检索结果-内蒙古大学图书馆

Real-time FPGA implementation of the Semi-Global Matching stereo vision algorithm for a 4K/UHD video stream

学校读者我要写书评

暂无评论

arXiv 2023年

作者： Grabowski, Mariusz Kryjak, Tomasz Embedded Vision Systems Group Computer Vision Laboratory Department of Automatic Control and Robotics Agh University of Science and Technology Krakow Poland

In this paper, we propose a real-time FPGA implementation of the Semi-Global Matching (SGM) stereo vision algorithm. The designed module supports a 4K/Ultra HD (3840 ×2160 pixels @ 30 frames per second) video stream in a 4 pixel per clock (ppc) format and a 64-pixel disparity range. The baseline SGM implementation had to be modified to process pixels in the 4ppc format and meet the timing constrains, however, our version provides results comparable to the original design. The solution has been positively evaluated on the Xilinx VC707 development board with a Virtex-7 FPGA device. © 2023, CC BY.

关键词： Field programmable gate arrays (FPGA)

A Position-Force control Approach for an Underwater Vehicle-Manipulator System

学校读者我要写书评

暂无评论

A Position-Force Control Approach for an Underwater Vehicle-...

Ocean Studies (ICOS), International Conference on

作者： Aleksey A. Kabanov Kirill V. Dementiev Department of Informatics and Control in Technical Systems Sevastopol State University Sevastopol Russian Federation Robotics and Intelligent Control System Laboratory Sevastopol State University Sevastopol Russian Federation

The paper proposes an approach to position-force control in the underwater vehicle-manipulator systems. There are two different strategies to realize position-force control systems: through a decoupled or coupled controller. The first one is more flexible and is used in the present article as a PID-controllers array. The paper considers the generalized system design methodology applicable to control objects of any kinematic structure. Compensation blocks for manipulator disturbances and calculation of additional end-effector offset to maintain a constant force on the operating point are proposed. The system also utilizes the desired underwater vehicle position calculation based on the relativity of the vehicle position and operating point for a heading set point. The modeling results showed that the system is able to perform qualitatively force operations in a given point of the working space. In addition, the graphs revealed the compensator unit efficiency and confirmed the necessity of the additional offset calculation unit.

关键词：

A Graph-Based Self-Calibration Technique for Cable-Driven Robots with Sagging Cable

学校读者我要写书评

暂无评论

A Graph-Based Self-Calibration Technique for Cable-Driven Ro...

IEEE/RSJ International Conference on Intelligent Robots and systems (IROS)

作者： M. R. Dindarloo A. S. Mirjalili S. A. Khalilpour R. Khorrambakht Stephan Weiss H. D. Taghirad Advanced Robotics and Automated Systems (ARAS) Faculty of Electrical Engineering K. N. Toosi University of Technology Tehran Iran Department of Electrical and Computer Engineering Tandon School of Engineering Control/Robotics Research Laboratory New York University Brooklyn NY Department of Smart Systems Technologies University of Klagenfurt Austria

ISBN: (数字)9798350377705

ISBN: (纸本)9798350377712

The efficient operation of large-scale Cable-Driven Parallel Robots (CDPRs) relies on precise calibration of kinematic parameters and the simplicity of the calibration process. This paper presents a graph-based self-calibration framework that explicitly addresses cable sag effects and facilitates the calibration procedure for large-scale CDPRs by only relying on internal sensors. A unified factor graph is proposed, incorporating a catenary cable model to capture cable sagging. The factor graph iteratively refines kinematic parameters, including anchor point locations and initial cable length, by considering jointly onboard sensor data and the robot’s kineto-static model. The applicability and accuracy of the proposed technique are demonstrated through Finite Element (FE) simulations, on both large and small-scale CDPRs subjected to significant initialization perturbations.

