Pub Date : 2023-05-01DOI: 10.1109/IDITR57726.2023.10145929
Yikai Guo, Jie Liu, Zhiyu Duan, Zhizhao Luo, Wei Li, Bowei Ning
PISA (Protocol Independent Switch Architecture) is one of the mainstream programmable chip architectures, allowing users to define the functions and behaviors of the data plane with PISA language. The PISA chip optimizes robot design by providing flexible data processing capabilities, reducing the difficulty of robot maintenance, and enhancing the interoperability of the robot with other devices or systems. However, the application of the PISA chip faces two challenges. (1) Due to the complex dependencies and constraints in the chip, the current chip resource allocation algorithms waste a lot of pipeline stages when executing basic blocks of programs. (2) The time complexity of existing resource scheduling algorithms is high, which affects the efficiency of the chip. To tackle these challenges, we introduce a low-complexity greedy tuning algorithm to schedule the resource of the chip. Firstly, we unify the data dependencies and control dependencies as path constraints, which are derived from the program flowchart and variable read-write information. Then the initial feasible schedule scheme is constructed under the given constraints. Finally, we optimize the initial feasible solution by designing a new data structure, namely segment flowchart, which efficiently solves the resource-sharing problem among basic blocks within the same pipeline stage. In a practical example, basic blocks are successfully placed into the pipeline under given constraints, saving 14.9% of pipeline stages compared with the sequential approach. Our code and dataset are publicly available.
{"title":"Greedy Tuning Algorithm for Resource Scheduling of PISA Chips","authors":"Yikai Guo, Jie Liu, Zhiyu Duan, Zhizhao Luo, Wei Li, Bowei Ning","doi":"10.1109/IDITR57726.2023.10145929","DOIUrl":"https://doi.org/10.1109/IDITR57726.2023.10145929","url":null,"abstract":"PISA (Protocol Independent Switch Architecture) is one of the mainstream programmable chip architectures, allowing users to define the functions and behaviors of the data plane with PISA language. The PISA chip optimizes robot design by providing flexible data processing capabilities, reducing the difficulty of robot maintenance, and enhancing the interoperability of the robot with other devices or systems. However, the application of the PISA chip faces two challenges. (1) Due to the complex dependencies and constraints in the chip, the current chip resource allocation algorithms waste a lot of pipeline stages when executing basic blocks of programs. (2) The time complexity of existing resource scheduling algorithms is high, which affects the efficiency of the chip. To tackle these challenges, we introduce a low-complexity greedy tuning algorithm to schedule the resource of the chip. Firstly, we unify the data dependencies and control dependencies as path constraints, which are derived from the program flowchart and variable read-write information. Then the initial feasible schedule scheme is constructed under the given constraints. Finally, we optimize the initial feasible solution by designing a new data structure, namely segment flowchart, which efficiently solves the resource-sharing problem among basic blocks within the same pipeline stage. In a practical example, basic blocks are successfully placed into the pipeline under given constraints, saving 14.9% of pipeline stages compared with the sequential approach. Our code and dataset are publicly available.","PeriodicalId":272880,"journal":{"name":"2023 2nd International Conference on Innovations and Development of Information Technologies and Robotics (IDITR)","volume":"47 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123900348","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 : 2023-05-01DOI: 10.1109/IDITR57726.2023.10145968
Miaomiao Cheng, Jiahui Zhang, Shuibing Long
Consistent hashing plays an important role in distributed systems as well as in distributed graph systems. It assumes the role of load balancing and data relocation for system expansion and contraction. Graph partitioning is used by many distributed graph systems, which can improve the system’s performance. In distributed systems, expansion and contraction cause data relocation. Data relocation consumes CPU and leads to the unreliability of online services.The system’s performance using the graph partition algorithm to optimize data distribution will inevitably decline with data relocation. Data migration due to capacity expansion is accidental, and system efficiency is a long-term indicator. We present GMaglev to find a way to balance the data migration and the fanout optimization. Experiments on multiple datasets show the effectiveness of GMaglev algorithm. In the case of only 5% increase in data relocation radio, fanout is reduced by 20%.
