检索结果-内蒙古大学图书馆

A Force-driven and Vision-driven Hybrid Control Method of Autonomous Laparoscope-Holding Robot

学校读者我要写书评

暂无评论

A Force-driven and Vision-driven Hybrid Control Method of Au...

IEEE International Conference on Robotics and Automation (ICRA)

作者： Jin Fang Ling Li Xiaojian Li Hangjie Mo Pengxin Guo Xilin Xiao Yanwei Qu School of Management Hefei University of Technology Hefei China Key Laboratory of Process Optimization and Intelligent Decision-Making (Ministry of Education) Hefei University of Technology Hefei China

ISBN: (数字)9798350384574

ISBN: (纸本)9798350384581

Laparoscope-holding robots significantly enhance the stability and precision of visualization in minimally invasive surgeries. Most existing robots of this kind depend on visual servo systems and struggle with efficient, rapid adjustments in the field-of-view (FOV), especially when identifying organs and needles outside the FOV. This paper presents a laparoscope-holding robot system capable of employing both vision-driven and force-driven mechanisms for continuous and large-scale FOV adjustments, respectively. The system features an integrated tactile handle, enabling the reception of human-robot interaction forces during surgical navigation. We propose a hybrid control method that leverages both force and vision inputs for laparoscopic FOV adjustments. This approach integrates a virtual wrench, generated from visual information, and an interaction wrench, obtained from the tactile handle, into the robot's dynamic model, which complies with remote center of motion constraints. The interaction wrench's gain is adjusted with the gripping force on the integrated tactile handle, ensuring that unintended movements caused by accidental contacts are prevented, thus safeguarding operational safety. The proposed method eliminates the need to switch control modes, enabling simultaneous visual tracking and tactile interaction guidance. Experimental results demonstrate that the proposed method not only allows for FOV adjustments with surgical instrument guiding but also adapts well to large-scale FOV adjustment tasks.

关键词： Laparoscopes Visualization Tracking Instruments Force Human-robot interaction Switches

Hierarchically-Trusted Evidential Fusion Method with Consistency Learning for Multimodal Language Understanding

学校读者我要写书评

暂无评论

SSRN 2024年

作者： Yang, Ying Yan-Qiu, Yang Ren, Gang School of Management Hefei University of Technology Hefei230009 China Key Laboratory of Process Optimization and Intelligent Decision-making Ministry of Education Hefei230009 China Ministry of Education Engineering Research Centre for Intelligent Decision-Making & Information System Technologies Hefei230009 China

Human communication in the real world is based on multimodal data. Multimodal language understanding tasks such as multimodal intention recognition and multimodal emotion recognition are key steps in many applications, such as dialogue systems and social media. Although many studies have explored multimodal fusion methods and applied these methods to multimodal language understanding tasks including intent recognition and sentiment analysis, it is still a knotty problem whether the classification results are trusted. Therefore, this paper proposes a hierarchically-trusted evidential fusion method (HTEF) for multimodal language understanding. We construct multiple views from multi-modal data in a hierarchical manner, and extract multiple pieces of evidence from them. Then, we introduce Dempster-Shafer evidence theory(DST) to fuse evidence obtained from multimodal data to ensure the reduction of overall uncertainty. In addition, we explore the feature-level consistency and decision-level consistency of multi-modal data to avoid the conflict across information captured from different modalities. The method is evaluated on the public multimodal intent recognition dataset MIntRec and the emotion recognition dataset MELD. The experimental results show that our model outperforms both existing single-modal and multimodal baseline models across multiple evaluation metrics, which verifies the effectiveness of the proposed method. © 2024, The Authors. All rights reserved.

关键词： Emotion Recognition

Predicting Helpfulness of Multimodal Reviews with Customer Confirmation Bias: A Hierarchically-Trusted Multi-View Deep Learning Method

学校读者我要写书评

暂无评论

SSRN 2024年

作者： Yang, Ying Tang, Si Ren, Gang School of Management Hefei University of Technology Hefei230009 China Key Laboratory of Process Optimization and Intelligent Decision-making Ministry of Education Hefei230009 China Ministry of Education Engineering Research Center for Intelligent Decision-Making & Information System Technologies Hefei230009 China

Accurately identifying high quality online-reviews is beneficial to both consumers and e-commerce platforms. Consumers usually exhibit confirmation bias phenomenon when evaluating the helpfulness of product reviews. However, there has been limited research on simulating and incorporating this information processing bias into predicting multimodal review helpfulness. To address this challenge, we combine confirmation bias theory with deep learning techniques, and propose a hierarchically-trusted multi-view deep learning method for predicting the helpfulness of multimodal reviews. Drawing upon the elaborate likelihood model and confirmation bias theory, multiple views are designed to represent consumers’ initial beliefs, central cues and peripheral cues of multimodal reviews. A hierarchically-trusted multi-view fusion strategy is designed to dynamically aggregate the different views considering their interactions. Experimental results on the *** review dataset demonstrate that the proposed method effectively improves the accuracy of evaluating the values of online reviews. This research contributes to the improvement of platform service operations and enhances consumer experience. © 2024, The Authors. All rights reserved.

