Pub Date : 2022-12-10DOI: 10.1109/STCR55312.2022.10009452
H. Rajaguru, R. Karthikamani
An electroencephalogram is a medical method that employs electrical signals to analyze brain activity. The (EEG) signal is commonly measured using Scalp electrodes, which is very useful in identifying a patient's brain status and epilepsy as well as supplementing CT scan measurements. EEG signals indirectly reveal the state of the brain. In this paper the performance of the classifiers are analyzed to detect Normal sleep and Seizure EEG signals. Features are extracted using six statistical features such as Mean, Variance, Skewness, Kurtosis, sample entropy, Pearson Correlation coefficient. The Detrend Fluctuation Analysis, Detrend Fluctuation Analysis Expectation Maximization, Detrend Fluctuation Analysis Firefly, Detrend Fluctuation Analysis with Gaussian Mixture Model, Detrend with Bayesian Linear Discriminant Classifiers are employed to detect the Normal sleep and Seizure from EEG signal. The hybrid classifier Detrend Fluctuation Analysis with EM achieved the highest accuracy of 98.96% for Seizure EEG signal and an accuracy of 97.66% using Detrend Fluctuation Analysis classifier for normal sleep EEG signal.
{"title":"Performance Analysis of the Classifier in the Classification of Normal-Sleep and Seizure from EEG Signal","authors":"H. Rajaguru, R. Karthikamani","doi":"10.1109/STCR55312.2022.10009452","DOIUrl":"https://doi.org/10.1109/STCR55312.2022.10009452","url":null,"abstract":"An electroencephalogram is a medical method that employs electrical signals to analyze brain activity. The (EEG) signal is commonly measured using Scalp electrodes, which is very useful in identifying a patient's brain status and epilepsy as well as supplementing CT scan measurements. EEG signals indirectly reveal the state of the brain. In this paper the performance of the classifiers are analyzed to detect Normal sleep and Seizure EEG signals. Features are extracted using six statistical features such as Mean, Variance, Skewness, Kurtosis, sample entropy, Pearson Correlation coefficient. The Detrend Fluctuation Analysis, Detrend Fluctuation Analysis Expectation Maximization, Detrend Fluctuation Analysis Firefly, Detrend Fluctuation Analysis with Gaussian Mixture Model, Detrend with Bayesian Linear Discriminant Classifiers are employed to detect the Normal sleep and Seizure from EEG signal. The hybrid classifier Detrend Fluctuation Analysis with EM achieved the highest accuracy of 98.96% for Seizure EEG signal and an accuracy of 97.66% using Detrend Fluctuation Analysis classifier for normal sleep EEG signal.","PeriodicalId":338691,"journal":{"name":"2022 Smart Technologies, Communication and Robotics (STCR)","volume":"155 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116589192","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 : 2022-12-10DOI: 10.1109/STCR55312.2022.10009526
Swetha Patil, Suparna H S, B. N, D. N
Agriculture being one of the domains in which application of modern technologies including Machine Learning(ML) is relatively less. Disease in rice leaf can cause huge loss to farmers and they need to be predicted accurately at early stage. This research work focuses on implementing a framework to detect the rice leaf disease using ML algorithms. Using the thermal images of rice leaves, 20 statistical features are extracted and weighted feature selection is done using Puzzle Optimization Algorithm. This Optimization Algorithm is applied for variety of applications but usage as weighted feature selection technique is relatively new. 636 thermal images of rice leaves belonging to six different classes namely blast, bacteria leaf blight, brown leaf spot, hispa, leaf folder and healthy leaves are retrieved from the publicly available website & considered in this analysis. Four classifiers namely extremely randomized trees classifier, Naïve bayes classifier, quadratic discriminant analysis classifier, and decision tree classifier are tested. Among them, extremely randomized trees classifier without any feature selection method offers the highest balanced accuracy score of 0.76 and it is raised to 0.84 when Puzzle optimization algorithm is used to select weighted features.
