检索结果-内蒙古大学图书馆

Spatial Graph-Based Localization and Navigation on Scaleless Floorplan

IEEE robotics and Automation Letters 2024年第4期9卷 3932-3939页

作者： Ewe, Zu Lin Chang, Fu-Hao Huang, Yi-Shiang Fu, Li-Chen National Taiwan University Department of Electrical Engineering Taipei100 Taiwan National Taiwan University NTU Center for Artificial Intelligence and Advanced Robotics Taipei100 Taiwan

Effective navigation in unfamiliar environments remains a critical challenge for successful deployment. Current navigation methods, which rely on autonomous or teleoperated exploration and map building, pose technical difficulties for inexperienced end-users. In contrast, humans can effectively navigate using abstract floorplans, suggesting the potential for service robots to leverage similar techniques. The practical application of floorplan-based navigation;however, is currently limited by methods that require pre-exploration or floorplan with accurate measurements or scale. This letter aims to address the aforementioned challenges and investigate the feasibility of floorplan-based navigation in unfamiliar environments. Specifically, we propose a novel scale-invariant floorplan localization method, enabling navigation without relying on precise scale information. Furthermore, we introduce an incremental graph augmentation approach that enriches the floorplan representation with traversability information derived from robot observations. Finally, we develop an efficient navigation framework capable of utilizing both the inherent structure of the floorplan and real-time observations. Experimental results demonstrate that our scale-invariant floorplan localization method outperforms baseline methods in most cases when floorplan scale information is unavailable, and our graph-based navigation system exhibits superior success and efficiency as compared to grid-based counterparts. The outcomes of this research contribute to the advancement of service robot deployment in unfamiliar environments, particularly in scenarios where extensive exploration and map building may be impractical or technically challenging for end-users. © 2016 IEEE.

关键词： Navigation systems

来源：评论

学校读者我要写书评

暂无评论

Design and Analysis of Programmed Face Monitoring System

Design and Analysis of Programmed Face Monitoring System

引用

2023 IEEE International Conference on Electrical, Electronics, Communication and Computers, ELEXCOM 2023

作者： Manjhi, Ashutosh Kewat Rajawat, Amit Singh Srivastava, Anshika Artificial Intelligence and Robotics Department of Information Technology India Gwalior474005 India

ISBN: (纸本)9798350305111

This study proposes a unique method for a programmed face monitoring system that makes use of a Proportional-Integral-Derivative (PID) controller and Mediapipe, a well-known computer vision framework. The system makes use of Mediapipe's real-time face detection technology and a PID controller algorithm to precisely monitor the identified face by dynamically adjusting the pan and tilt angles of a servo motor system. The usefulness of the proposed approach is demonstrated by the graphical representation depicting setting time, steady-state error and the oscillation type. The results were compared between Proportional-Integral-Derivative (PID), Proportional-Derivative (PD) and Proportional (P) controllers. The PID controller showed favorable results among all three of the controllers utilized in the face monitoring system. For PID controller-based faced monitoring system, the setting time of the system designed is obtained as 1 second, the steady-state error is low and the oscillation type observed is critically damped. The usage of this method for programmed face monitoring systems has a potential application in a number of fields, including robotics, human-computer interaction, and surveillance systems. © 2023 IEEE.

关键词： Face recognition

来源：评论

学校读者我要写书评

暂无评论

Radar Dead-reckoning Based EKF-SLAM Using Virtual Line Segment

引用

Journal of Institute of Control, robotics and Systems 2024年第9期30卷 996-1003页

作者： Lee, Min-Ho Seo, Kyeong-Wook Hwang, Dong-Yun Song, Jin-Woo Do, Hoang-Viet Department of Intelligent Mechatronics Engineering and Convergence Engineering for Intelligent Drone Sejong University Korea Republic of Department of Artificial Intelligence and Robotics and Convergence Engineering for Intelligent Drone Sejong University Korea Republic of

