检索结果-内蒙古大学图书馆

RSI/ISM International Conference on robotics and Mechatronics (ICROM)

作者： Asal Roudbari Mahdi Dalvand Navid Pasiar Navid Asadi Mohammad Yousefi Mehdi Tale Masouleh Electrical Engineering Department Sharif University of Technology Tehran Iran Human & Robot Interaction Lab University of Tehran Tehran Iran Human & Robot Interaction Lab School of Electrical&Computer Engineering University of Tehran Tehran Iran

Autonomous robotic grasping has emerged as a valuable capability for automating everyday household tasks like table setting (bossing), cleaning, and meal preparation. Still, accurately detecting and manipulating diverse everyday objects in unstructured home environments remains a significant challenge. In this study, an integrated robotic system combining YOLOv8s-seg deep learning for object detection and segmentation with computer vision techniques to determine optimal grasping points is proposed. The accuracy of this model has also been compared with other models through their mAP0.5, mAP.05-.95, precision and recall. A custom dataset of over 1000 images containing common household objects was created, with manual annotation and augmentation to train the YOLOv8s-seg segmentation model. The approach processes input images to understand desired arrangements and identify target objects and their locations for grasping. The detected objects are localized using camera-to-robot calibration and aligned with estimated orientations from computer vision techniques. Extensive tests have proven that representative household objects in cluttered arrangements have a grasping success rate of over 90% using a 3DOF Delta Parallel robot with a 2-finger gripper guided entirely by visual perception. By overcoming real-world challenges like transparent and reflective materials, the perception-guided robots provide precise and efficient robotic grasping in unstructured home environments. This integrated approach combines robust perception and localization, providing promising evidence for reliable automation of repetitive household organization tasks 1 . 1 A video of real-time grasp performance is uploaded as supplementary material.

关键词：

来源：评论

学校读者我要写书评

暂无评论

AGILE: Approach-based Grasp Inference Learned from Element Decomposition

AGILE: Approach-based Grasp Inference Learned from Element D...

引用

RSI/ISM International Conference on robotics and Mechatronics (ICROM)

作者： MohammadHossein Koosheshi Hamed Hosseini Mehdi Tale Masouleh Ahmad Kalhor Mohammad Reza Hairi Yazdi Human and Robot Interaction Laboratory School of Mechanical Engineering University of Tehran Human and Robot Interaction Laboratory School of Electrical and Computer Engineering University of Tehran School of Mechanical Engineering College of Engineering University of Tehran

humans, this species expert in grasp detection, can grasp objects by taking into account hand-object positioning information. This work proposes a method to enable a robot manipulator to learn the same, grasping objects in the most optimal way according to how the gripper has approached the object. Built on deep learning, the proposed method consists of two main stages. In order to generalize the network on unseen objects, the proposed Approach-based Grasping Inference involves an element decomposition stage to split an object into its main parts, each with one or more annotated grasps for a particular approach of the gripper. Subsequently, a grasp detection network utilizes the decomposed elements by Mask R-CNN and the information on the approach of the gripper in order to detect the element the gripper has approached and the most optimal grasp. In order to train the networks, the study introduces a robotic grasping dataset collected in the Coppeliasim simulation environment. The dataset involves 10 different objects with annotated element decomposition masks and grasp rectangles. The proposed method acquires a 90% grasp success rate on seen objects and 78% on unseen objects in the Coppeliasim simulation environment. Lastly, simulation-to-reality domain adaptation is performed by applying transformations on the training set collected in simulation and augmenting the dataset, which results in a 70% physical grasp success performance using a Delta parallel robot and a 2 -fingered gripper.

关键词：

来源：评论

学校读者我要写书评

暂无评论

HUGGA: human-like Grasp Generation with Gripper’s Approach State Using Deep Learning

HUGGA: Human-like Grasp Generation with Gripper’s Approach ...

