Pub Date : 2022-08-01DOI: 10.1109/CCPQT56151.2022.00009
Ting-dong Hu
Wireless Mesh Network (WMN) requires end-to-end performance of data transmission while reducing routing overhead. Therefore, an improved DSR (Dynamic Source Routing) protocol based on the MarKov model is designed. First, WCETT (Weighted Cumulative Expected Transmission Time) is applied to the DSR protocol as a routing parameter to select the link quality optimal path with the smallest bottleneck channel for data transmission, and locally optimize the traditional DSR protocol into an improved DSR (IDSR) protocol. Then, the MarKov model is used in the IDSR protocol to predict the geographic location of the node at the next moment, and determine if the original route is invalid base on this location, reroute to the next best route of WCETT if the original route is no longer valid and so on, until the optimal route available is selected. The performance is evaluated using NS2 simulation software, and the simulation results show that under the same wireless transmission and network size conditions, the IDSR protocol based on the MarKov model results in higher packet delivery rate, significantly reduced end-to-end average delay, route initiation frequency and routing overhead, when the node movement speed is accelerated, the IDSR protocol improves the network performance more significantly as the nodes move faster.
{"title":"Research on an Improved DSR Protocol Based on the MarKov Model","authors":"Ting-dong Hu","doi":"10.1109/CCPQT56151.2022.00009","DOIUrl":"https://doi.org/10.1109/CCPQT56151.2022.00009","url":null,"abstract":"Wireless Mesh Network (WMN) requires end-to-end performance of data transmission while reducing routing overhead. Therefore, an improved DSR (Dynamic Source Routing) protocol based on the MarKov model is designed. First, WCETT (Weighted Cumulative Expected Transmission Time) is applied to the DSR protocol as a routing parameter to select the link quality optimal path with the smallest bottleneck channel for data transmission, and locally optimize the traditional DSR protocol into an improved DSR (IDSR) protocol. Then, the MarKov model is used in the IDSR protocol to predict the geographic location of the node at the next moment, and determine if the original route is invalid base on this location, reroute to the next best route of WCETT if the original route is no longer valid and so on, until the optimal route available is selected. The performance is evaluated using NS2 simulation software, and the simulation results show that under the same wireless transmission and network size conditions, the IDSR protocol based on the MarKov model results in higher packet delivery rate, significantly reduced end-to-end average delay, route initiation frequency and routing overhead, when the node movement speed is accelerated, the IDSR protocol improves the network performance more significantly as the nodes move faster.","PeriodicalId":235893,"journal":{"name":"2022 International Conference on Computing, Communication, Perception and Quantum Technology (CCPQT)","volume":"39 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117121482","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-01DOI: 10.1109/CCPQT56151.2022.00058
Can Zeng, Qiao Kang, Pcngfei Hu, Mintao Dong, F. Dong, Kewei Chen
Aiming at the problems of low manual detection efficiency and low automatic detection accuracy of cashmere and wool, a method is proposed to realize the image detection of cashmere and wool single fiber by using large kernel attention(LKA) mechanism and deep convolutional neural network. Based on the ConvNeXt network structure paradigm with simple structure, good scalability and high accuracy on ImageNet large datasets, the inverted residual structure improvement avoids information loss, and large kernel attention mechanism is added to make the model more accurate to pay attention to differentiated regions on the feature map space and channel, while considering the sparse amount of fiber image information, in order to avoid redundant training parameters and overfitting, the network is lightweight while maintaining the proportion of the ConvNeXt network hierarchy, and the LKA-RConvNeXt model is established. Finally, after training on 15,000 cashmere and wool datasets, the highest classification accuracy can reach 96.1%. Through ablation experiments and model comparison analysis, it is verified that the improved method used is beneficial to the accuracy of the model. The model can be used for cashmere and wool in automatic classification tasks, and contributes to backbone network for the subsequent fiber object detection task.