{"title":"Puzzle Optimization Algorithm based Weighted Feature Selection for Identification of Rice Leaf Disease Through Thermal Images","authors":"Swetha Patil, Suparna H S, B. N, D. N","doi":"10.1109/STCR55312.2022.10009526","DOIUrl":"https://doi.org/10.1109/STCR55312.2022.10009526","url":null,"abstract":"Agriculture being one of the domains in which application of modern technologies including Machine Learning(ML) is relatively less. Disease in rice leaf can cause huge loss to farmers and they need to be predicted accurately at early stage. This research work focuses on implementing a framework to detect the rice leaf disease using ML algorithms. Using the thermal images of rice leaves, 20 statistical features are extracted and weighted feature selection is done using Puzzle Optimization Algorithm. This Optimization Algorithm is applied for variety of applications but usage as weighted feature selection technique is relatively new. 636 thermal images of rice leaves belonging to six different classes namely blast, bacteria leaf blight, brown leaf spot, hispa, leaf folder and healthy leaves are retrieved from the publicly available website & considered in this analysis. Four classifiers namely extremely randomized trees classifier, Naïve bayes classifier, quadratic discriminant analysis classifier, and decision tree classifier are tested. Among them, extremely randomized trees classifier without any feature selection method offers the highest balanced accuracy score of 0.76 and it is raised to 0.84 when Puzzle optimization algorithm is used to select weighted features.","PeriodicalId":338691,"journal":{"name":"2022 Smart Technologies, Communication and Robotics (STCR)","volume":"438 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"122882865","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 : 2022-12-10DOI: 10.1109/STCR55312.2022.10009133
S. Saranya, G. Flora, S. Malini, S. Sanjay
The novel power management is focused on using IOT to regulate a hybrid energy system. There are many different types of energy that are all alternatives to one another, such as solar energy, wind energy, biofuel, and fuel cells. When a hybrid energy system is constructed for personal or commercial use, however, it is necessary to control it. At this moment, IOT plays a critical function in system control. The major objectives were to transition between the two energy sources, solar and wind, short of causing any inconvenience via a website utilizing a Wi-Fi module. The information is sent wirelessly to the ESP8266, which regulates the energy.
{"title":"A Web-based Optimized Hybrid Power Management with IoT","authors":"S. Saranya, G. Flora, S. Malini, S. Sanjay","doi":"10.1109/STCR55312.2022.10009133","DOIUrl":"https://doi.org/10.1109/STCR55312.2022.10009133","url":null,"abstract":"The novel power management is focused on using IOT to regulate a hybrid energy system. There are many different types of energy that are all alternatives to one another, such as solar energy, wind energy, biofuel, and fuel cells. When a hybrid energy system is constructed for personal or commercial use, however, it is necessary to control it. At this moment, IOT plays a critical function in system control. The major objectives were to transition between the two energy sources, solar and wind, short of causing any inconvenience via a website utilizing a Wi-Fi module. The information is sent wirelessly to the ESP8266, which regulates the energy.","PeriodicalId":338691,"journal":{"name":"2022 Smart Technologies, Communication and Robotics (STCR)","volume":"197 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128358528","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 : 2022-12-10DOI: 10.1109/STCR55312.2022.10009282
Muzamil Rouf, Mir Nazish, Ishfaq Sultan, M. T. Banday
The ARM Cortex-M series of cores incorporate a variety of configurable performance preferences, which help the designers to use desired cores for their applications. Although it is much more common to find ARM Cortex-M cores implemented in Microcontroller Units (MCUs) with memories, clocks and peripherals integrated within the MCU, FPGA implementation of ARM Cortex-M as a soft core can be used to design optimised cores. The ARM DesignStart program provides the CPU and Physical IP solutions that have enabled thousands of organisations worldwide to access, assess, and create System on Chips (SoCs) with ARM IPs. This paper reports the optimisation of the ARM Cortex-M1 and -M3 cores for area and power metrics. The designed cores have been tested for the Digilent Arty A7 FPGA platform. The results report an 8% and 24% reduction in LUT utilisation for Cortex-M1 and -M3 cores, respectively. In addition, the power consumption of the Cortex-M1 and -M3 cores decreases by 25% and 5%, respectively. These results justify using the optimised cores for resource-constrained IoT applications and help designers build power and area-efficient SoCs for low-end devices. Furthermore, the Cortex M-23 and Cortex M-33 cores have been implemented on ARM V2M MPS and tested for secure and non-secure modes in the Keil MDK software development platform.