关键词： Deep learning

A deep learning model integrating domain-specific features for enhanced glaucoma diagnosis

学校读者我要写书评

暂无评论

BMC medical informatics and decision making 2025年第1期25卷 195页

作者： Jie Xu Erkang Jing Yidong Chai Beijing Institute of Ophthalmology Beijing Tongren Eye Center Beijing Tongren Hospital Capital Medical University Beijing China. School of Management of Hefei Key Laboratory of Process Optimization and Intelligence Decision Making University of Technology Minister of Education Hefei China. School of Management of Hefei Key Laboratory of Process Optimization and Intelligence Decision Making University of Technology Minister of Education Hefei China. chaiyd@***.

Glaucoma is a group of serious eye diseases that can cause incurable blindness. Despite the critical need for early detection, over 60% of cases remain undiagnosed, especially in less developed regions. Glaucoma diagnosis is a costly task and some models have been proposed to automate diagnosis based on images of the retina, specifically the area known as the optic cup and the associated disc where retinal blood vessels and nerves enter and leave the eye. However, diagnosis is complicated because both normal and glaucoma-affected eyes can vary greatly in appearance. Some normal cases, like glaucoma, exhibit a larger cup-to-disc ratio, one of the main diagnostic criteria, making it challenging to distinguish between them. We propose a deep learning model with domain features (DLMDF) to combine unstructured and structured features to distinguish between glaucoma and physiologic large cups. The structured features were based upon the known cup-to-disc ratios of the four quadrants of the optic discs in normal, physiologic large cups, and glaucomatous optic cups. We segmented each cup and disc using a fully convolutional neural network and then calculated the cup size, disc size, and cup-to-disc ratio of each quadrant. The unstructured features were learned from a deep convolutional neural network. The average precision (AP) for disc segmentation was 98.52%, and for cup segmentation it was also 98.57%. Thus, the relatively high AP values enabled us to calculate the 15 reliable features from each segmented disc and cup. In classification tasks, the DLMDF outperformed other models, achieving superior accuracy, precision, and recall. These results validate the effectiveness of combining deep learning-derived features with domain-specific structured features, underscoring the potential of this approach to advance glaucoma diagnosis.

关键词： Deep learning Disease diagnosis Glaucoma diagnosis Image segmentation Medical image analysis Physiologic large cups

Exploring the impact of COVID-19 on travel behavior based on EV charging big data 5

学校读者我要写书评

暂无评论

Exploring the impact of COVID-19 on travel behavior based on...

5th IEEE Conference on Energy Internet and Energy System Integration, EI2 2021

作者： Zhou, Kaile Hu, Dingding Li, Fangyi Key Laboratory of Process Optimization and Intelligent Decision-making of Ministry of Education School of Management Hefei University of Technology Hefei230009 China

ISBN: (纸本)9781665434256

The COVID-19 pandemic has forced many governments around the world to implement strict lockdown measures and order citizens to stay at home, which has caused a major change in travel patterns. This study leveraged electric vehicle charging big data in Hefei, Anhui Province, China to estimate electric vehicle charging demand in the absence of the COVID-19 pandemic using multi-layer perceptron model, which quantified the impact of the COVID-19 pandemic. In addition, we employed the vector autoregressive model to investigate the dynamic relationships between the changes in charging demand and various explanatory factors. The results suggest that the daily average charging demand in Hefei decreased by 78.3% compared to the predicted value during the pandemic. Furthermore, according to the variance decomposition and impulse response function analysis, national confirmed COVID-19 cases play a dominant role in reducing charging demand. The number of daily hospitalizations and Migration Scale Index also have significant and robust effect on the decrease in charging demand. The Air Quality Index and Baidu Index are susceptible to external factors and do not have a direct impact on the change in charging demand. Findings support a better understanding of changes in travel behavior during the pandemic and provide policy makers with references to deal with similar events. © 2021 IEEE

关键词： COVID-19

Generative Adversarial Network-Based Optimal Energy Storage Capacity Configuration for Newly–Build Microgrids

学校读者我要写书评

暂无评论

SSRN 2024年

作者： Zhou, Kunshu Zhou, Kaile Yang, Shanlin School of Management Hefei University of Technology Hefei230009 China Key Laboratory of Process Optimization and Intelligent Decision-making of Ministry of Education Hefei University of Technology Hefei230009 China

In the initial stage of microgrid construction, historical data is limited. It is challenging to address the issues of the small amount of data and high uncertainty of renewable energy when proceeding with capacity configuration. This study performs Wasserstein distance-based generative adversarial network with gradient penalty (WGAN-GP) to enhance the original 30 days PV power generation scenario data. Furthermore, the improved data is used to build a distributionally robust optimization model based on various discrete scenarios to determine the optimal capacity configuration of energy storage in newly–built microgrids. The results obtained from the solution are compared with those of traditional stochastic optimization and robust optimization, and the robustness of the results is verified by sensitivity analysis. The comparison results demonstrate that the strategy employed in this study can better address the issues of limited data and uncertainty in renewable energy in the capacity configuration process of newly-built microgrids. © 2024, The Authors. All rights reserved.