{"title":"Cashmere and Wool Classification with Large Kernel Attention and Deep Learning","authors":"Can Zeng, Qiao Kang, Pcngfei Hu, Mintao Dong, F. Dong, Kewei Chen","doi":"10.1109/CCPQT56151.2022.00058","DOIUrl":"https://doi.org/10.1109/CCPQT56151.2022.00058","url":null,"abstract":"Aiming at the problems of low manual detection efficiency and low automatic detection accuracy of cashmere and wool, a method is proposed to realize the image detection of cashmere and wool single fiber by using large kernel attention(LKA) mechanism and deep convolutional neural network. Based on the ConvNeXt network structure paradigm with simple structure, good scalability and high accuracy on ImageNet large datasets, the inverted residual structure improvement avoids information loss, and large kernel attention mechanism is added to make the model more accurate to pay attention to differentiated regions on the feature map space and channel, while considering the sparse amount of fiber image information, in order to avoid redundant training parameters and overfitting, the network is lightweight while maintaining the proportion of the ConvNeXt network hierarchy, and the LKA-RConvNeXt model is established. Finally, after training on 15,000 cashmere and wool datasets, the highest classification accuracy can reach 96.1%. Through ablation experiments and model comparison analysis, it is verified that the improved method used is beneficial to the accuracy of the model. The model can be used for cashmere and wool in automatic classification tasks, and contributes to backbone network for the subsequent fiber object detection task.","PeriodicalId":235893,"journal":{"name":"2022 International Conference on Computing, Communication, Perception and Quantum Technology (CCPQT)","volume":"134 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116363005","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-01DOI: 10.1109/CCPQT56151.2022.00072
Lu Cai, Jun Liu, Fu-cheng Xiang, Shan Li
A highly sensitive liquid refractive index (RI) sensor with a single-mode fiber-multimode fiber-tapered thin-core fiber-single-mode fiber (SMTTS) structure can be simulated and the method is the beam propagation method (BPM). The modal perturbation between core and cladding mode happened because of modal mismatch and the transmission spectrum is calculated by BPM. Liquid RI and spectral dip wavelength shift for demodulating RI values is obtained and their relationship is approximately linear. To make a comparison, transmissions of single mode fiber-thin-core fiber-single mode fiber (STS) as well as single mode fiber-multimode fiber-thin-core fiber-single mode fiber (SMTS) are also calculated. It is significantly shown that the interference is improved by the segment of multimode fiber (MMF). And the tapered thin-core fiber (TCF) has the functions not only further enhance the fringe visibility, but also increase the RI sensitivity, which is higher with the thinner thin-core taper diameter. The average RI sensitivity would reach up to −116.7nm/RIU within the RI range of 1.33 to 1.39 if the taper diameter is $53.5 mumathrm{m}$. This is much higher than the value of SMTS structure. And in the practice application, it could be a slimmer taper in the SMTTS structure to improve the sensitivity just achieving by a simple fabrication process. All process could be finished using fusion splicer, cutter and hydrogen flame.
对单模光纤-多模光纤-锥形薄芯光纤-单模光纤(SMTTS)结构的高灵敏度液体折射率(RI)传感器进行了仿真，其方法为光束传播法(BPM)。模态失配导致芯模与包层模之间产生模态扰动，利用BPM计算透射谱。得到了解调RI值的液体RI和光谱倾角波长移，两者近似线性关系。为了进行比较，还计算了单模光纤-薄芯光纤-单模光纤(STS)以及单模光纤-多模光纤-薄芯光纤-单模光纤(SMTS)的传输量。结果表明，多模光纤的分段可以明显地改善干扰。锥形薄芯光纤(TCF)不仅具有进一步提高条纹可见度的功能，而且还具有提高RI灵敏度的功能，并且随着薄芯锥度直径的增加，RI灵敏度更高。在1.33 ~ 1.39的RI范围内，当锥径为53.5 mu mathm {m}$时，平均RI灵敏度可达- 116.7nm/RIU。这远远高于SMTS结构的值。在实际应用中，可以在SMTTS结构中采用更细的锥度，通过简单的制作工艺来提高灵敏度。所有工序均可通过熔接机、切割机和氢火焰完成。
{"title":"Theoretical Analysis of an In-line Thin-core Fiber Based Refractometer","authors":"Lu Cai, Jun Liu, Fu-cheng Xiang, Shan Li","doi":"10.1109/CCPQT56151.2022.00072","DOIUrl":"https://doi.org/10.1109/CCPQT56151.2022.00072","url":null,"abstract":"A highly sensitive liquid refractive index (RI) sensor with a single-mode fiber-multimode fiber-tapered thin-core fiber-single-mode fiber (SMTTS) structure can be simulated and the method is the beam propagation method (BPM). The modal perturbation between core and cladding mode happened because of modal mismatch and the transmission spectrum is calculated by BPM. Liquid RI and spectral dip wavelength shift for demodulating RI values is obtained and their relationship is approximately linear. To make a comparison, transmissions of single mode fiber-thin-core fiber-single mode fiber (STS) as well as single mode fiber-multimode fiber-thin-core fiber-single mode fiber (SMTS) are also calculated. It is significantly shown that the interference is improved by the segment of multimode fiber (MMF). And the tapered thin-core fiber (TCF) has the functions not only further enhance the fringe visibility, but also increase the RI sensitivity, which is higher with the thinner thin-core taper diameter. The average RI sensitivity would reach up to −116.7nm/RIU within the RI range of 1.33 to 1.39 if the taper diameter is $53.5 mumathrm{m}$. This is much higher than the value of SMTS structure. And in the practice application, it could be a slimmer taper in the SMTTS structure to improve the sensitivity just achieving by a simple fabrication process. All process could be finished using fusion splicer, cutter and hydrogen flame.","PeriodicalId":235893,"journal":{"name":"2022 International Conference on Computing, Communication, Perception and Quantum Technology (CCPQT)","volume":"61 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126388558","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-01DOI: 10.1109/CCPQT56151.2022.00012