{"title":"GMaglev: Graph-friendly Consistent Hashing for Distributed Social Graph Partition","authors":"Miaomiao Cheng, Jiahui Zhang, Shuibing Long","doi":"10.1109/IDITR57726.2023.10145968","DOIUrl":"https://doi.org/10.1109/IDITR57726.2023.10145968","url":null,"abstract":"Consistent hashing plays an important role in distributed systems as well as in distributed graph systems. It assumes the role of load balancing and data relocation for system expansion and contraction. Graph partitioning is used by many distributed graph systems, which can improve the system’s performance. In distributed systems, expansion and contraction cause data relocation. Data relocation consumes CPU and leads to the unreliability of online services.The system’s performance using the graph partition algorithm to optimize data distribution will inevitably decline with data relocation. Data migration due to capacity expansion is accidental, and system efficiency is a long-term indicator. We present GMaglev to find a way to balance the data migration and the fanout optimization. Experiments on multiple datasets show the effectiveness of GMaglev algorithm. In the case of only 5% increase in data relocation radio, fanout is reduced by 20%.","PeriodicalId":272880,"journal":{"name":"2023 2nd International Conference on Innovations and Development of Information Technologies and Robotics (IDITR)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125817756","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 : 2023-05-01DOI: 10.1109/IDITR57726.2023.10145819
Yu Cao, Zhou Huang, Xingchen Zhang, Gang Liu, R. Y. Hou
Over the past two decades, building an intelligent transportation system has become a popular and challenging research topic. As a key role in such a system, accurate traffic speed forecasting is critical. Although many powerful prediction methods have been proposed, they have not considered the application of models in real situations, that is, on various types of roads with different characteristics. So, we apply some representative and state-of-the-art methods on different types of roads to help people select the appropriate prediction method to construct an intelligent transportation system. First, we use the traffic data of 214 roads in Guangzhou in 61 days as the data set, and select four typical roads according to their characteristics of the roads. Then we use feature engineering to enhance the quality of the data set. After that, we apply Autoregressive Integrated Moving Average (ARIMA), Exponential Smoothing, short-term and long-term memory (LSTM) neural networks, and Informer on selected roads to make comparisons. The performance of models varies significantly in different types of roads: The Mean Absolute Error (MAE) for low mean and low variance roads is around 2, but the MAE for high mean and high variance roads is about 5. Notably, the Holt-Winters model shows the best performance in short-period prediction, and the Informer model offers the best performance in long-period prediction in the benchmarking.
{"title":"Traffic Speed Forecasting: Comparison of Modeling Approaches","authors":"Yu Cao, Zhou Huang, Xingchen Zhang, Gang Liu, R. Y. Hou","doi":"10.1109/IDITR57726.2023.10145819","DOIUrl":"https://doi.org/10.1109/IDITR57726.2023.10145819","url":null,"abstract":"Over the past two decades, building an intelligent transportation system has become a popular and challenging research topic. As a key role in such a system, accurate traffic speed forecasting is critical. Although many powerful prediction methods have been proposed, they have not considered the application of models in real situations, that is, on various types of roads with different characteristics. So, we apply some representative and state-of-the-art methods on different types of roads to help people select the appropriate prediction method to construct an intelligent transportation system. First, we use the traffic data of 214 roads in Guangzhou in 61 days as the data set, and select four typical roads according to their characteristics of the roads. Then we use feature engineering to enhance the quality of the data set. After that, we apply Autoregressive Integrated Moving Average (ARIMA), Exponential Smoothing, short-term and long-term memory (LSTM) neural networks, and Informer on selected roads to make comparisons. The performance of models varies significantly in different types of roads: The Mean Absolute Error (MAE) for low mean and low variance roads is around 2, but the MAE for high mean and high variance roads is about 5. Notably, the Holt-Winters model shows the best performance in short-period prediction, and the Informer model offers the best performance in long-period prediction in the benchmarking.","PeriodicalId":272880,"journal":{"name":"2023 2nd International Conference on Innovations and Development of Information Technologies and Robotics (IDITR)","volume":"53 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130509833","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 : 2023-05-01DOI: 10.1109/IDITR57726.2023.10145817
Yifan Shi, Xiao Jing, Lushi Liu
In this paper, we optimize the locomotion speed of a micro-bristle robot using three surrogate model optimization methods: Kriging method, Bayesian method, and Deep Neural Network. Moreover, the current most popular optimization algorithm in the micro-robot optimization field, the genetic algorithm, is used as the baseline method for comparison. The four methods’ performances are tested in MATLAB, during which a state-of-art dynamic model is used. Then we 3D print the robot designs obtained from these methods and test these robot designs’ real performances. This is the first time that surrogate model optimization methods are applied on micro-robot design field. The MATLAB optimization results and the robot experimental results show that applying proper surrogate model optimization methods, especially Bayesian method will be able to obtain a satisfying robot design 5-6 times faster than the time spent by genetic algorithm. The paper provides an efficient guidance on micro-robot optimization field.