{"title":"Implementation of Area and Power Optimised ARM Cortex-M Cores on FPGA","authors":"Muzamil Rouf, Mir Nazish, Ishfaq Sultan, M. T. Banday","doi":"10.1109/STCR55312.2022.10009282","DOIUrl":"https://doi.org/10.1109/STCR55312.2022.10009282","url":null,"abstract":"The ARM Cortex-M series of cores incorporate a variety of configurable performance preferences, which help the designers to use desired cores for their applications. Although it is much more common to find ARM Cortex-M cores implemented in Microcontroller Units (MCUs) with memories, clocks and peripherals integrated within the MCU, FPGA implementation of ARM Cortex-M as a soft core can be used to design optimised cores. The ARM DesignStart program provides the CPU and Physical IP solutions that have enabled thousands of organisations worldwide to access, assess, and create System on Chips (SoCs) with ARM IPs. This paper reports the optimisation of the ARM Cortex-M1 and -M3 cores for area and power metrics. The designed cores have been tested for the Digilent Arty A7 FPGA platform. The results report an 8% and 24% reduction in LUT utilisation for Cortex-M1 and -M3 cores, respectively. In addition, the power consumption of the Cortex-M1 and -M3 cores decreases by 25% and 5%, respectively. These results justify using the optimised cores for resource-constrained IoT applications and help designers build power and area-efficient SoCs for low-end devices. Furthermore, the Cortex M-23 and Cortex M-33 cores have been implemented on ARM V2M MPS and tested for secure and non-secure modes in the Keil MDK software development platform.","PeriodicalId":338691,"journal":{"name":"2022 Smart Technologies, Communication and Robotics (STCR)","volume":"10 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117064605","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 : 2022-12-10DOI: 10.1109/STCR55312.2022.10009589
M. Moses, T. Perarasi, M. R. Raja, S. k, L. Lino
The massive growth in mobile devices and machine type communication devices which demands higher performance leads to higher data traffic and spectrum scarcity problem. Multiplexing of time frequency components with code division enhances better capacity in data transmission that are best suited for Non-orthogonal multiple access (NOMA). Deployment of NOMA helps to explode data traffic in a heterogeneous network and that plays an important model for next generation wireless networks. Joint optimization problems over the channel assignment, user decoding, and allocation of power are formulated to maximize system throughput. Not only throughput, achievable sum rate is also maximized at power allocation. Mixed-integer non-linear problem are resolved for continuous variables as an optimization sub-problem (P1) and integer variables as a matching sub-problem (P2). A power control scheme is focused for resource allocation policy to improvise average performance of sum rates with aid of reinforcement learning. With this keyhole, approaches are addressed with two issues for allocation of resource in NOMA namely dynamic user allocation and resource blocks and network traffic balancing. Results validate the proposed scheme can significantly enhance user sum data rates and thus utilities are compared with known Q-learning based strategy.
{"title":"Energy Efficient QoS Aware Machine Learning Model for Scheduling Users in NOMA Heterogeneous Networks","authors":"M. Moses, T. Perarasi, M. R. Raja, S. k, L. Lino","doi":"10.1109/STCR55312.2022.10009589","DOIUrl":"https://doi.org/10.1109/STCR55312.2022.10009589","url":null,"abstract":"The massive growth in mobile devices and machine type communication devices which demands higher performance leads to higher data traffic and spectrum scarcity problem. Multiplexing of time frequency components with code division enhances better capacity in data transmission that are best suited for Non-orthogonal multiple access (NOMA). Deployment of NOMA helps to explode data traffic in a heterogeneous network and that plays an important model for next generation wireless networks. Joint optimization problems over the channel assignment, user decoding, and allocation of power are formulated to maximize system throughput. Not only throughput, achievable sum rate is also maximized at power allocation. Mixed-integer non-linear problem are resolved for continuous variables as an optimization sub-problem (P1) and integer variables as a matching sub-problem (P2). A power control scheme is focused for resource allocation policy to improvise average performance of sum rates with aid of reinforcement learning. With this keyhole, approaches are addressed with two issues for allocation of resource in NOMA namely dynamic user allocation and resource blocks and network traffic balancing. Results validate the proposed scheme can significantly enhance user sum data rates and thus utilities are compared with known Q-learning based strategy.","PeriodicalId":338691,"journal":{"name":"2022 Smart Technologies, Communication and Robotics (STCR)","volume":"50 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128064046","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 : 2022-12-10DOI: 10.1109/STCR55312.2022.10009378
N. M, K. C, Shineka Varshini A P, Tharunika N, Sakthipriya S A
The evaporator is vital to the pharmaceutical industry as it is used to refine finished pharmaceutical products. The evaporator is used to remove excess water during the manufacturing of pharmaceuticals. The SISO evaporator is used to determine the dry matter content by measuring temperature. System identification is used to create a mathematical model of the evaporator in a pharmaceutical factory. Adjusting the temperature of an evaporator is a laborious process. Thus, we build and implement a Neural network predictive controller for usage in the pharmaceutical industry. To fine-tune the evaporator’s control signal, it can be utilised to predict the device’s future performance. The effectiveness of the controller is evaluated using error metrics like ISE, IAE, ITSE, and ITAE. Time-domain criteria such as rising time, settling time, and overshoot are utilised to better appreciate controller functionality. Based on these analyses, it is clear that the predictive controller is superior than the more common PID controller in use in the pharmaceutical sector.