In this paper, we propose a dead reckoning–based radar simultaneous localization and mapping (SLAM) system utilizing constrained virtual line segments. In particular, we utilize Doppler velocity from radar for odometry while constructing a map using 3D radar point-cloud measurements. Unlike the conventional approach where accelerometers are employed, our system integrates radar velocity for pose calculation, thereby preventing double-integration and accelerometer bias and potentially improving the localization accuracy. Additionally, we analyze radar 3D point cloud characteristics and employ a recursive random sample consensus algorithm to eliminate outliers. The resulting set of inlier point clouds is then used to predict virtual line segments. Using virtual lines instead of feature points mitigates the sparse and noisy nature of radar point clouds. Moreover, considering the typical characteristics of indoor environments where walls intersect perpendicularly, we impose constraints on estimates to enhance performance further. We evaluate the performance of the proposed algorithm through experiments conducted in real indoor environments using an unmanned ground vehicle platform. Our results demonstrate significant improvements over traditional odometer-based systems, with reductions of approximately 41% in 2D trajectory error and 49.5% in heading angle error. © ICROS 2024.

关键词： Accelerometers

来源：评论

学校读者我要写书评

暂无评论

Healthcare Monitoring Companion: The Social Robot 3

Healthcare Monitoring Companion: The Social Robot

引用

3rd International Conference on artificial intelligence For Internet of Things, AIIoT 2024

作者： Goraksha, Rahul Bongale, Arunkumar Sayyad, Sameer Kumar, Satish Symbiosis Institute of Technology Department of Robotics and Automation Pune412115 India Symbiosis Centre for Applied Artificial Intelligence Department of Robotics and Automation Pune412115 India

ISBN: (纸本)9798350372120

There is a rising scarcity of healthcare professionals to meet the demand of taking care of aged people. Social robots are becoming more and more popular because of the benefits they provide;during the pandemic, in particular, several of their primary applications are advantageous. The purpose of this work is to build the prototype of social robots and carry out the integration of those robots into healthcare monitoring using Thing Speak and Raspberry Pi. This prototype of social robots with Thing Speak and Raspberry Pi hosts an array of sensors (for measuring a multitude of data from the body of humans) like heartbeat sensor, Dallas temperature sensor, Radio-Frequency Identification (RFID), ultrasonic sensor, and Infrared (IR) sensor along with other components like motor controlling module and DC motor. The data from these sensors are transmitted to the cloud platform Thing Speak, offering remote monitoring and data analysis capabilities and the Raspberry Pi acts as the central processing element in our design, which additionally helps in triggering corresponding alerts regarding the patient status. Finally, we experimentally validate our design using 2 cases and then we present the assessment of the patient alerting feature. © 2024 IEEE.

关键词： Radio frequency identification (RFID)

来源：评论

学校读者我要写书评

暂无评论

Integrating fuzzy metrics and negation operator in FCM algorithm via genetic algorithm for MRI image segmentation

引用

Neural Computing and Applications 2024年第27期36卷 17057-17077页

作者： Kutlu, Fatih Ayaz, İbrahim Garg, Harish Department of Artificial Intelligence and Robotics Van Yüzüncü Yıl University Van Turkey Department of Computer Technologies Bitlis Eren University Bitlis Turkey Punjab Patiala147004 India

In this study, we redefine FCM algorithm by integrating fuzzy set theory, fuzzy metrics, and Sugeno negation principles. This innovative approach overcomes the limitations inherent in conventional machine learning models, especially in situations characterized by uncertainty, noise, and ambiguity. Our model utilizes the membership degrees from fuzzy set theory, and transforms the concept of proximity defined by fuzzy metrics into a minimization problem. This transformation is achieved using a linguistic negation operator, which is crucial for optimizing FCM algorithm's objective function. A significant innovation in our research is the use of GA for optimizing parameters within the contexts of fuzzy metrics and Sugeno negation. The precise optimization capabilities of GA greatly enhance the sensitivity and adaptability of FCM algorithm, thereby improving overall performance. By leveraging the meticulous parameter adjustments provided by GA, our approach has shown superior results in practical applications, such as brain MRI image segmentation, surpassing traditional methods. Experimental results highlight the considerable enhancements our proposed FCM algorithms bring over existing methods across various performance metrics. In conclusion, this study makes a valuable addition to the field of fuzzy-based machine learning methodologies. It combines the optimization strength of GA with the flexible classification capabilities of fuzzy logic. The integration of Sugeno negation and fuzzy metrics not only improves the accuracy and precision of FCM algorithm but also provides significant benefits in handling complex and ambiguous datasets. This research signifies a major advance in machine learning and fuzzy logic, setting the stage for future applications and studies. © The Author(s) 2024.