关键词： Sensitivity analysis

Competition in durable goods markets with product failures into consideration 9

学校读者我要写书评

暂无评论

Competition in durable goods markets with product failures i...

9th International Symposium on Project Management, ISPM 2021

作者： Tian, Jianfei Qian, Xiaofei School of management Hefei University of Technology Anhui China Key Laboratory of Process Optimization and Intelligent Decision-making of the Ministry of Education China

ISBN: (纸本)9781921712777

Few studies on competition in dual channel for durable goods take the product failures into consideration. However, the factors related to preventive maintenance (PM) and minimal maintenance (MM) have an important impact on the optimal profits of the manufacturer and retailer. Consequently, we develop three game models (Single Selling channel (SS), Manufacturer Stackelberg (MS) and Cooperation Model (CM)) to study the players’ equilibrium solutions and analyze the impact of the initial point, numbers and level of preventive maintenance on equilibrium price and profits. We find that the equilibrium wholesale and retail price under SS tend to increase when the manufacturer decides to raise the repair charge, while in MS the manufacturer and retailer will select a higher level for the wholesale price, retail price and rental charge when manufacturer chooses to put off the initial point for PM. In addition, numerical examples are provided to supplement our conclusions. The results show that the manufacturer in MS will set the numbers and level of PM to the maximum threshold in order to maximize its profit. © ISPM *** right reserved.

关键词： Preventive maintenance

A Two-Dimensional Time-Aware Cloud Service Recommendation Approach with Enhanced Similarity and Trust

学校读者我要写书评

暂无评论

SSRN 2023年

作者： Tang, Chunhua Zhao, Shuangyao Chen, Binbin Lu, Xiaonong Zhang, Qiang School of Management Hefei University of Technology Hefei230009 China Key Laboratory of Process Optimization and Intelligent Decision-making Ministry of Education Hefei230009 China

Owing to an increasing number of functionally equivalent services on the cloud computing platform, quality-of-service (QoS) prediction and service recommendation have been developed using collaborative filtering (CF) techniques. However, each individual QoS is related to service status and user perceptions and thus instantly varies over time. QoS instantaneity often makes service predicted QoS deviate from actual values in traditional CF approaches. In addition, existing CF approaches ignore inauthentic QoS values given by untrustworthy users. To address these problems, we develop a two-dimensional time-aware and trust-aware service recommendation approach (TaTruSR). First, considering QoS instantaneity, both time weight (time dimension) and temporal certainty (QoS dimension) are designed to determine the contribution of co-invoked services. Second, a set of most similar and trusted neighbors can be identified from the view of the time-aware similarity model and trust model. In models, the direct similarity and local trust are calculated based on the QoS ratings and contribution of co-invoked services to improve the prediction accuracy and eliminate unreliable QoS. The indirect similarity and global trust are estimated based on the social network to alleviate the data sparsity problem. Finally, missing QoS prediction and service recommendation for the active user can be achieved based on enhanced similarity and trust. A case study and experimental evaluation on real-world datasets demonstrate the practicality and accuracy of the proposed approach. © 2023, The Authors. All rights reserved.

关键词： Forecasting

Learning Adaptive Mapless Navigation Skills Through Evidential Deep Learning

学校读者我要写书评

暂无评论

Learning Adaptive Mapless Navigation Skills Through Evidenti...

IEEE International Conference on Real-time Computing and Robotics (RCAR)

作者： Cancan Zhao Yufeng Yan Xuhao Liu Shuai Zhang Bo Ouyang School of Management Hefei University of Technology Hefei China The Key Laboratory of Process Optimization and Intelligent Decision-Making(Ministry of Education) Hefei University of Technology Hefei China

Traditional navigation methods tend to design and train navigation skills based on the consideration of multi-scenario applicability and perform well in general scenarios. However, this results in sub-optimal navigation performance in some context-specific scenarios, such as navigating through open or crowded scenarios. While some existing navigation skills can achieve satisfactory performance in context-specific scenarios, e.g., using PID in open scenarios and reinforcement learning-based training methods in crowded scenarios. Most methods currently do without consideration of the integration of navigation skills and contextual understanding. To address this gap, we propose an adaptive mapless navigation method to utilize the optimal capabilities of each skill corresponding to the context-specific scenario. Specifically, we package several existing navigation skills and train a scenario classifier to enable the agent to automatically select the most appropriate skill based on the current scenario. Furthermore, we incorporate evidential deep learning (EDL) to assess classification uncertainty and ensure the safe switching of navigation skills. Our method was tested in both simulated and real scenarios, and the results of the experiments confirmed the effectiveness and generalization of our method. Moreover, the proposed method exhibits notably enhanced, context-sensitive efficiency within a hybrid multi-scenario setting, in comparison to a single navigation skill.

关键词：