关键词： Meters Accuracy Perturbation methods Kinematics Robot sensing systems End effectors Software Calibration Sensors Finite element analysis

Enhancing Thin-Film Wafer Inspection With A Multi-Sensor Array And Robot Constraint Maintenance

学校读者我要写书评

暂无评论

arXiv 2025年

作者： Sánchez-Arriaga, Néstor Eduardo Canzini, Ethan Espley-Plumb, Nathan John Farnsworth, Michael Pope, Simon Leyland, Adrian Tiwari, Ashutosh Department of Automatic Control & Systems Engineering University of Sheffield Sheffield United Kingdom Department of Materials Science & Engineering University of Sheffield Sheffield United Kingdom Airbus Robotics Airbus UK Chester United Kingdom

Thin-film inspection on large-area substrates in coating manufacture remains a critical parameter to ensure product quality;however, extending the inspection process precisely over a large area presents major challenges, due to the limitations of the available inspection equipment. An additional manipulation problem arises when automating the inspection process, as the silicon wafer requires movement constraints to ensure accurate measurements and to prevent damage. Furthermore, there are other increasingly important large-area industrial applications, such as Roll-to-Roll (R2R) manufacturing where coating thickness inspection introduces additional challenges. This paper presents an autonomous inspection system using a robotic manipulator with a novel learned constraint manifold to control a wafer to its calibration point, and a novel multi-sensor array with high potential for scalability into large substrate areas. We demonstrate that the manipulator can perform required motions whilst adhering to movement constraints. We further demonstrate that the sensor array can perform thickness measurements statically with an error of 0.5◦ from the calibration point whilst monitoring the RMSE and R2 over 1406 data points. These features are potentially useful for detecting displacement variations in R2R manufacturing processes. © 2025, CC BY-NC-ND.

关键词： Silicon wafers

6-DOF AUV Identification with MIMO-NARX

学校读者我要写书评

暂无评论

6-DOF AUV Identification with MIMO-NARX

International Conference on Industrial Engineering, Applications and Manufacturing (ICIEAM)

作者： Aleksey A. Kabanov Ivan Y. Lipko Kirill V. Dementiev Department of Informatics and Control in Technical Systems Sevastopol State University Sevastopol Russian Federation Robotics and Intelligent Control System Laboratory Sevastopol State University Sevastopol Russian Federation

This paper applies the MIMO NARX neural model to identify AUV with six degrees of freedom. Strongly nonlinear dependencies are modeled using a MIMO NARX model trained using «input-output» experimental data. This results in a grey box model with a partially known internal structure that can be supervised. This is achieved by specifying a certain regressors type, which is feedback to the model input. Obtained results show that the neural model trained by the Levenberg–Marquardt numerical algorithm provides acceptable performance and accuracy. Validation of the model was performed using data from full-scale experiments, which allowed to identify the weaknesses and advantages of the resulting system more accurately.

关键词：

An Efficient Learning control Framework with Sim-to-Real for String-Type Artificial Muscle-Driven Robotic systems

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Tao, Jiyue Zhang, Yunsong Rajendran, Sunil Kumar Zhang, Feitian The Robotics and Control Laboratory Department of Advanced Manufacturing and Robotics College of Engineering the State Key Laboratory of Turbulence and Complex Systems Peking University Beijing100871 China The BSS Technologies Inc. Takoma ParkMD20912 United States

Robotic systems driven by artificial muscles present unique challenges due to the nonlinear dynamics of actuators and the complex designs of mechanical structures. Traditional model-based controllers often struggle to achieve desired control performance in such systems. Deep reinforcement learning (DRL), a trending machine learning technique widely adopted in robot control, offers a promising alternative. However, integrating DRL into these robotic systems faces significant challenges, including the requirement for large amounts of training data and the inevitable sim-to-real gap when deployed to real-world robots. This paper proposes an efficient reinforcement learning control framework with sim-to-real transfer to address these challenges. Bootstrap and augmentation enhancements are designed to improve the data efficiency of baseline DRL algorithms, while a sim-to-real transfer technique, namely randomization of muscle dynamics, is adopted to bridge the gap between simulation and real-world deployment. Extensive experiments and ablation studies are conducted utilizing two string-type artificial muscle-driven robotic systems including a two degree-of-freedom robotic eye and a parallel robotic wrist, the results of which demonstrate the effectiveness of the proposed learning control strategy. Copyright © 2024, The Authors. All rights reserved.