Xiaobo Yang, Ruonan Zhang, Hongmei Xie, Huakui Sun, Huanling Li
Intelligent transportation systems (ITS) are designed to provide efficient and comfortable transportation. The development of ITS has brought new communication challenges, which require faster and more reliable transmission of information. In this paper, we investigate the modulation mode recognition method of communication signals based on a complex-valued neural network (CVNN). By combining a complex-valued convolutional neural network (CVCNN) with complex-valued long short-term memory (CVLSTM) and adding a residual learning unit, a modulation recognition model is established. The model can automatically learn from complex-valued signals without manual feature extraction and can recognize 11 modulation modes (3 analog modulation modes and 8 digital modulation modes) with a signal-to-noise ratio (SNR) between −20 dB and 18 dB. We design a Gaussian filter, and divide the signal to be identified into high SNR signal and low SNR signal through SNR estimation. The low SNR signal is Gaussian filtered before modulation recognition, so as to improve its modulation recognition accuracy. The algorithm proposed in this paper directly recognizes the modulation mode of the complex-valued signal without any preprocessing, and the recognition accuracy is better than the existing algorithms. This work is of great significance to the improvement of information transmission speed and the construction of ITS.
{"title":"Automatic Modulation Mode Recognition of Communication Signals Based on Complex-Valued Neural Network","authors":"Xiaobo Yang, Ruonan Zhang, Hongmei Xie, Huakui Sun, Huanling Li","doi":"10.1109/CCPQT56151.2022.00012","DOIUrl":"https://doi.org/10.1109/CCPQT56151.2022.00012","url":null,"abstract":"Intelligent transportation systems (ITS) are designed to provide efficient and comfortable transportation. The development of ITS has brought new communication challenges, which require faster and more reliable transmission of information. In this paper, we investigate the modulation mode recognition method of communication signals based on a complex-valued neural network (CVNN). By combining a complex-valued convolutional neural network (CVCNN) with complex-valued long short-term memory (CVLSTM) and adding a residual learning unit, a modulation recognition model is established. The model can automatically learn from complex-valued signals without manual feature extraction and can recognize 11 modulation modes (3 analog modulation modes and 8 digital modulation modes) with a signal-to-noise ratio (SNR) between −20 dB and 18 dB. We design a Gaussian filter, and divide the signal to be identified into high SNR signal and low SNR signal through SNR estimation. The low SNR signal is Gaussian filtered before modulation recognition, so as to improve its modulation recognition accuracy. The algorithm proposed in this paper directly recognizes the modulation mode of the complex-valued signal without any preprocessing, and the recognition accuracy is better than the existing algorithms. This work is of great significance to the improvement of information transmission speed and the construction of ITS.","PeriodicalId":235893,"journal":{"name":"2022 International Conference on Computing, Communication, Perception and Quantum Technology (CCPQT)","volume":"7 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126432671","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-01DOI: 10.1109/CCPQT56151.2022.00008
Baihua Yang, Yue Zhang
Cybersecurity threat identification and analysis of wind farm industrial control systems (ICS-WF) is an important part of security work. The article innovatively proposes the STRIDE threat model as the basic structure to establish a framework of information system cybersecurity hierarchical threat analysis model for identifying and analyzing cyber security threats to ICS-WF. The framework model system covers three levels and six dimensions. The three levels refer to the Data Stream Layer, Context Layer, and Component Layer, which correspond to three parts of System Devices, Trust Boundary, and Application Process respectively; the six dimensions refer to six dimensions of Spoofing, Tampering, Repudiation, Information Disclosure, Denial of Service and Elevation of Privilege. Through this model framework for system analysis in a typical wind farm industrial control network environment, not only can we accurately identify and analyze cybersecurity threats, but also provide an effective basis for threat abatement and system design through the visualization of threats.
