检索结果-内蒙古大学图书馆

Learning high-speed flight in the wild

学校读者我要写书评

暂无评论

arXiv 2021年

作者： Loquercio, Antonio Kaufmann, Elia Ranftl, René Müller, Matthias Koltun, Vladlen Scaramuzza, Davide Robotics and Perception Group UZH Zurich Switzerland Intelligent Systems Lab Intel Munich Germany Intelligent Systems Lab Intel Santa ClaraCA United States

Quadrotors are agile. Unlike most other machines, they can traverse extremely complex environments at high speeds. To date, only expert human pilots have been able to fully exploit their capabilities. Autonomous operation with onboard sensing and computation has been limited to low speeds. State-of-the-art methods generally separate the navigation problem into subtasks: sensing, mapping, and planning. Although this approach has proven successful at low speeds, the separation it builds upon can be problematic for high-speed navigation in cluttered environments. The subtasks are executed sequentially, leading to increased processing latency and a compounding of errors through the pipeline. Here we propose an end-to-end approach that can autonomously fly quadrotors through complex natural and human-made environments at high speeds, with purely onboard sensing and computation. The key principle is to directly map noisy sensory observations to collision-free trajectories in a receding-horizon fashion. This direct mapping drastically reduces processing latency and increases robustness to noisy and incomplete perception. The sensorimotor mapping is performed by a convolutional network that is trained exclusively in simulation via privileged learning: imitating an expert with access to privileged information. By simulating realistic sensor noise, our approach achieves zero-shot transfer from simulation to challenging real-world environments that were never experienced during training: dense forests, snow-covered terrain, derailed trains, and collapsed buildings. Our work demonstrates that end-to-end policies trained in simulation enable high-speed autonomous flight through challenging environments, outperforming traditional obstacle avoidance pipelines. Copyright © 2021, The Authors. All rights reserved.

关键词： Pipelines

On the Use of Torque Measurement in Centroidal State Estimation

学校读者我要写书评

暂无评论

On the Use of Torque Measurement in Centroidal State Estimat...

IEEE International Conference on robotics and Automation (ICRA)

作者： Shahram Khorshidi Ahmad Gazar Nicholas Rotella Maximilien Naveau Ludovic Righetti Maren Bennewitz Majid Khadiv Max-Planck Institute for Intelligent Systems Tübingen Germany Humanoid Robots Lab University of Bonn Germany Agility Robotics Oregon USA LAAS-CNRS Toulouse France Tandon School of Engineering New York University USA

State-of-the-art legged robots are either capable of measuring torque at the output of their drive systems, or have transparent drive systems which enable the computation of joint torques from motor currents. In either case, this sensor modality is seldom used in state estimation. In this paper, we propose to use joint torque measurements to estimate the centroidal states of legged robots. To do so, we project the whole-body dynamics of a legged robot into the nullspace of the contact constraints, allowing expression of the dynamics independent of the contact forces. Using the constrained dynamics and the centroidal momentum matrix, we are able to directly relate joint torques and centroidal states dynamics. Using the resulting model as the process model of an Extended Kalman Filter (EKF), we fuse the torque measurement in the centroidal state estimation problem. Through real-world experiments on a quadruped robot executing different gaits, we demonstrate that the estimated centroidal states from our torque-based EKF drastically improve the recovery of these quantities compared to direct computation.

关键词：

Design and Validation of Tunable Stiffness Actuator using Soft-Rigid Combined Layer Jamming Mechanism

学校读者我要写书评

暂无评论

Design and Validation of Tunable Stiffness Actuator using So...

IEEE International Conference on Soft robotics (RoboSoft)

作者： Seoyeon Ham Brian Byunghyun Kang Kevin Abishek Hyunglae Lee Wansoo Kim Department of Interdisciplinary Robot Engineering Systems Hanyang University Gyeonggi-do Republic of Korea Neuromuscular Control and Human Robotics Lab. Arizona States University AZ USA Department of Intelligent Mechatronics Engineering Sejong University Seoul Republic of Korea Robotics Department Ansan-si Gyeonggi-do Republic of Korea

This paper presents a tunable stiffness actuator with a soft-rigid combined layer jamming mechanism. The tunable stiffness actuator is aimed to be integrated into an exosuit to prevent ankle sprain and avoid or mitigate the development of chronic ankle instability. The main purpose of the soft-rigid layer jamming mechanism is to produce large stiffness with a small volume and achieve a linear stiffness characteristic. To this end, the actuator is designed to include rigid retainer pieces within the soft silicone layers, and each soft-rigid layer is jammed to induce stiffness changes. To validate the stiffness characteristics of the proposed soft-rigid actuator, a series of experiments were performed and stiffness changes were investigated for varying jamming states from unjammed to fully jammed states. Increasing the number of jamming layer effectively increased the actuator stiffness, which was consistent with expected results from the analytical model. Soft-rigid actuator's stiffness at the fully jammed state was 212.1% and 123.1% higher than the unjammed state for one-side and both sides anchored conditions, respectively. Compared to the soft actuator without the rigid retainer, the soft-rigid actuator exhibited a more linear characteristic (Pearson correlation coefficient = 0.990 and 0.997 for one-side and both sides anchored conditions, respectively). Moreover, the soft-rigid actuator achieved significantly higher stiffness than the soft actuator in all jamming states (at least 41.3% increase in each jamming state). The results suggest a potential use of the tunable stiffness actuator to develop a soft ankle exosuit with highly variable but linear stiffness characteristics.

