检索结果-内蒙古大学图书馆

Self‐supervised monocular depth estimation via asymmetric convolution block

IET Cyber-Systems and Robotics 2022年第2期4卷 131-138页

作者： Lingling Hu Hao Zhang Zhuping Wang Chao Huang Changzhu Zhang Department of Control Science and Engineering Tongji UniversityShanghaiChina

Without the dependence of depth ground truth,self‐supervised learning is a promising alternative to train monocular depth *** builds its own supervision signal with the help of other tools,such as view synthesis and pose ***,more training parameters and time consumption may be *** paper proposes a monocular depth prediction framework that can jointly learn the depth value and pose transformation between images in an end‐to‐end *** depth network creatively employs an asymmetric convolution block instead of every square kernel layer to strengthen the learning ability of extracting image features when *** infer-ence time,the asymmetric kernels are fused and converted to the original network to predict more accurate image depth,thus bringing no extra computations *** network is trained and tested on the KITTI monocular *** evaluated results demonstrate that the depth model outperforms some State of the Arts(SOTA)ap-proaches and can reduce the inference time of depth ***,the pro-posed model performs great adaptability on the Make3D dataset.

关键词： asymmetric convolution block(ACB) KITTI dataset self‐supervised depth estimation

来源：评论

学校读者我要写书评

暂无评论

Parameter Identification Based on Generalized Orthonormal Basis Function Without Persistent Excitation: A Learning-Based Paradigm 22

Parameter Identification Based on Generalized Orthonormal Ba...

引用

22nd IEEE International Conference on Industrial Informatics, INDIN 2024

作者： Li, Kuan Su, Weijie Li, Xingyong He, Yang Huang, Minchang Luo, Hao Shanghai Aerospace Control Technology Institute Shanghai China Department of Control Science and Engineering Harbin Institute of Technology Harbin China

ISBN: (纸本)9798331527471

This paper presents the parameter identification based on generalized orthonormal basis function (GOBF) without persistent excitation via two learning-based paradigms. The problem is formulated as a two-stage identification of poles and weight coefficients of GOBFs. The GOBF is designed based on the two-parameter Kautz basis function for the compatiblity with complex poles. The two learning-based paradigms are established on the nature-inspired meta-heuristic optimization and deep reinforcement learning, respectively. The ability of identification without strict persistent excitation is discussed as well. The effectiveness of the two paradigms are verified through a simulation study on a 4-order transfer function model. © 2024 IEEE.

关键词： Deep reinforcement learning

来源：评论

学校读者我要写书评

暂无评论

Adaptive Particle Swarm Optimization-Based Tuning of Fractional Order PID for Moisture control in an Industrial Rotary Dryer System 11th

Adaptive Particle Swarm Optimization-Based Tuning of Fractio...

引用

11th EAI International Conference on Advancement of science and Technology, ICAST 2023

作者： Bosera, Adisu Safo Sori, Raji Ababe Bayesa, Ashebir Berhanu Department of Electrical Power and Control Engineering Adama Science and Technology University Adama Ethiopia

ISBN: (纸本)9783031641503

In this paper, a dynamic model of an industrial rotary dryer was formulated, and the model of the system was linearized using the Jacobian method. Adaptive particle swarm and gray wolf optimization-based tuning of parameters of fractional order proportional, integral, and derivative (FOPID) and conventional Ziegler-Nichols proportional, integral, and derivative (ZN-PID) was assessed with simulation for a solid moisture control rotary dryer system. The main feature of fractional order PID is that it may have a greater degree of freedom compared to conventional PID. Here, particle swarm and gray wolf optimization algorithms will be used for tuning the parameters of the FOPID controller against the kinds of disturbances that exist in this plant. The performance of the controller is compared to that of typical PID and gray wolf optimizer (GWO) tuning of fractional order PID controllers (FOPID). The performance of the developed controller is evaluated by assessing various time domain metrics. Using the application of the PSO-FOPID control, a satisfactory result was obtained with an integral time square error (ITSE) of 0.0032 compared to 0.0039 and 0.0366 in the cases of GWO-FOPID and ZN-FOPID, returning the expected and desired output moisture. In order to determine the robustness of the proposed control system, random linear density of solids and drying air, inlet temperature of solids, inlet moisture of solids, and linear velocity of solids are subjected to the controller. Therefore, from the simulation result, the performance of the PSO-FOPID controller will show good time domain specifications, such as an overshoot of 16.14%, a rise time of 0.1378 s and a settling time of 5.39 s. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Moisture control

来源：评论

学校读者我要写书评

暂无评论

Affinity Based Semantic Collaborative Hashing for Image Retrieval 11th

Affinity Based Semantic Collaborative Hashing for Image Ret...