{"title":"Micro-Bristle Robot Design Via Different Surrogate Model Optimization Methods","authors":"Yifan Shi, Xiao Jing, Lushi Liu","doi":"10.1109/IDITR57726.2023.10145817","DOIUrl":"https://doi.org/10.1109/IDITR57726.2023.10145817","url":null,"abstract":"In this paper, we optimize the locomotion speed of a micro-bristle robot using three surrogate model optimization methods: Kriging method, Bayesian method, and Deep Neural Network. Moreover, the current most popular optimization algorithm in the micro-robot optimization field, the genetic algorithm, is used as the baseline method for comparison. The four methods’ performances are tested in MATLAB, during which a state-of-art dynamic model is used. Then we 3D print the robot designs obtained from these methods and test these robot designs’ real performances. This is the first time that surrogate model optimization methods are applied on micro-robot design field. The MATLAB optimization results and the robot experimental results show that applying proper surrogate model optimization methods, especially Bayesian method will be able to obtain a satisfying robot design 5-6 times faster than the time spent by genetic algorithm. The paper provides an efficient guidance on micro-robot optimization field.","PeriodicalId":272880,"journal":{"name":"2023 2nd International Conference on Innovations and Development of Information Technologies and Robotics (IDITR)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121375710","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 : 2023-05-01DOI: 10.1109/IDITR57726.2023.10145975
Le Kang, Zhihua Chen, Tianyuan Yuan, Yunjiang Cheng
Aiming at the problems of complex working conditions, diverse fault modes and high-test costs of the underwater vehicle, a fault injector is designed based on TMS320F28335 digital signal processor. By analyzing the functional requirements, the hardware functional modules and software workflow of the fault injector are designed. After that, a fault injection experimental verification platform is built, and the effectiveness of the basic functions of the fault injector is verified by taking the speed sensor of the underwater vehicle as an example. The results show that the fault injector can meet the fault injection requirements of different fault modes, fault injection nodes and fault injection time, which is of great significance to the reliability research of the underwater vehicle.
{"title":"Design of Fault Injector for Underwater Vehicle Sensors","authors":"Le Kang, Zhihua Chen, Tianyuan Yuan, Yunjiang Cheng","doi":"10.1109/IDITR57726.2023.10145975","DOIUrl":"https://doi.org/10.1109/IDITR57726.2023.10145975","url":null,"abstract":"Aiming at the problems of complex working conditions, diverse fault modes and high-test costs of the underwater vehicle, a fault injector is designed based on TMS320F28335 digital signal processor. By analyzing the functional requirements, the hardware functional modules and software workflow of the fault injector are designed. After that, a fault injection experimental verification platform is built, and the effectiveness of the basic functions of the fault injector is verified by taking the speed sensor of the underwater vehicle as an example. The results show that the fault injector can meet the fault injection requirements of different fault modes, fault injection nodes and fault injection time, which is of great significance to the reliability research of the underwater vehicle.","PeriodicalId":272880,"journal":{"name":"2023 2nd International Conference on Innovations and Development of Information Technologies and Robotics (IDITR)","volume":"66 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132718954","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 : 2023-05-01DOI: 10.1109/IDITR57726.2023.10145820
Sheng-Fu Wang, Jing Huang, Baohua Zhang, Jia Li
Recently, a growing number of Named Entity Recognition (NER) methods utilize cross-sentence features (also known as contexts) to improve the performance of NER models, instead of using single-sentence information alone. As far as we know, most NER models choose to exploit pre- and post-sentences to capture cross-sentence features. Generally, current NER studies focus only on the model architecture to capture better token representations. However, there is no in-depth exploration on how to better model cross-sentence features. In this paper, based on the span classification model, we investigate the effect of cross-sentence features under different settings. Specifically, we evaluate the impact of context stitching, context window size, context window padding, and classifier token of pre-trained language model (PLM) on model performance. Comparative experimental results show that appropriate incorporation of document-level contexts can considerably improve the NER metrics. Furthermore, we find that several factors can be used to improve the performance of NER models: (1) use domain-specific PLMs, but not classifier tokens; (2) use only preceding contexts for generic text, and random contexts for specialized text; (3) truncate overly long contexts when the context window is small, and preserve sentence integrity when the window is large; (4) set the context window size to about 200 for the basic size PLM.