关键词： Reinforcement learning

Optimising Graph Representation for Hardware Implementation of Graph Convolutional Networks for Event-based Vision

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Jeziorek, Kamil Wzorek, Piotr Blachut, Krzysztof Pinna, Andrea Kryjak, Tomasz Embedded Vision Systems Group Department of Automatic Control and Robotics AGH University of Krakow Poland Sorbonne Université CNRS LIP6 ParisF-75005 France

Event-based vision is an emerging research field involving processing data generated by Dynamic Vision Sensors (neuromorphic cameras). One of the latest proposals in this area are Graph Convolutional Networks (GCNs), which allow to process events in its original sparse form while maintaining high detection and classification performance. In this paper, we present the hardware implementation of a graph generation process from an event camera data stream, taking into account both the advantages and limitations of FPGAs. We propose various ways to simplify the graph representation and use scaling and quantisation of values. We consider both undirected and directed graphs that enable the use of PointNet convolution. The results obtained show that by appropriately modifying the graph representation, it is possible to create a hardware module for graph generation. Moreover, the proposed modifications have no significant impact on object detection performance, only 0.08% mAP less for the base model and the N-Caltech data set. Finally, we describe the proposed hardware architecture of the graph generation module. © 2024, CC BY.

关键词： Field programmable gate arrays (FPGA)

Robot Teleoperation via Virtual Reality: A Unity and ROS Approach

学校读者我要写书评

暂无评论

Robot Teleoperation via Virtual Reality: A Unity and ROS App...

Automation, robotics and Computer Engineering (ICARCE), International Conference on

作者： Matheus Nicolás Hernandez Brenno Henrique Ramos Roberto Simoni Doglas Negri Diego De Souza Luís Gonzaga Trabasso Robotics Development Department SENAI Innovation Institute for Manufacturing Systems Joinville Brazil Domingues Automation and Control Department SENAI Innovation Institute for Manufacturing Systems Joinville Brazil POSECM Universidade Federal de Santa Catarina UFSC Joinville Brazil

ISBN: (数字)9798331529505

ISBN: (纸本)9798331529512

This paper explores the integration of virtual reality (VR) with teleoperation systems to enhance human-robot interaction. With the rise of remote operations, traditional methods often rely on two-dimensional interfaces, making it challenging to control robots in complex three-dimensional environments. This research presents a VR teleoperation system using the Meta Quest 2 headset, enabling immersive control of a climbing robot with six degrees-of-freedom and magnetic adhesion. The system architecture integrates Unity for virtual simulation and the Robot Operating System (ROS) for real-time communication and control. A virtual environment allows the user to control the virtual robot model through the VR headset and controllers, with ROS serving as the communication bridge between the virtual and physical robots, enabling effective operations in various applications, including high-risk environments (A video is available at https://***/MiZDOV4IfLA).

关键词： Headphones Solid modeling Operating systems Virtual environments systems architecture Human-robot interaction Streaming media Climbing robots Real-time systems Robots

Embedded Graph Convolutional Networks for Real-Time Event Data Processing on SoC FPGAs

学校读者我要写书评

暂无评论

arXiv 2024年

作者： Jeziorek, Kamil Wzorek, Piotr Blachut, Krzysztof Pinna, Andrea Kryjak, Tomasz The Embedded Vision Systems Group Department of Automatic Control and Robotics AGH University of Krakow Poland The Sorbonne Université CNRS LIP6 ParisF-75005 France

The utilisation of event cameras represents an important and swiftly evolving trend aimed at addressing the constraints of traditional video systems. Particularly within the automotive domain, these cameras find significant relevance for their integration into embedded real-time systems due to lower latency and energy consumption. One effective approach to ensure the necessary throughput and latency for event processing systems is through the utilisation of graph convolutional networks (GCNs). In this study, we introduce a series of hardware-aware optimisations tailored for PointNet++, a GCN architecture designed for point cloud processing. The proposed techniques result in more than a 100-fold reduction in model size compared to Asynchronous Event-based GNN (AEGNN), one of the most recent works in the field, with a relatively small decrease in accuracy (2.3% for N-Caltech101 classification, 1.7% for N-Cars classification), thus following the TinyML trend. Based on software research, we designed a custom EFGCN (Event-Based FPGA-accelerated Graph Convolutional Network) and we implemented it on ZCU104 SoC FPGA platform, achieving a throughput of 13.3 million events per second (MEPS) and real-time partially asynchronous processing with a latency of 4.47 ms. We also address the scalability of the proposed hardware model to improve the obtained accuracy score. To the best of our knowledge, this study marks the first endeavour in accelerating PointNet++ networks on SoC FPGAs, as well as the first hardware architecture exploration of graph convolutional networks implementation for real-time continuous event data processing. We publish both software and hardware source code in an open repository: https://***/vision-agh/***. © 2024, CC BY.

关键词： Field programmable gate arrays (FPGA)