{"title":"Cybersecurity Analysis of Wind Farm Industrial Control System Based on Hierarchical Threat Analysis Model Framework","authors":"Baihua Yang, Yue Zhang","doi":"10.1109/CCPQT56151.2022.00008","DOIUrl":"https://doi.org/10.1109/CCPQT56151.2022.00008","url":null,"abstract":"Cybersecurity threat identification and analysis of wind farm industrial control systems (ICS-WF) is an important part of security work. The article innovatively proposes the STRIDE threat model as the basic structure to establish a framework of information system cybersecurity hierarchical threat analysis model for identifying and analyzing cyber security threats to ICS-WF. The framework model system covers three levels and six dimensions. The three levels refer to the Data Stream Layer, Context Layer, and Component Layer, which correspond to three parts of System Devices, Trust Boundary, and Application Process respectively; the six dimensions refer to six dimensions of Spoofing, Tampering, Repudiation, Information Disclosure, Denial of Service and Elevation of Privilege. Through this model framework for system analysis in a typical wind farm industrial control network environment, not only can we accurately identify and analyze cybersecurity threats, but also provide an effective basis for threat abatement and system design through the visualization of threats.","PeriodicalId":235893,"journal":{"name":"2022 International Conference on Computing, Communication, Perception and Quantum Technology (CCPQT)","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125626799","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-01DOI: 10.1109/CCPQT56151.2022.00043
Xianyu Zhang, Tao Liang, K. An, Xiaoqiang Qiao, Xiaoyu Wang, Xiaoli Sun
This paper studied the effect of phase noise on secrecy performance of the cell-free massive MIMO network with multi-antenna user terminals, where a multi-antenna eavesdropper will actively contaminate the uplink training. Specifically, using discrete-time Wiener model and random matrix theory, tractable expression for the achievable secrecy rate is derived, which can provide precise evaluation of key factors including phase noise, multi-antenna and arbitrary system configuration. Analytical results illustrate that strong phase noise and high wiretapping power would lead to performance degradation while the network can acquire multiplexing gain by equipping multiple antennas. Also, it indicates that phase noise at the user terminals imposes more rigorous influence than that on access points (APs). Simulations are presented to corroborate the derived results.
{"title":"Impact of Phase Noise on Secrecy Performance of Cell-Free Massive MIMO Networks with Multi-Antenna Users","authors":"Xianyu Zhang, Tao Liang, K. An, Xiaoqiang Qiao, Xiaoyu Wang, Xiaoli Sun","doi":"10.1109/CCPQT56151.2022.00043","DOIUrl":"https://doi.org/10.1109/CCPQT56151.2022.00043","url":null,"abstract":"This paper studied the effect of phase noise on secrecy performance of the cell-free massive MIMO network with multi-antenna user terminals, where a multi-antenna eavesdropper will actively contaminate the uplink training. Specifically, using discrete-time Wiener model and random matrix theory, tractable expression for the achievable secrecy rate is derived, which can provide precise evaluation of key factors including phase noise, multi-antenna and arbitrary system configuration. Analytical results illustrate that strong phase noise and high wiretapping power would lead to performance degradation while the network can acquire multiplexing gain by equipping multiple antennas. Also, it indicates that phase noise at the user terminals imposes more rigorous influence than that on access points (APs). Simulations are presented to corroborate the derived results.","PeriodicalId":235893,"journal":{"name":"2022 International Conference on Computing, Communication, Perception and Quantum Technology (CCPQT)","volume":"3 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132015345","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-01DOI: 10.1109/CCPQT56151.2022.00027
Haoshuang Zhao, Wen-hui Zhang, Xiuqiang Wu, Hongning Li, Liu He
With the expansion of the market of new energy vehicles and the requirement of low carbon emission, the resource reuse of the Internet of Vehicles (IoV) has attracted much attention. Cognitive radio technology is used to solve the problem of spectrum shortage of IoV. As a classic attack, spectrum sensing data falsification attack can mislead others to make wrong decisions. Traditional solutions mainly consider fixed network, which is difficult to apply to the fast moving IoV. This paper proposes a distributed collaboration method based on outlier detection, which uses outlier detection, trust management and block chain to defense spectrum sensing data falsification attack. Simulation results show that compared with traditional distributed cooperative spectrum sensing schemes, this method has certain performance advantages and can effectively defend against spectrum sensing data falsification attack launched by malicious users in IoV