引用

RSI/ISM International Conference on robotics and Mechatronics (ICROM)

作者： Elnaz Balazadeh Mehdi Tale Masouleh Ahmad Kalhor Human and Robot Interaction Laboratory University of Tehran Tehran Iran Human and Robot Interaction Laboratory School of Electrical and Computer Engineering University of Tehran Tehran Iran

In this study, an attempt has been made to develop a new method for producing a successful grasp based on human behavior, by considering the gripper’s approach state in the final grasp of an object. The dataset collected for the purpose of this paper consists of 14 different daily usage objects with different orientations (In total 2526 images) which a human approaches and grasps them. Data collection has been done based on a protocol shown in TableI to cover all achievable states. From Fig1, it can be seen that the dataset includes one image from a single object without the user’s hand. The second image contains the human hand approach and its configuration regarding the object which is subject to be grasped and the third image represents the object which is grasped by the human hand and thus evidently contains both object and human hand grasp configuration. It is important to acknowledge that providing a static hand image does not suffice to facilitate the acquisition of human-like grasping capabilities by robot grippers, so, a mapping from hand fingers to a two-finger gripper should be implemented to fill this gap. In the collected dataset, all grasps are performed by the index finger and thumb to make it possible to map the coordinates of each finger on a gripper finger. Using MediaPipe, which operates a single shot detection network, with only an RGB image containing a human hand the index finger and thumb landmarks were extracted. Finally, a detailed human approach and grasp information were availab.e, and with the collected data, a two-input and one-output deep neural network was trained. The designed model produced a grasp with a human pattern by receiving only the RGB image of the object and the approach vector to the object. Ultimately, through testing and evaluation based on existing criteria, the trained network successfully produced a grasp with 61% accuracy in grasp rectangle overlap, 97% in grasp angle and overall 60.5% met two established grasp crite

关键词：

来源：评论

学校读者我要写书评

暂无评论

Grasp the Graph (GtG): A Super Light Graph-RL Framework for robotic Grasping

Grasp the Graph (GtG): A Super Light Graph-RL Framework for ...

引用

RSI/ISM International Conference on robotics and Mechatronics (ICROM)

作者： Ali Rashidi Moghadam Mehdi Tale Masouleh Ahmad Kalhor Human and Robot Interaction Laboratory School of Electrical and Computer Engineering University of Tehran Tehran Iran

This paper introduces Grasp the Graph, an innovative framework addressing challenges posed by processing point clouds with conventional neural networks for robotic grasping. Such challenges often lead to the conversion of point clouds to multi-view images or voxel grids, requiring large networks with millions of parameters. Grasp the Graph formulates robotic grasping as a one-step reinforcement learning problem and leverages graph representation and graph neural networks to efficiently process geometric relationships in point clouds. The lightweight design with only 67k parameters facilitates swift training and strong generalization to novel objects, offering the capability to propose multiple valid grasping candidates. Empirical testing demonstrates the effectiveness of the proposed approach, achieving a 90.4% success rate on test objects while being trained with only 10 objects. Even with a three-camera setup during training, reducing the cameras during evaluation minimally impacts performance, resulting in less than a 3% decrease. 1

关键词：

来源：评论

学校读者我要写书评

暂无评论

Developing a Reinforcement Learning-Based Actor-Critic Neural Network Controller for General Form of Serial Chain robots with Torque Constraints

SSRN

引用

SSRN 2023年

作者： Bahrami, Vahid Masouleh, Mehdi Tale Kalhor, Ahmad School of Electrical and Computer Engineering University of Tehran Human and Robot Interaction Laboratory Tehran Iran

This paper presents a regulation control system for Serial Chain robots (SCRs) using Reinforcement Learning (RL) algorithms. The main challenge is the high dimensionality of the state and action spaces due to the multiple degrees-of-freedom in SCRs. Traditional optimal control methods rely on mathematical models and solving differential equations, while RL offers a computationally intensive alternative. In this study, a controller is proposed with key features: reduced reliance on robot dynamics knowledge, applicability to various robots (SCRs and Cable-Driven Serial Chain robots (CDSCRs)), and the ability to limit torque. This method avoids complex computations of Coriolis vectors by modifying variables in the robot's dynamics to achieve a zero equilibrium point for regulation error. Actor-critic Neural Network (NN) based on RL methods is used for the adaptation laws of the regulator torque, while a MLP-NN is employed for the desired steady-state torque. Stability of the controller designs is ensured using suitable Lyapunov functions. Simulations on a 4-DoF CDSCR, representing a finger segment of the human body, validate the proposed control method's effectiveness. © 2023, The Authors. All rights reserved.