{"title":"A Comparative Study of Cross-Sentence Features for Named Entity Recognition","authors":"Sheng-Fu Wang, Jing Huang, Baohua Zhang, Jia Li","doi":"10.1109/IDITR57726.2023.10145820","DOIUrl":"https://doi.org/10.1109/IDITR57726.2023.10145820","url":null,"abstract":"Recently, a growing number of Named Entity Recognition (NER) methods utilize cross-sentence features (also known as contexts) to improve the performance of NER models, instead of using single-sentence information alone. As far as we know, most NER models choose to exploit pre- and post-sentences to capture cross-sentence features. Generally, current NER studies focus only on the model architecture to capture better token representations. However, there is no in-depth exploration on how to better model cross-sentence features. In this paper, based on the span classification model, we investigate the effect of cross-sentence features under different settings. Specifically, we evaluate the impact of context stitching, context window size, context window padding, and classifier token of pre-trained language model (PLM) on model performance. Comparative experimental results show that appropriate incorporation of document-level contexts can considerably improve the NER metrics. Furthermore, we find that several factors can be used to improve the performance of NER models: (1) use domain-specific PLMs, but not classifier tokens; (2) use only preceding contexts for generic text, and random contexts for specialized text; (3) truncate overly long contexts when the context window is small, and preserve sentence integrity when the window is large; (4) set the context window size to about 200 for the basic size PLM.","PeriodicalId":272880,"journal":{"name":"2023 2nd International Conference on Innovations and Development of Information Technologies and Robotics (IDITR)","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115850466","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 : 2023-05-01DOI: 10.1109/IDITR57726.2023.10145941
Xusen Lang, Chenyao Bai, Yunlong Zhu
Electronic consumer products are closely related to life. With the increase of demand, various consumer electronic products have emerged as the times require. For some high-precision equipment, such as in all aspects of the chip manufacturing process, the requirements for PCB are relatively high. In the PCB manufacturing process, it is often due to various problems caused by improper operation of certain links in the process. Many defects may appear on the PCB, such as bubbles appearing when the film is not firmly attached to the ground during the film attachment process, and bubbles appear during the exposure process. Negative film scratches, excessive pressure during etching, unevenness during plating, etc. Printed board surface defects vary in size. This inconsistency of multi-scale features will cause the pooling operation of the network model to lose some fine-grained spatial features. In the field of object detection, in the past, the detection effect in the field of PCB defect detection was poor, and the natural defects were few and small, and faced some bottlenecks in engineering. Aiming at this problem, a PCB defect detection method based on RDIDet is proposed. The experimental results prove that the improved network has obvious performance advantages over the previous classic model, with an accuracy rate of 98.3%, and a better detection effect on PCB defects.