{"title":"Pathway Guided Deep Neural Network towards Interpretable and Predictive Modeling and Drug Preparation","authors":"N. M, K. C, Shineka Varshini A P, Tharunika N, Sakthipriya S A","doi":"10.1109/STCR55312.2022.10009378","DOIUrl":"https://doi.org/10.1109/STCR55312.2022.10009378","url":null,"abstract":"The evaporator is vital to the pharmaceutical industry as it is used to refine finished pharmaceutical products. The evaporator is used to remove excess water during the manufacturing of pharmaceuticals. The SISO evaporator is used to determine the dry matter content by measuring temperature. System identification is used to create a mathematical model of the evaporator in a pharmaceutical factory. Adjusting the temperature of an evaporator is a laborious process. Thus, we build and implement a Neural network predictive controller for usage in the pharmaceutical industry. To fine-tune the evaporator’s control signal, it can be utilised to predict the device’s future performance. The effectiveness of the controller is evaluated using error metrics like ISE, IAE, ITSE, and ITAE. Time-domain criteria such as rising time, settling time, and overshoot are utilised to better appreciate controller functionality. Based on these analyses, it is clear that the predictive controller is superior than the more common PID controller in use in the pharmaceutical sector.","PeriodicalId":338691,"journal":{"name":"2022 Smart Technologies, Communication and Robotics (STCR)","volume":"104 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134192862","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 : 2022-12-10DOI: 10.1109/STCR55312.2022.10009582
G. N, G. C, V. B, Agathiyan S, Abi Nandha P, A. S, A. S
Automatic Number Plate Detection is an established method to interpret the letters in the number plates. In the last 5-10 years, the number of active vehicles has reached a tremendous growth, the growth has also resulted in increase of the illegal activities. It is hard to keep track of a vehicle due to rapid increase of the vehicles. It is crucially important to keep track of all vehicles by the belonging authorities. In this paper, we use technology open source platform called Tensor flow for machine learning. Primarily, the first step is to give the image of the car. Generally, the given image of the car is in low resolution and has satirical deficit in edge data. So, we need to process pictures which are present, it requires the high level precision. Secondly, this technology henceforth used to retrieve the pictures of the automobile, board which indicate it’s identify in the extracted picture also in a way cropped and converted into grayscale. Final output thus converted into grayscale so that the noise level of the image is reduced and the number plates of different colors also detected. So that the computer doesn’t need different algorithms for different colors. The letters of number plate in the image which is processed is extracted to text using optical character recognition. The extracted text is saved in Excel document, which can be used for future purposes. Assist more when compared with the cutting edge plate acknowledgment approach, the normal change is 3.6%. At long last, we propose a crossover chain of command classification framework relying somewhat using vector technique and the Bayesian rule-three methodology.