关键词： robots

来源：评论

学校读者我要写书评

暂无评论

Developing a Reinforcement Learning-Based Actor-Critic Neural Network Controller for General Form of Serial Chain robots with Torque Constraints

SSRN

引用

SSRN 2023年

作者： Bahrami, Vahid Masouleh, Mehdi Tale Kalhor, Ahmad School of Electrical and Computer Engineering University of Tehran Human and Robot Interaction Laboratory Tehran Iran

关键词： robots

来源：评论

学校读者我要写书评

暂无评论

Experimental Study on Chess Board Setup Using Delta Parallel robot Based on Deep Learning

Experimental Study on Chess Board Setup Using Delta Parallel...

引用

RSI/ISM International Conference on robotics and Mechatronics (ICROM)

作者： Parsa Yarmohammadi Navid Asadi Khomami Mehdi Tale Masouleh Mohammad Reza Zakerzadeh Human and Robot Interaction Laboratory School of Mechanical Engineering University of Tehran Tehran Iran Human and Robot Interaction Laboratory School of Electrical and Computer Engineering University of Tehran Tehran Iran School of Mechanical Engineering University of Tehran Tehran Iran

In recent years, robotics and object detection applications have seen remarkable improvements in executing routine tasks. In the near future, it is expected that most of the tasks can be performed by means of these two foregoing tools. This paper presents a system that combines these two approaches in order to automate the setup of a chessboard using a Delta parallel robot equipped with a two-finger gripper designed for grasping small objects. This system utilizes the so-called YOLO model for object detection in order to detect the chess pieces. The system begins by taking a picture of a scene, where the chess pieces are placed randomly around a board. The location of the chessboard is predefined and YOLO is used to detect and locate chess pieces in the given picture. The final object detection model used to execute this task is YOLOv5m, although this model encountered some challenges in detecting four types of chess pieces from a top view, these challenges were addressed by increasing the instances of those objects in the dataset. Upon extracting the location of each piece, the coordinates are sent to the robot, and the Delta parallel robot arranges the initial setup of a chess game. The model achieved an overall mAP of 0.99, indicating excellent precision and the grasp success rate in the process of grasping the chess pieces was 95%. This study demonstrates the potential of automating everyday tasks by exploring the intersection of robotics and artificial intelligence and can be well extended to tasks, namely, sorting medications and pills in hospitals, sorting different wastes like plastic and papers for recycling, identifying and removing spoiled fruits in agricultural industries, etc.

关键词：

来源：评论

学校读者我要写书评

暂无评论

A 2D Geometry Based Grasping Pose Generation Algorithm for a Two-Finger robot Hand

A 2D Geometry Based Grasping Pose Generation Algorithm for a...

引用

Iranian Conference on electrical engineering (ICEE)

作者： Arash Akbari Arman Akbari Mehdi Tale Masouleh Department of Computer Science University of Tehran Tehran Iran Human and Robot Interaction Laboratory School of Electrical and Computer Engineering Tehran Iran

robot grasping is of paramount importance in industrial and service robotics. In recent years, various data-driven algorithms have been proposed to solve the problem of grasp detection and a part of them are based on reinforcement learning (RL) approaches. In a variety of proposed algorithms, random key points are being employed which will make the learning process inefficient and time-consuming. In this paper, a geometry-based algorithm is presented which can find grasp poses based on the geometry of the unknown object and propose the ones which may lead to successful grasping. For the grasp contacts computation part, the presented algorithm produces a finite number of key points based on the 2D shape of the object from a specific point of view. Afterward, it will narrow down the candidate points and output a finite number of successful grasp poses based on three grasp quality metrics for various unknown objects. Three approaches are proposed in order to achieve center points which can describe different parts of a 2D shape. Then, the obtained points are used as the center of circles which are tangent to the 2D shape contour. Also, a new grasp quality metric is proposed. The time of the grasp and the amount of object disorientation after grasping are considered as a metric to evaluate the successfulness of the grasp. Simulation results demonstrate that the proposed algorithm for unknown object grasping can find a finite number of successful grasp poses for different seen or unseen objects without using any random point,

关键词：

来源：评论

学校读者我要写书评

暂无评论

Experimental Study on Reducing the Oscillations of a Cable-Suspended Parallel robot for Video Capturing Purposes using Simulated Annealing and Path Planning 7

Experimental Study on Reducing the Oscillations of a Cable-S...

