词： Machine learning

POMDP-Guided Active Force-Based Search for Robotic Insertion

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Wang, Chen Luo, Haoxiang Zhang, Kun Chen, Hua Pan, Jia Zhang, Wei School of System Design and Intelligent Manufacturing Southern University of Science and Technology Shenzhen China Department of Computer Science The University of Hong Kong Hong Kong Cheng Kar-Shun Robotics Institute Hong Kong University of Science and Technology Clear Water Bay Hong Kong

In robotic insertion tasks where the uncertainty exceeds the allowable tolerance, a good search strategy is essential for successful insertion and significantly influences efficiency. The commonly used blind search method is time-consuming and does not exploit the rich contact information. In this paper, we propose a novel search strategy that actively utilizes the information contained in the contact configuration and shows high efficiency. In particular, we formulate this problem as a Partially Observable Markov Decision Process (POMDP) with carefully designed primitives based on an in-depth analysis of the contact configuration's static stability. From the formulated POMDP, we can derive a novel search strategy. Thanks to its simplicity, this search strategy can be incorporated into a Finite-State-Machine (FSM) controller. The behaviors of the FSM controller are realized through a low-level Cartesian Impedance Controller. Our method is based purely on the robot's proprioceptive sensing and does not need visual or tactile sensors. To evaluate the effectiveness of our proposed strategy and control framework, we conduct extensive comparison experiments in simulation, where we compare our method with the baseline approach. The results demonstrate that our proposed method achieves a higher success rate with a shorter search time and search trajectory length compared to the baseline method. Additionally, we show that our method is robust to various initial displacement errors. Copyright © 2024, The Authors. All rights reserved.

关键词： robotics

A joint image compression and encryption scheme based on a novel coupled map lattice system and DNA operations

学校读者我要写书评

暂无评论

Frontiers of Information Technology & Electronic Engineering 2023年第6期24卷 813-827页

作者： Yuanyuan LI Xiaoqing YOU Jianquan LU Jungang LOU Department of Applied Mathematics College of ScienceNanjing Forestry UniversityNanjing 210037China School of Cyber Science and Engineering Southeast UniversityNanjing 210096China Department of Systems Science School of MathematicsSoutheast UniversityNanjing 210096China Yangtze Delta Region(Huzhou)Institute of Intelligent Transportation Huzhou UniversityHuzhou 313000China School of Computer Science and Technology Zhejiang Normal UniversityJinhua 321004China

In this paper,an efficient image encryption scheme based on a novel mixed linear–nonlinear coupled map lattice(NMLNCML)system and DNA operations is *** proposed NMLNCML system strengthens the chaotic characteristics of the system,and is applicable for image *** main advantages of the proposed method are embodied in its extensive key space;high sensitivity to secret keys;great resistance to chosen-plaintext attack,statistical attack,and differential attack;and good robustness to noise and data *** image cryptosystem adopts the architecture of scrambling,compression,and ***,a plain image is transformed to a sparsity coefficient matrix by discrete wavelet transform,and plaintext-related Arnold scrambling is performed on the coefficient ***,semi-tensor product(STP)compressive sensing is employed to compress and encrypt the coefficient ***,the compressed coefficient matrix is diffused by DNA random encoding,DNA addition,and bit XOR *** NMLNCML system is applied to generate chaotic elements in the STP measurement matrix of compressive sensing and the pseudo-random sequence in DNA *** SHA-384 function is used to produce plaintext secret keys and thus makes the proposed encryption algorithm highly sensitive to the original *** results and performance analyses verify the security and effectiveness of our scheme.

关键词： Compressive sensing Coupled map lattice(CML) DNA operations Semi-tensor product

Producing Lunar Steel and Oxygen using Molten Lunar Regolith Electrolysis

学校读者我要写书评

暂无评论

Producing Lunar Steel and Oxygen using Molten Lunar Regolith...

Accelerating Space Commerce, Exploration, and New Discovery Conference, ASCEND 2023

作者： Patel, Palak B. Soto, Jose Adams, Zachary K. Koller, Scarlett Latyshev, Kir Boothe, Evan Cooper, Amber Fontanez, Jonatan Jalil, Laman Ma, Beverly Meza, Cesar Misquitta, Kristoff Na, Hyein Nguyen, Tom Pendley, Hezekiah Sangabattula, Lokesh Usey, Helena Zehner, Alice Allanore, Antoine Culpepper, Martin De Weck, Olivier Lordos, George Hoffman, Jeffrey Bergman, Dean Department of Mechanical Engineering Massachusetts Institute of Technology CambridgeMA02139 United States Department of Aeronautics and Astronautics Massachusetts Institute of Technology CambridgeMA02139 United States Department of Electrical Engineering and Computer Science Massachusetts Institute of Technology CambridgeMA02139 United States Department of Aeronautics and Astronautics Massachusetts Institute of Technology Sloan School of Management CambridgeMA02139 United States Department of Chemical Engineering Massachusetts Institute of Technology CambridgeMA02139 United States Department of Material Science and Engineering Massachusetts Institute of Technology CambridgeMA02139 United States Honeybee Robotics Exploration Systems AltadenaCA91001 United States