{"title":"Outlier Detection and Trust based Distributed Cooperative Spectrum Sensing in Internet of Vehicles","authors":"Haoshuang Zhao, Wen-hui Zhang, Xiuqiang Wu, Hongning Li, Liu He","doi":"10.1109/CCPQT56151.2022.00027","DOIUrl":"https://doi.org/10.1109/CCPQT56151.2022.00027","url":null,"abstract":"With the expansion of the market of new energy vehicles and the requirement of low carbon emission, the resource reuse of the Internet of Vehicles (IoV) has attracted much attention. Cognitive radio technology is used to solve the problem of spectrum shortage of IoV. As a classic attack, spectrum sensing data falsification attack can mislead others to make wrong decisions. Traditional solutions mainly consider fixed network, which is difficult to apply to the fast moving IoV. This paper proposes a distributed collaboration method based on outlier detection, which uses outlier detection, trust management and block chain to defense spectrum sensing data falsification attack. Simulation results show that compared with traditional distributed cooperative spectrum sensing schemes, this method has certain performance advantages and can effectively defend against spectrum sensing data falsification attack launched by malicious users in IoV","PeriodicalId":235893,"journal":{"name":"2022 International Conference on Computing, Communication, Perception and Quantum Technology (CCPQT)","volume":"61 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130805632","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-01DOI: 10.1109/CCPQT56151.2022.00030
L. Wang, Bo Jiang, Zhe Sun
In the path search problem under the roadless environmen such as the wild or remote areas, the existing path planning algorithms still have shortcomings such as slow speed and long planning path; The A* algorithm is efficient and optimal when dealing with node search problems, so it is widely used, but if you want to use the A* algorithm in a roadless environment, you need to divide the environment into a series of square grids or triangular grids. Compared with square grids, triangular grids can better adapt to various terrains and make full use of obstacle data; therefore, this paper proposes an improved A* algorithm based on triangular grids (referred to as TIA* algorithm), hope it can solve the dilemma of the current roadless algorithm. The experimental results show that the TIA* algorithm can realize the path planning under the roadless environmen, and compared with the traditional square grid-based A* algorithm and the RRT algorithm, it can shorten the planning time and the length of path under the premise of ensuring a high success rate.
{"title":"An Improved A* Algorithm for Roadless Networks Based on Triangular Mesh","authors":"L. Wang, Bo Jiang, Zhe Sun","doi":"10.1109/CCPQT56151.2022.00030","DOIUrl":"https://doi.org/10.1109/CCPQT56151.2022.00030","url":null,"abstract":"In the path search problem under the roadless environmen such as the wild or remote areas, the existing path planning algorithms still have shortcomings such as slow speed and long planning path; The A* algorithm is efficient and optimal when dealing with node search problems, so it is widely used, but if you want to use the A* algorithm in a roadless environment, you need to divide the environment into a series of square grids or triangular grids. Compared with square grids, triangular grids can better adapt to various terrains and make full use of obstacle data; therefore, this paper proposes an improved A* algorithm based on triangular grids (referred to as TIA* algorithm), hope it can solve the dilemma of the current roadless algorithm. The experimental results show that the TIA* algorithm can realize the path planning under the roadless environmen, and compared with the traditional square grid-based A* algorithm and the RRT algorithm, it can shorten the planning time and the length of path under the premise of ensuring a high success rate.","PeriodicalId":235893,"journal":{"name":"2022 International Conference on Computing, Communication, Perception and Quantum Technology (CCPQT)","volume":"551 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133171660","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-01DOI: 10.1109/CCPQT56151.2022.00018
Huilin Chen, Shengsong Yang, Ting Lyu
A multi-task semantic segmentation network architecture based on adaptive multi-scale feature fusion is proposed, which improves segmentation target edge details and small-scale target segmentation accuracy by combining boundary detection tasks and semantic segmentation tasks. The critical component of the architecture is the adaptive multi-scale feature fusion module, which can adaptively fuse the semantic feature information and boundary feature information of different scales, extract semantic features that contain more spatial data, and reduce the loss of spatial information of small-scale targets, thereby enhancing the network's ability to learn small-scale target features and boundary features. Experiments show that our designed network architecture can improve the segmentation accuracy of small-scale objects and optimize the edge details of segmented objects.