引用

7th International Conference on Signal Processing and Intelligent Systems, ICSPIS 2021

作者： Ghanatian, Mohammad Yazdi, Mohammad Reza Hairi Masouleh, Mehdi Tale School of Mechanical Engineering University of Tehran Human and Robot Interaction Laboratory Tehran Iran School of Mechanical Engineering University of Tehran Tehran Iran School of Electrical and Computer Engineering University of Tehran Human and Robot Interaction Laboratory Tehran Iran

ISBN: (纸本)9781665409384

This paper aims to smoothen the movements of an under-constrained cable-suspended parallel robot which carries a camera for video capturing purposes, especially for video capturing of football games. This goal is achieved by means of an accurate while simple PID controller optimized by the Simulated Annealing algorithm and implemented on the joint space. Moreover, a planning strategy is considered for the joint space trajectory of the robot which guarantees the zero jerk, acceleration, and velocity at the start and the end of motions. On this regard, a septic function for each joint of the robot is considered and the corresponding boundary conditions are applied to the function to make the end-effector movements less oscillatory. This method is implemented and tested on an experimental setup while the end-effector oscillation data is recorded using an IMU sensor attached to the end-effector of the robot. Applying frequency analysis on the oscillatory data of the end-effector reveals that this simple method, on average, resulted in a 33.8% reduction in the average amplitude of the end-effector oscillations. Moreover, the maximum joint space error was decreased by 76.68% when using septic joint profile compared to the ordinary linear Cartesian trajectory planning approach. Upon applying the proposed strategy, the error of the controller has been reduced by 92.26% with respect to the previous research performed on this experimental setup. Without requiring any knowledge on the dynamic model of the robot or the natural frequencies of the end-effector or using any complex controller, this method significantly increased the smoothness and accuracy of the robot movements. The proposed method can be regarded as a definitive asset when this robot is used for video capturing purposes. © 2021 IEEE.

关键词： Simulated annealing

来源：评论

学校读者我要写书评

暂无评论

Practical Implementation of Real-Time Waste Detection and Recycling based on Deep Learning for Delta Parallel robot

Practical Implementation of Real-Time Waste Detection and Re...

引用

International eConference on computer and Knowledge engineering (ICCKE)

作者： Hasan Jalali Shaya Garjani Ahmad Kalhor Mehdi Tale Masouleh Parisa Yousefi School of Electrical and Computer Engineering Human and Robot Interaction Laboratory University of Tehran Tehran Iran School of Computer Engineering Imam Reza International University Mashhad Iran

Intelligent robots play an essential role in waste management and recycling due to their high speed and a wide variety of applications. In this paper, two methods for waste detection and accurate pick-and-place based on computer vision and neural networks are presented. The suggested methods have been put into practical application on a 3-DOF Delta parallel robot to show the accuracy and fastness of the foregoing method for real intelligence systems. The first method, Multi-Stage Detection, consists of two stages to detect the waste objects, namely, object localization and segmentation, and classification. The second method, known as One-Stage object detectors, such as YOLOv5, has the capability to simultaneously localize and classify the waste objects. The dataset utilized in this paper relies on the TrashNet dataset as its foundation. In order to improve the classification capabilities in the multi-stage method, a larger dataset was created by utilizing data augmentation. Also, for the one-stage method, a new multi-lab.l dataset is constructed based on the TrashNet dataset. Additionally, the results of the experimental implementation were compared based on time and evaluation metrics for detection and classification. The ResNet50 model achieved the highest accuracy in the multi-stage method, with 99.31% accuracy. In the one-stage detection method, the YOLOv5x model achieved the best mAP (@IoU = 0.75) of 97.4%, which outperformed the YOLOv5s model by 0.8 percent; however, the inference speed of the YOLOv5x in comparison with the YOLOv5s models was six times as slow. Therefore, the YOLOv5s model was employed in real-time online waste detection, which resulted in 82.1% mAP (@IoU = 0.5) after being trained on real images from the waste-sorting platform.

关键词：

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：