引用

11th International Joint Conference on Rough Sets, IJCRS 2025

作者： Li, Yi Fan, Wentao Meng, Ziqi Li, Huaxiong Department of Control Science and Intelligence Engineering Nanjing University Jiangsu Nanjing210093 China

ISBN: (纸本)9783031927409

Hashing methods have captured increasing attention in image retrieval due to their computational efficiency and low storage cost. Most existing supervised hashing approaches focus on preserving the semantic similarity for discriminative hash codes learning. However, they may not discover intrinsic latent features embedded in high-dimensional feature space, and ignore the underlying affinity structures of data. In this paper, we proposed a novel Affinity based Semantic Collaborative Hashing (ASCH) method for image retrieval, which obtains the hash codes by the collaboration of latent discriminative features and semantic labels. Specifically, ASCH explores the data correlations by learning a self-representation based affinity matrix. To facilitate the semantic latent features learning, an asymmetric strategy is designed by factorizing an affinity matrix into the inner product of low-dimensional label matrix and latent feature matrix. Thus, the learned latent features not only explore the underlying low-rank data structure, but also encode the semantic class information. The hash codes are jointly learned in a unified framework by collaboratively projecting the latent features and labels. Based on the hash codes, a flexible regression model with adaptive marginalization is used to learn the hash functions for out-of-sample extension. Our proposed model can be efficiently solved with linear time complexity. We compare the ASCH with state-of-the-art methods across five benchmark datasets. The experimental results validate the feasibility and superiority of our proposed method. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

关键词： Supervised learning

来源：评论

学校读者我要写书评

暂无评论

High-Precision control of Humanoid Muscle-skeleton Robotic Arm Using Reinforcement Learning and Large Language Models 50

High-Precision Control of Humanoid Muscle-skeleton Robotic A...

引用

50th Annual Conference of the IEEE Industrial Electronics Society, IECON 2024

作者： Wang, Yan Wang, Qiang Fan, Jianyin Department of Control Science and Engineering Harbin Institute of Technology Harbin150001 China

ISBN: (纸本)9781665464543

In recent years, humanoid robotics have achieved significant advancements, including those embodied by muscle-skeleton robots. Due to the nonlinear characteristics of these robots, reinforcement learning is a popular method for muscle-skeleton robot control. However, the design of the reward function and the adjustment of hyper parameters is a difficult task when using reinforcement learning. In this paper, we combine reinforcement learning with a large language model to achieve precise control of a humanoid muscle-skeleton robotic arm. We first tell the large language model task, states, and hyper parameters, and then continuously feedback the training results. We can finally get a controller that can control the robot to complete the target-tracking task. The experimental results show that the large language model can propose a suitable reward function based on experience, and adjust the hyper parameters to better training results within several questions and answers. © 2024 IEEE.

关键词： Musculoskeletal system

来源：评论

学校读者我要写书评

暂无评论

Design and Implementation of Steering Torque Feedback System for Steer-by-Wire Vehicles 43

Design and Implementation of Steering Torque Feedback System...

引用

43rd Chinese control Conference, CCC 2024

作者： Ma, Chao Zhao, Linhui Harbin Institute of Technology Department of Control Science and Engineering Harbin150001 China

ISBN: (纸本)9789887581581

Steer-by-wire (SBW) is an advanced technology of great significance for the intelligent driving. To implement the function of the road feeling feedback for SBW, this paper presents a design of the road feeling feedback system, including the mechanical structure and the control algorithm. The target feedback torque is planned by the method of function fitting, and the controller is designed to track the target torque by combining the feedback control and the feedforward control. This paper designs an experimental bench and validates the performance of the system by hardware in-loop (HIL) experiment. The results show that the system can provide an approximate feedback torque and a comfortable driving experience. © 2024 Technical Committee on control Theory, Chinese Association of Automation.

关键词： Feedback control

来源：评论

学校读者我要写书评

暂无评论

A Simplified Method for Convex Hull Pricing 4

A Simplified Method for Convex Hull Pricing

引用

4th International Conference on Intelligent Power and Systems, ICIPS 2024

作者： Wu, Tongqi Zou, Bin Shanghai University Dept. Department of Control Science and Engineering Shanghai China

ISBN: (纸本)9798331532765

In recent years, the development of new energy sources has led to a reduction in the proportion of thermal power generation, which in turn has highlighted the issue of non-convexities in electricity markets. Convex hull pricing is a well-documented method for coping with the non-existence of uniform clearing prices in electricity markets with non-convex costs and constraints. This paper revisits the primal and dual approaches for calculating convex hull prices, discusses the Dantzig-Wolfe decomposition method, and highlights its limitations. Building on this, we propose a simplified approach to convex hull pricing. Additionally, we provide empirical evidence about the favorable performance of our algorithm on large test instances based on RTS system. © 2024 IEEE.

关键词： convex hull pricing level method simplified approach uniform clearing prices

来源：评论

学校读者我要写书评

暂无评论

A Reinforcement Learning Solution to the Nonlinear Spacecraft Pursuit-Evasion Game Problem 14

A Reinforcement Learning Solution to the Nonlinear Spacecraf...