{"title":"Multitask Semantic Segmentation Network Using Adaptive Multiscale Feature Fusion","authors":"Huilin Chen, Shengsong Yang, Ting Lyu","doi":"10.1109/CCPQT56151.2022.00018","DOIUrl":"https://doi.org/10.1109/CCPQT56151.2022.00018","url":null,"abstract":"A multi-task semantic segmentation network architecture based on adaptive multi-scale feature fusion is proposed, which improves segmentation target edge details and small-scale target segmentation accuracy by combining boundary detection tasks and semantic segmentation tasks. The critical component of the architecture is the adaptive multi-scale feature fusion module, which can adaptively fuse the semantic feature information and boundary feature information of different scales, extract semantic features that contain more spatial data, and reduce the loss of spatial information of small-scale targets, thereby enhancing the network's ability to learn small-scale target features and boundary features. Experiments show that our designed network architecture can improve the segmentation accuracy of small-scale objects and optimize the edge details of segmented objects.","PeriodicalId":235893,"journal":{"name":"2022 International Conference on Computing, Communication, Perception and Quantum Technology (CCPQT)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128853456","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Pub Date : 2022-08-01DOI: 10.1109/CCPQT56151.2022.00034
Haojie Lin, Hong Wen, Wenjing Hou, Wenxin Lei
Federated learning enables distributed devices to conduct cooperative training models while protecting data privacy, so it is widely promoted in big data scenario and the scope of the Internet of Things. Federated learning in multi-tier computing can integrate the resources of the device-edge-fog-cloud layer to interact and cooperate. For example, in addition to offloading training locally, the tasks of the device layer can also be uploaded to the edge layer or the fog layer for training, while the global aggregation node can be selected at the edge or fog or cloud. However, due to the uncertainty of network bandwidth, computing resources and terminal training tasks at each layer, it brings challenges to resource allocation and task offloading under federated learning in multi-tier computing. Therefore, we propose a TD3-based algorithm which aims to solve how to select training nodes and aggregation nodes during joint training on multi-tier federated learning to minimize the average task delay. Numerical experiments show that our method has better performance in terms of energy consumption and delay compared with edge federated learning and traditional federated learning.
{"title":"TD3-based Algorithm for Node Selection on Multi-tier Federated Learning","authors":"Haojie Lin, Hong Wen, Wenjing Hou, Wenxin Lei","doi":"10.1109/CCPQT56151.2022.00034","DOIUrl":"https://doi.org/10.1109/CCPQT56151.2022.00034","url":null,"abstract":"Federated learning enables distributed devices to conduct cooperative training models while protecting data privacy, so it is widely promoted in big data scenario and the scope of the Internet of Things. Federated learning in multi-tier computing can integrate the resources of the device-edge-fog-cloud layer to interact and cooperate. For example, in addition to offloading training locally, the tasks of the device layer can also be uploaded to the edge layer or the fog layer for training, while the global aggregation node can be selected at the edge or fog or cloud. However, due to the uncertainty of network bandwidth, computing resources and terminal training tasks at each layer, it brings challenges to resource allocation and task offloading under federated learning in multi-tier computing. Therefore, we propose a TD3-based algorithm which aims to solve how to select training nodes and aggregation nodes during joint training on multi-tier federated learning to minimize the average task delay. Numerical experiments show that our method has better performance in terms of energy consumption and delay compared with edge federated learning and traditional federated learning.","PeriodicalId":235893,"journal":{"name":"2022 International Conference on Computing, Communication, Perception and Quantum Technology (CCPQT)","volume":"131 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-08-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114662620","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：