关键词： Image segmentation

来源：评论

学校读者我要写书评

暂无评论

Leveraging Dual Timestamps to Enhance Process Model Construction Using Time-Infused Inductive Miner

引用

IEEE Access 2025年 13卷 88535-88556页

作者： Effendi, Yutika Amelia Kim, Minsoo Airlangga University Faculty of Advanced Technology and Multidiscipline Department of Robotics and Artificial Intelligence Engineering Surabaya60115 Indonesia Pukyong National University College of Engineering Department of Industrial and Data Engineering Busan48547 Korea Republic of

In business process management, the increasing complexity of modern business processes underscores the need for accurately understanding parallel activities to efficiently allocate and utilize limited corporate resources. However, many existing event log analysis algorithms focus solely on the sequence of executed activities, limiting their ability to model processes involving simultaneous activities effectively. This research seeks to enhance the inductive miner, which is known for its flexibility, formal guarantees, and scalability, by enabling it to better capture and represent the complexities of real-world business processes. The core modification involves leveraging dual timestamps, which utilize both the start and end times of activities. This approach enhances the algorithm's ability to understand parallel structures within process models. Through experimentation, limitations of single timestamp were identified, including reduced accuracy in modeling parallel activities. The proposed approach also demonstrates the capability to effectively minimize dummy activities when discovering process models containing both sequential and parallel activities. Experimental results highlight the improved performance of the proposed algorithm, with metrics showing a fitness score of 0.815, precision of 0.827, generalization of 0.846, and simplicity of 0.834. Moreover, the algorithm achieves the highest conformance results with fewer traces required compared to the alpha miner, fuzzy miner, heuristic miner, and the original inductive miner. © 2013 IEEE.

关键词： Enterprise resource planning

来源：评论

学校读者我要写书评

暂无评论

Using Generative AI for Human-robot Handover

引用

Journal of Institute of Control, robotics and Systems 2025年第5期31卷 496-501页

作者： Park, Mu Yeol Park, Kyu Min Department of Artificial Intelligence and Robotics Sejong University Korea Republic of

Recent advancements in human-robot interaction (HRI) have highlighted the importance of natural and efficient human-robot handover techniques, particularly in the field of industrial and service robotics. Traditional learning-based handover methods rely on large-scale data collection, which is costly and time-consuming. In this paper, we explore the use of generative AI, specifically OpenAI's Sora, to generate synthetic video for training a handover recognition model. Based on the synthetic training data, we propose a framework that can be used for the following three key handover scenarios: recognizing H2R (human-to-robot) handover initiation, recognizing R2H (robot-to-human) handover initiation, and object searching for R2H handover. The model is based on the YOLO11n architecture, achieving high accuracy in recognizing the user’s hand and object within the three handover scenarios. Experimental results demonstrate that the model trained only with synthetic data can effectively generalize to real-world handover scenarios with a precision of 0.94 and recall of 0.97. This paper highlights the potential of generative AI to change the data collection paradigm for robot learning, in the context of efficiency and expansion to various scenarios. Future work will focus on extending our approach to diverse objects, users, and handover scenarios. Copyright© ICROS 2025.

关键词： Social robots

来源：评论

学校读者我要写书评

暂无评论

A Study on Application of Self-supervised Learning for Multispectral Pedestrian Detection

引用

Journal of Institute of Control, robotics and Systems 2025年第5期31卷 526-535页

作者： Heo, Jae-Yeon Jung, Ui-Cheol Hwang, Yujin Choi, Yukyung Department of Artificial Intelligence and Robotics Sejong University Korea Republic of

Pedestrian detection has long been a prominent area of focus in computer vision due to its extensive applications, particularly in the development of autonomous driving systems. While color image-based pedestrian detection has achieved significant progress, its vulnerability to illumination changes has prompted more research into multispectral pedestrian detection, which leverages both visible and infrared modalities for robust detection in varying lighting conditions. Despite advances in multispectral detection models, research on applying self-supervised learning for feature representation in this area remains limited. In this study, we extend existing self-supervised learning approaches for color image data to multispectral pedestrian detection. By pre-training with self-supervised learning methods and fine-tuning, we evaluate their extensibility to the multispectral domain. Experimental results demonstrate that self-supervised learning enhances multispectral pedestrian detection, especially in nighttime conditions, outperforming models pre-trained only on color modalities. Furthermore, we analyze the framework of self-supervised learning method that perform well on multispectral data, providing insights into which framework elements enhance multispectral pedestrian detection. This study underscores the potential of self-supervised learning in multispectral pedestrian detection and its role in advancing detection robustness for critical applications such as autonomous driving. © ICROS 2025.