{"title":"RDIDI: Recognition of Defect Image Detection in Industry","authors":"Xusen Lang, Chenyao Bai, Yunlong Zhu","doi":"10.1109/IDITR57726.2023.10145941","DOIUrl":"https://doi.org/10.1109/IDITR57726.2023.10145941","url":null,"abstract":"Electronic consumer products are closely related to life. With the increase of demand, various consumer electronic products have emerged as the times require. For some high-precision equipment, such as in all aspects of the chip manufacturing process, the requirements for PCB are relatively high. In the PCB manufacturing process, it is often due to various problems caused by improper operation of certain links in the process. Many defects may appear on the PCB, such as bubbles appearing when the film is not firmly attached to the ground during the film attachment process, and bubbles appear during the exposure process. Negative film scratches, excessive pressure during etching, unevenness during plating, etc. Printed board surface defects vary in size. This inconsistency of multi-scale features will cause the pooling operation of the network model to lose some fine-grained spatial features. In the field of object detection, in the past, the detection effect in the field of PCB defect detection was poor, and the natural defects were few and small, and faced some bottlenecks in engineering. Aiming at this problem, a PCB defect detection method based on RDIDet is proposed. The experimental results prove that the improved network has obvious performance advantages over the previous classic model, with an accuracy rate of 98.3%, and a better detection effect on PCB defects.","PeriodicalId":272880,"journal":{"name":"2023 2nd International Conference on Innovations and Development of Information Technologies and Robotics (IDITR)","volume":"13 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125055983","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 : 2023-05-01DOI: 10.1109/IDITR57726.2023.10145970
Jin Li, Z. Lv, Jinhui You, Wei Guo
Crane hoisting is an important part of smelting operations. The traditional operation method is that the ground dispatcher and the driver work together, which is not only inefficient but also has great potential safety hazards. To solve the above problems, this paper designs a video monitoring system. The PTZ camera is used for video monitoring of the crane hook operation, the laser sensor and the rotary encoder collect data, and the Kalman-based detection adaptive filtering algorithm is used for filtering processing to obtain the precise motion trajectory of the hook. On the basis of the traditional PID controller, the DMC controller based on predictive control is designed to solve the control lag problem in the control process of PTZ cameras. The experimental results show that the monitoring system can clearly display the hook in the center of the screen, help the crane operator to grasp the status of the hook, improve the production efficiency of the workshop, and ensure the safety of hoisting.
{"title":"Crane Hook Motion Monitoring Based on Trajectory Filtering and Tracking Video Monitoring","authors":"Jin Li, Z. Lv, Jinhui You, Wei Guo","doi":"10.1109/IDITR57726.2023.10145970","DOIUrl":"https://doi.org/10.1109/IDITR57726.2023.10145970","url":null,"abstract":"Crane hoisting is an important part of smelting operations. The traditional operation method is that the ground dispatcher and the driver work together, which is not only inefficient but also has great potential safety hazards. To solve the above problems, this paper designs a video monitoring system. The PTZ camera is used for video monitoring of the crane hook operation, the laser sensor and the rotary encoder collect data, and the Kalman-based detection adaptive filtering algorithm is used for filtering processing to obtain the precise motion trajectory of the hook. On the basis of the traditional PID controller, the DMC controller based on predictive control is designed to solve the control lag problem in the control process of PTZ cameras. The experimental results show that the monitoring system can clearly display the hook in the center of the screen, help the crane operator to grasp the status of the hook, improve the production efficiency of the workshop, and ensure the safety of hoisting.","PeriodicalId":272880,"journal":{"name":"2023 2nd International Conference on Innovations and Development of Information Technologies and Robotics (IDITR)","volume":"41 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133135736","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 : 2023-05-01DOI: 10.1109/IDITR57726.2023.10145973
Yongqiang Gu, Shangrong Wang, Shaolong Han, Ruixu Han
The speed control of the climbing belt motor of the cut tobacco feeder does not match the demand of the cut tobacco feeding flow, and there are a series of problems such as large fluctuation of the mass flow, cut tobacco shortage, cut tobacco blockage, climbing belt motor start or stop frequently, large equipment loss and energy consumption. An improved speed control method of the climbing belt motor is proposed to solve the above problems. Firstly, the relationship between motor frequency and mass flow is modeled and analyzed; Then, a reasonable frequency modulation control logic is proposed; Finally, the model and control logic are verified and further improved based on production data. The experimental results show that: After improvement, the average starting or stopping times of the motor are reduced by 98.61%, the average high-frequency triggering times are reduced by 96.48%, and average value of flow range in the head stage is reduced by 75.78%. While improving the stability of the cut tobacco mass flow, the stability of the climbing belt motor has also been greatly improved, which not only eliminates the hidden danger of cut tobacco shortage and blockage, but also reduces the equipment loss and energy consumption.