ISBN: (纸本)9781624107054

With the current Artemis missions planned, by 2028, NASA and its international and commercial partners may be establishing a permanent presence at the lunar South pole. The Artemis Base Camp will be the locus of international and public-private collaboration under the Artemis Accords, helping to bootstrap a lunar economy. As the lunar economy grows, traffic of people and goods will grow with it. Thus, we envision that NASA and its international and commercial partners will have increasing long-term needs for producing metal from in situ resources such as lunar regolith. These metals can be used to produce many end products on the moon including large pressure vessels, both for spacious habitats and industrial-scale storage and operations. In this paper, as a part of the NASA 2023 BIG Idea Challenge ‘Lunar Forge: Producing Metal Products on the Moon’, we design and manufacture a molten regolith electrolysis (MRE) reactor capable of producing various alloys of steel and other metal products from lunar regolith, with oxygen as a useful byproduct. We have built ARTEMIS Steelworks (Advancing Reactor Technologies for Electrolytic Manufacturing of In situ Steel), which consists of an electrolytic reactor, supporting the addition of alloying elements to produce molten steel from lunar regolith simulant. To prepare for a lunar technology demonstration, a preliminary test of the reaction on Earth was performed in a vacuum, a sonicator was tested to dislodge oxygen bubbles from the anode to increase electrode life, and an automated iron and slag collector was made to reduce the need for astronaut operation. The project’s technical objectives are to demonstrate steel-making capability while also quantifying energy efficiency, steel quality, and the expected useful life of the electrodes and vessel under different configurations relative to the state of the art. The comprehensive testing plan presented here includes functional subsystem and integrated testing in ambient conditio

关键词： Electrolysis

HOOD: Hierarchical Graphs for Generalized Modelling of Clothing Dynamics

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Grigorev, Artur Thomaszewski, Bernhard Black, Michael J. Hilliges, Otmar ETH Zurich Department of Computer Science Switzerland Max Planck Institute for Intelligent Systems Tubingen Germany

We propose a method that leverages graph neural networks, multi-level message passing, and unsupervised training to enable efficient prediction of realistic clothing dynamics. Whereas existing methods based on linear blend skinning must be trained for specific garments, our method, called HOOD, is agnostic to body shape and applies to tight-fitting garments as well as loose, free-flowing clothing. Furthermore, HOOD handles changes in topology (e.g., garments with buttons or zippers) and material properties at inference time. As one key contribution, we propose a hierarchical message-passing scheme that efficiently propagates stiff stretching modes while preserving local detail. We empirically show that HOOD outperforms strong baselines quantitatively and that its results are perceived as more realistic than state-of-the-art methods. Copyright © 2022, The Authors. All rights reserved.

关键词： Message passing

Terrain Classification Using Inside-Wheel Cameras Based on Wheel-Terrain Interaction Characteristics

学校读者我要写书评

暂无评论

Terrain Classification Using Inside-Wheel Cameras Based on W...

IEEE International Conference on robotics and Biomimetics

作者： Longteng Hu Feng Xue Chen Yao Yunzhou Li Jin Wei Peichen Wang Zheng Zhu Zhenzhong Jia Department of Mechanical and Energy Engineering Shenzhen Key Laboratory of Biomimetic Robotics and Intelligent Systems Southern University of Science and Technology (SUSTech) Shenzhen China Harbin of Institute of Technology (HIT) Harbin China

The wheeled rover is widely used in off-road planetary exploration tasks nowadays. However, complex operating environment like soft deformable terrains on the planet surface demands the high performance of rover. In order to reduce danger, environmental perception technologies need to be used on the rover. Terrain classification is a critical method in environmental perception. In this research, a wheel-vision-based terrain classification method is proposed. This method uses wheel-terrain interaction parameters obtained from inside-wheel cameras which are strongly related to the terrain as input. We use single-wheel testbed to collect and process data to form a dataset. The data is obtained under different slip ratios and different terrains. Then, we train four terrain classification models using machine learning algorithms and analyze their performance. From the results, all performance criteria of the models achieve over 95%, with DT being the optimal method. The feasibility of our method in the case of fewer training samples and the rationality of dataset selection are also verified through the results of different training set sizes and the ablation experiment.

关键词：