引用

14th IEEE International Conference on Cyber Technology in Automation, control, and Intelligent Systems, CYBER 2024

作者： Huang, Haoqi Ran, Guangtao Lyu, Yueyong Ma, Guangfu Harbin Institute of Technology Department of Control Science and Engineering Harbin150001 China

ISBN: (纸本)9798331506056

The pursuit-evasion game of non-cooperative spacecrafts under nonlinear dynamics is currently a hot topic in orbital gaming. We describe the above pursuit-evasion game model using differential game theory, transforming the gaming problem into a bilateral optimal control problem. Using elliptical orbit line-of-sight (LOS) dynamics with simple field-of-view constrains as the system model, we solve the Nash equilibrium solution for the two-body pursuit-evasion under the assumption of complete information. Due to the difficulty in obtaining analytical solutions for the Nash equilibrium, we adopt a reinforcement learning (RL)-based adaptive dynamic programming method. We obtain the approximate Nash equilibrium solution with RL method eventually and provide a successful simulation example. © 2024 IEEE.

关键词： Orbits

来源：评论

学校读者我要写书评

暂无评论

Reinforcement Learning with Neighborhood Search and Self-learned Rules 13

Reinforcement Learning with Neighborhood Search and Self-lea...

引用

13th IEEE Data Driven control and Learning Systems Conference, DDCLS 2024

作者： Yin, Xiao Zhu, Yuanyang Chen, Chunlin Nanjing University Department of Control Science and Intelligent Engineering Nanjing210093 China

ISBN: (纸本)9798350361674

Reinforcement learning (RL), induced by the exploration-exploitation dilemma, usually requires enough interaction experiences with the environment to improve performance. Rule-based methods utilize the internal expertise of tasks to benefit the learning process of the target task and can be shared with similar goal-setting tasks, which is also one form of knowledge transfer with better interpretability. However, symbolic rules or prior knowledge of the task may not be obtained in advance for most control tasks. In this paper, we propose an effective approach named neighborhood search and self-learned rules (RL-NSSR), where the neighborhood search library and the rules can be learned from the samples of a set of collected trajectories. In the RL- NSSR, first, the expert policy is obtained by training the RL agent in the discretized continuous state-action space task for collecting the trajectories data. Second, we extract the rules using the collected data via ensemble decision tree methods, and select the states of the trajectories with higher cumulative rewards as the neighborhood states to speed up the learning process. Third, we employ both the neighborhood search and the self-learned symbolic rules in a fine-grained way and a rough way, respectively, to guide the exploration strategy for accelerating the learning process. Experiments conducted on several groups of locomotion tasks by the MuJoCo physics engine demonstrate that the proposed RL-NSSR approach can achieve improved control performances for both simple and complex tasks. © 2024 IEEE.

关键词： Self-supervised learning

来源：评论

学校读者我要写书评

暂无评论

Cotransport of nanoplastics with nZnO in saturated porous media:From brackish water to seawater

引用

Journal of Environmental sciences 2025年第2期148卷 541-552页

作者： Yuanzhang Hou Yange Wang Yidi Zhang Zhibo Lu Zhenbo Zhang Zhiqiang Dong Yuping Qiu Department of Environmental Science College of Environmental Science and EngineeringTongji UniversityState Key Laboratory of Pollution Control and Resources ReuseShanghai Institute of Pollution Control and Ecological SecurityShanghai 200092China School of Medicine Tongji UniversityShanghai 200065China Municipal Environmental Protection Engineering Co. Ltd of CERC Shanghai GroupShanghai 201906China China Railway Engineering Group Co. Beijing 100039China

The ocean serves as a repository for various types of artificial ***(NPs)and nano zinc oxide(nZnO),which are frequently employed in personal care products and food packaging materials,are likely simultaneously released and eventually into the ocean with surface ***,their mutual influence and shared destiny in marine environment cannot be *** study examined how nanomaterials interacted and transported through sea sand in various salinity *** showed that NPs remained dispersed in brine,while nZnO formed *** seawater of 35 practical salinity units(PSU),nZnO formed heteroaggregates with NPs,inhibiting NPs mobility and decreasing the recovered mass percentage(Meff)from 24.52%to 12.65%.In 3.5 PSU brackish water,nZnO did not significantly aggregate with NPs,and thus barely affected their ***,NPs greatly enhanced nZnO transport with Meff increasing from 14.20%to 25.08%,attributed to the carrier effect of higher mobility *** from brackishwater to seawater was simulated in salinity change experiments and revealed a critical salinity threshold of 10.4 PSU,below which the mobility of NPs was not affected by coexisting nZnO and above which nZnO strongly inhibited NP *** study highlights the importance of considering the mutual influence and shared destiny of artificial nanoparticles in the marine environment and how their interaction and cotransport are dependent on changes in seawater salinity.

关键词： Homoaggregate Heteroaggregate Transport Sea sand Salinity threshold

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：