{"title":"Design of a Constant Flow Control System for Cut Tobacco Feeder","authors":"Yongqiang Gu, Shangrong Wang, Shaolong Han, Ruixu Han","doi":"10.1109/IDITR57726.2023.10145973","DOIUrl":"https://doi.org/10.1109/IDITR57726.2023.10145973","url":null,"abstract":"The speed control of the climbing belt motor of the cut tobacco feeder does not match the demand of the cut tobacco feeding flow, and there are a series of problems such as large fluctuation of the mass flow, cut tobacco shortage, cut tobacco blockage, climbing belt motor start or stop frequently, large equipment loss and energy consumption. An improved speed control method of the climbing belt motor is proposed to solve the above problems. Firstly, the relationship between motor frequency and mass flow is modeled and analyzed; Then, a reasonable frequency modulation control logic is proposed; Finally, the model and control logic are verified and further improved based on production data. The experimental results show that: After improvement, the average starting or stopping times of the motor are reduced by 98.61%, the average high-frequency triggering times are reduced by 96.48%, and average value of flow range in the head stage is reduced by 75.78%. While improving the stability of the cut tobacco mass flow, the stability of the climbing belt motor has also been greatly improved, which not only eliminates the hidden danger of cut tobacco shortage and blockage, but also reduces the equipment loss and energy consumption.","PeriodicalId":272880,"journal":{"name":"2023 2nd International Conference on Innovations and Development of Information Technologies and Robotics (IDITR)","volume":"98 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115727658","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 : 2023-05-01DOI: 10.1109/IDITR57726.2023.10145988
Xi-kun Zhang, Jie Hou
With the continuous advancement of the life sciences and the advancement of ultra-high resolution technology, people can observe biological information such as organelles and molecules, and can understand their internal structure and interaction from the acquired cell diffraction images. However, cells contain various types of organelles, which have high heterogeneity, and the cell structures of different types of cells have certain differences. Therefore, the study of cell diffraction image classification is of great significance in many fields such as cell morphology and cell biology. The research task of image classification is to extract useful feature information from the image, and then distinguish the images of different attributes, and finally divide the image targets of different categories. Deep learning techniques are used in a variety of industries, including picture categorization, as a result of the development of deep learning. Among these, there has been a notable improvement in the accuracy of picture classification using deep neural networks. The classification accuracy can be further increased in the real cell diffraction image classification procedure, though. This study proposes a deep neural network-based classification strategy for cell diffraction pictures. For the cell diffraction images, some interference pictures are created by cell debris or impurities. In this work, the produced diffraction pictures are preprocessed using a clustering approach and a support vector machine (SVM).After that, a Gray Level Co-occurrence Matrix(GLCM) is used to extract the texture features of the diffraction image. This research proposes an enhanced deep neural network-based picture classification algorithm. Max-margin Minimum Classification Error (M3CE) is introduced during the training of deep neural networks, and cross-entropy is used to build the loss function. Finally, the Ramos and Jurkat cell experiments' findings support the high precision of the categorization approach.
{"title":"Classification of Cell Diffraction Images Based on Deep Neural Network","authors":"Xi-kun Zhang, Jie Hou","doi":"10.1109/IDITR57726.2023.10145988","DOIUrl":"https://doi.org/10.1109/IDITR57726.2023.10145988","url":null,"abstract":"With the continuous advancement of the life sciences and the advancement of ultra-high resolution technology, people can observe biological information such as organelles and molecules, and can understand their internal structure and interaction from the acquired cell diffraction images. However, cells contain various types of organelles, which have high heterogeneity, and the cell structures of different types of cells have certain differences. Therefore, the study of cell diffraction image classification is of great significance in many fields such as cell morphology and cell biology. The research task of image classification is to extract useful feature information from the image, and then distinguish the images of different attributes, and finally divide the image targets of different categories. Deep learning techniques are used in a variety of industries, including picture categorization, as a result of the development of deep learning. Among these, there has been a notable improvement in the accuracy of picture classification using deep neural networks. The classification accuracy can be further increased in the real cell diffraction image classification procedure, though. This study proposes a deep neural network-based classification strategy for cell diffraction pictures. For the cell diffraction images, some interference pictures are created by cell debris or impurities. In this work, the produced diffraction pictures are preprocessed using a clustering approach and a support vector machine (SVM).After that, a Gray Level Co-occurrence Matrix(GLCM) is used to extract the texture features of the diffraction image. This research proposes an enhanced deep neural network-based picture classification algorithm. Max-margin Minimum Classification Error (M3CE) is introduced during the training of deep neural networks, and cross-entropy is used to build the loss function. Finally, the Ramos and Jurkat cell experiments' findings support the high precision of the categorization approach.","PeriodicalId":272880,"journal":{"name":"2023 2nd International Conference on Innovations and Development of Information Technologies and Robotics (IDITR)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2023-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131122950","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：