关键词： Correlation coefficient Actuators Analytical models Composite materials Mechanical cables Soft robotics Jamming

Classification of Activities of Daily Living in Subjects with Parkinson’s Disease using Artificial Neural Networks

学校读者我要写书评

暂无评论

Classification of Activities of Daily Living in Subjects wit...

Pan American Health Care Exchanges, PAHCE

作者： L. R. Montero J. A. Bastian A. I. P. SanPablo Universidad La Salle Mexico City Mexico Groups of Applied Intelligent Systems and Collaborative Assistance Robotics Human Motion Analysis and Rehabilitation Engineering Lab National Institute of Rehabilitation Luis Guillermo Ibarra Ibarra Mexico City Mexico

Parkinson’s disease (PD) is a progressive disease characterized by slow movements, tremors, and postural instability [1]. Inertial sensors provide valuable information to assess the quality of movement in patients with PD [2]. The present work proposes improving the follow-up and support systems issued by specialists in patients with PD by identifying and adequately classifying the activities of daily living (ADL) carried out by these patients by applying programmed artificial neuronal networks (ANN) in Python language. The proposed supervised learning method allowed the classification of 6 ADLs through 64 signals from inertial sensors that made up the database (dataset) to train a multilayer perceptron (MLP), achieving an accuracy of 99%. Also, by applying dimensionality reduction to dataset, it was possible to reach 80% of accuracy in the classification. These results can potentially help the specialist to reduce costs and achieve a more accurate evaluation, seeking to improve monitoring and quality in the rehabilitation of these patients.

关键词：

Deep Domain Adaptation Regression for Force Calibration of Optical Tactile Sensors

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Chen, Zhuo Ou, Ni Jiang, Jiaqi Luo, Shan Robot Perception Lab Centre for Robotics Research Department of Engineering King’s College London LondonWC2R 2LS United Kingdom State Key Laboratory of Intelligent Control and Decision of Complex Systems Beijing Institute of Technology Beijing100081 China

Optical tactile sensors provide robots with rich force information for robot grasping in unstructured environments. The fast and accurate calibration of three-dimensional contact forces holds significance for new sensors and existing tactile sensors which may have incurred damage or aging. However, the conventional neural-network-based force calibration method necessitates a large volume of force-labeled tactile images to minimize force prediction errors, with the need for accurate Force/Torque measurement tools as well as a time-consuming data collection process. To address this challenge, we propose a novel deep domain-adaptation force calibration method, designed to transfer the force prediction ability from a calibrated optical tactile sensor to uncalibrated ones with various combinations of domain gaps, including marker presence, illumination condition, and elastomer modulus. Experimental results show the effectiveness of the proposed unsupervised force calibration method, with lowest force prediction errors of 0.102N (3.4% in full force range) for normal force, and 0.095N (6.3%) and 0.062N (4.1%) for shear forces along the x-axis and y-axis, respectively. This study presents a promising, general force calibration methodology for optical tactile sensors. Copyright © 2024, The Authors. All rights reserved.

关键词： Forecasting

TransForce: Transferable Force Prediction for Vision-based Tactile Sensors with Sequential Image Translation

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Chen, Zhuo Ou, Ni Zhang, Xuyang Luo, Shan The Robot Perception Lab Centre for Robotics Research Department of Engineering King's College London LondonWC2R 2LS United Kingdom The State Key Laboratory of Intelligent Control and Decision of Complex Systems Beijing Institute of Technology Beijing100081 China

Vision-based tactile sensors (VBTSs) provide high-resolution tactile images crucial for robot in-hand manipulation. However, force sensing in VBTSs is underutilized due to the costly and time-intensive process of acquiring paired tactile images and force labels. In this study, we introduce a transferable force prediction model, TransForce, designed to leverage collected image-force paired data for new sensors under varying illumination colors and marker patterns while improving the accuracy of predicted forces, especially in the shear direction. Our model effectively achieves translation of tactile images from the source domain to the target domain, ensuring that the generated tactile images reflect the illumination colors and marker patterns of the new sensors while accurately aligning the elastomer deformation observed in existing sensors, which is beneficial to force prediction of new sensors. As such, a recurrent force prediction model trained with generated sequential tactile images and existing force labels is employed to estimate higher-accuracy forces for new sensors with lowest average errors of 0.69N (5.8% in full work range) in x-axis, 0.70N (5.8%) in y-axis, and 1.11N (6.9%) in z-axis compared with models trained with single images. The experimental results also reveal that pure marker modality is more helpful than the RGB modality in improving the accuracy of force in the shear direction, while the RGB modality show better performance in the normal direction. Copyright © 2024, The Authors. All rights reserved.