关键词： Self-supervised learning

来源：评论

学校读者我要写书评

暂无评论

A novel piezoelectric motor design and its controller

A novel piezoelectric motor design and its controller

引用

4th International Conference on Smart and Sustainable Developments in Materials, Manufacturing and Energy Engineering, SME 2022

作者： Veeresha, R.K. Bhat, Krishna Muralidhara Department of Robotics and Artificial Intelligence Udupi574110 India Department of Mechanical Engineering Udupi574110 India

The paper aims to develop a novel piezoelectric motor and its controller;here the motor shaft needs to be rotated fully by 360° using two sets of Parallelly Prestressed Actuators (PPA) and one set of lock release actuators (LR). The first set of PPA drives the driving element and produces a displacement of 39 lm for an applied input voltage of 150 V & rotates by an angle of 0.145°, thus taking 2483 steps to complete one complete rotation. The second set of PPA is deployed to transfer the rotation caused by the first set of PPA to the output shaft. The lock release actuator will arrest the rotation that occurred and avoids the shaft from moving back to its original position till the next cycle begins. The above set of operations is continued to obtain continuous rotation of the piezoelectric motor shaft. There is a FAP used to hold the shaft during the power-off condition. The mathematical controller is developed to obtain both forward and reverse rotation of the shaft by applying an input voltage, further the developed model is simulated using MATLAB Simulink software to illustrate the working of a piezoelectric motor. Copyright © 2023 Elsevier Ltd. All rights reserved.

关键词： Rotation

来源：评论

学校读者我要写书评

暂无评论

A Novel Approach to Continual Knowledge Transfer in Multilingual Neural Machine Translation Using Autoregressive and Non-Autoregressive Models for Indic Languages

引用

IEEE Access 2025年 13卷 89775-89810页

作者： Sheshadri, Shailashree K. Gupta, Deepa Paul, Biswajit Bhavani, J. Siva Amrita School of Computing Department of Computer Science and Engineering Amrita Vishwa Vidyapeetham Bengaluru560035 India DRDO CV Raman Nagar Centre for Artificial Intelligence and Robotics Bangalore India

Recent progress in multilingual pre-trained models has significantly improved translation quality for Indic languages. However, extending these models to new languages via fine-tuning or retraining remains computationally costly and often leads to parameter interference, degrading performance on previously learned or typologically distant languages. While continual learning offers a promising alternative for incremental language addition, its application in Indic contexts is still limited and faces challenges in generalization across diverse linguistic settings. To overcome these issues, we propose a Continual Knowledge Transfer (CKT) framework for efficient and scalable multilingual adaptation. CKT is realized in both autoregressive (MNMT) and non-autoregressive (Switch-GLAT) architectures, yielding two variants: MNMT+CKT and Switch-GLAT+CKT. Rather than retraining the entire model, CKT freezes the multilingual base and updates only parameters relevant to the newly added language. Key innovations include gradient-based knowledge pruning, sequential teacher integration, and dynamic vocabulary expansion for minimizing interference and maximizing cross-lingual retention. Comprehensive evaluations on the IN22-Conv and IN22-Gen benchmark datasets demonstrate that both MNMT+CKT and Switch-GLAT+CKT consistently outperform established baselines, such as IndicTrans2, Google Translate, GPT-4-32K, LLaMA-2-17B, and NLIP-LAB-IITH. The proposed multi-step distillation approach, MNMT+CKT, consistently outperforms conventional fine-tuning and Knowledge Transfer (MNMT+KT) strategies for incremental adaptation of linguistically diverse Indic languages. On IN22-Conv, BLEU improvements range from +4.93 (Kashmiri → English) to +11.48 (Assamese → English) and similar improvements are seen for IN22-Gen. The method also achieves substantial reductions in trainable parameters - 19.80% (Nepali) to 66.87% (Kannada) - while enabling up to 4x faster inference when integrated with the Switch-GL

关键词： Neural machine translation

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：