关键词： RGB color model

Contextual Multi-Armed Bandit-Based Dynamic Cooperative Link Configuration for AUV in UWASNs With Energy Harvesting

学校读者我要写书评

暂无评论

IEEE Transactions on Cognitive Communications and Networking 2025年

作者： Dai, Jun Li, Xinbin Han, Song Yu, Junzhi Liu, Zhixin Yanshan University Key Lab of Industrial Computer Control Engineering of Hebei Province Qinhuangdao066004 China Yanshan University Key Laboratory of Intelligent Rehabilitation and Neuromodulation of Hebei Province Qinhuangdao066004 China Peking University State Key Laboratory for Turbulence and Complex Systems Department of Advanced Manufacturing and Robotics BIC-ESAT College of Engineering Beijing100871 China

This paper investigates the cooperative link configuration problem for Autonomous Underwater Vehicle (AUV) in Underwater Acoustic (UWA) sensor networks with Energy Harvesting (EH), which aims to maximize long-term cumulative capacity by jointly optimizing cooperation relay, AUV transmission power, and relay transmission power. Subject to unknown time-varying Channel State Information (CSI), unpredictable stochastic EH, and variable communication topology caused by AUV mobility, the proposed problem is an unknown dynamic combination decision optimization problem. To resist the change of the dynamic UWA communication topology with unknown time-varying CSI and EH, a novel collaborative contextual multi-armed bandit learning framework is proposed, which allows relay nodes to learn cooperatively and enriches the learning information of relays. Consequently, the proposed learning framework can resist the change of the dynamic UWA communication topology efficiently, thereby achieving the superior link configuration quickly. Besides, a learning-rate adjustment rule for the dynamic UWA communication topology is proposed to adaptively balance the exploration-exploitation, thereby avoiding missing the real superior joint relay-power configuration. Finally, simulation results show the significantly superiority of the proposed scheme. © 2015 IEEE.

关键词： Cooperative communication

Applying Pre-trained Multilingual BERT in Embeddings for Improved Malicious Prompt Injection Attacks Detection

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Rahman, Md Abdur Shahriar, Hossain Wu, Fan Cuzzocrea, Alfredo Dept. of Intelligent Systems and Robotics University of West Florida PensacolaFL United States Center for Cybersecurity University of West Florida PensacolaFL United States Department of Computer Science Tuskegee University TuskegeeAL United States iDEA Lab University of Calabria Rende Italy

Large language models (LLMs) are renowned for their exceptional capabilities, and applying to a wide range of applications. However, this widespread use brings significant vulnerabilities. Also, it is well observed that there are huge gap which lies in the need for effective detection and mitigation strategies against malicious prompt injection attacks in large language models, as current approaches may not adequately address the complexity and evolving nature of these vulnerabilities in real-world applications. Therefore, this work focuses the impact of malicious prompt injection attacks which is one of most dangerous vulnerability on real LLMs applications. It examines to apply various BERT (Bidirectional Encoder Representations from Transformers) like multilingual BERT, DistilBert for classifying malicious prompts from legitimate prompts. Also, we observed how tokenizing the prompt texts and generating embeddings using multilingual BERT contributes to improve the performance of various machine learning methods: Gaussian Naive Bayes, Random Forest, Support Vector Machine, and Logistic Regression. The performance of each model is rigorously analyzed with various parameters to improve the binary classification to discover malicious prompts. Multilingual BERT approach to embed the prompts significantly improved and outperformed the existing works and achieves an outstanding accuracy of 96.55% by Logistic regression. Additionally, we investigated the incorrect predictions of the model to gain insights into its limitations. The findings can guide researchers in tuning various BERT for finding the most suitable model for diverse LLMs vulnerabilities. © 2024, CC BY.

关键词： Embeddings

Documentation and Modeling of ROS systems

学校读者我要写书评

暂无评论

Documentation and Modeling of ROS Systems

IEEE Southeastcon

作者： William R. Drumheller David C. Conner Capable Humanitarian Robotics and Intelligent Systems Lab (CHRISLab) Christopher Newport University Newport News VA

The Robot Operating System (ROS) is frequently used to develop complex robotic systems with many interacting software components. While ROS provides several tools for visualizing active ROS deployments, it lacks tools for Model-Based Engineering (MBE) during the development stage, and ROS lacks tools for easily documenting the interfaces and deployed components other than specific configuration and launch files themselves. This paper presents a software tool chain - dubbed CHRIS ROS Modeling - that enables users to document an existing ROS deployment. The information is stored as instances of a simplified metamodel in portable formats. The resulting information serves to document an existing system and will serve as the foundation for future work to develop an integrated MBE system. This paper compares our approach to similar approaches currently in development.

关键词： Visualization System testing Computational modeling Documentation Tools US Department of Transportation Portable document format