{"title":"Automatic Number Plate Detection using Deep Learning","authors":"G. N, G. C, V. B, Agathiyan S, Abi Nandha P, A. S, A. S","doi":"10.1109/STCR55312.2022.10009582","DOIUrl":"https://doi.org/10.1109/STCR55312.2022.10009582","url":null,"abstract":"Automatic Number Plate Detection is an established method to interpret the letters in the number plates. In the last 5-10 years, the number of active vehicles has reached a tremendous growth, the growth has also resulted in increase of the illegal activities. It is hard to keep track of a vehicle due to rapid increase of the vehicles. It is crucially important to keep track of all vehicles by the belonging authorities. In this paper, we use technology open source platform called Tensor flow for machine learning. Primarily, the first step is to give the image of the car. Generally, the given image of the car is in low resolution and has satirical deficit in edge data. So, we need to process pictures which are present, it requires the high level precision. Secondly, this technology henceforth used to retrieve the pictures of the automobile, board which indicate it’s identify in the extracted picture also in a way cropped and converted into grayscale. Final output thus converted into grayscale so that the noise level of the image is reduced and the number plates of different colors also detected. So that the computer doesn’t need different algorithms for different colors. The letters of number plate in the image which is processed is extracted to text using optical character recognition. The extracted text is saved in Excel document, which can be used for future purposes. Assist more when compared with the cutting edge plate acknowledgment approach, the normal change is 3.6%. At long last, we propose a crossover chain of command classification framework relying somewhat using vector technique and the Bayesian rule-three methodology.","PeriodicalId":338691,"journal":{"name":"2022 Smart Technologies, Communication and Robotics (STCR)","volume":"53 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"134355907","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 : 2022-12-10DOI: 10.1109/STCR55312.2022.10009388
Nidish Kumar P, Dineshkumar V, A. M, S. R, E. S.
Micro-electronics are now very important in every part of a person's life in the age of present technology. Due to this, the demand for their components and their availability decreases the amount of time they can be made and raises the failure rate of the final product. Therefore methods that increase the of hardware design and verification effectiveness and efficiency are extremely valuable. Advanced Microcontroller Bus Architecture (AMBA) is an open-standard, on-chip interface inter connect. That provides the standard set of rules to achieve the communication inside the system on chip. AXI-Advanced Extensible interconnect comes under the AMBA family and is used to communicate (transfer) the data from high-speed IP cores (master-slave). High frequency and high performance system designs are offered by AXI. It is a protocol for on-chip communication. It is appropriate for low-delay designs with large bandwidth and frequency. It is compatible with current APB and AHB interfaces. The AXI protocols unique address, data phases and control are one of its defining characteristics. The work involved in the design of AXI protocol in an effective manner using the System Verilog. The design is verified using the QuestaSim tool.
{"title":"Design and Verification of AMBA AXI3 Protocol for High Speed Communication","authors":"Nidish Kumar P, Dineshkumar V, A. M, S. R, E. S.","doi":"10.1109/STCR55312.2022.10009388","DOIUrl":"https://doi.org/10.1109/STCR55312.2022.10009388","url":null,"abstract":"Micro-electronics are now very important in every part of a person's life in the age of present technology. Due to this, the demand for their components and their availability decreases the amount of time they can be made and raises the failure rate of the final product. Therefore methods that increase the of hardware design and verification effectiveness and efficiency are extremely valuable. Advanced Microcontroller Bus Architecture (AMBA) is an open-standard, on-chip interface inter connect. That provides the standard set of rules to achieve the communication inside the system on chip. AXI-Advanced Extensible interconnect comes under the AMBA family and is used to communicate (transfer) the data from high-speed IP cores (master-slave). High frequency and high performance system designs are offered by AXI. It is a protocol for on-chip communication. It is appropriate for low-delay designs with large bandwidth and frequency. It is compatible with current APB and AHB interfaces. The AXI protocols unique address, data phases and control are one of its defining characteristics. The work involved in the design of AXI protocol in an effective manner using the System Verilog. The design is verified using the QuestaSim tool.","PeriodicalId":338691,"journal":{"name":"2022 Smart Technologies, Communication and Robotics (STCR)","volume":"12 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133871560","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 : 2022-12-10DOI: 10.1109/STCR55312.2022.10009275
Akshath Mahajan, Deap Daru, Aditya Thaker, M. Narvekar, Debajyoti Mukhopadhyay
India is prone to tropical cyclones annually, originating from the North Indian Ocean basin. Tropical cyclones are destructive and sudden natural occurrences that annually wreak havoc by taking a huge toll on human lives and property. This engenders a need for accurately forecasting the scale of such mass-destructive events, to provide us with enough time to take precautionary measures that can reduce the death toll and minimize costs. Using the CyINSAT dataset, which gives a multimodal and temporal resolution for TCs occurred from 2014 to 2022, this paper employs and compares multiple techniques to solve the wind speed forecasting issue. All models involve recurrent networks along with image feature extractors, which are used together to predict the next wind speeds from a sequence of images. The architectural differences between these models mainly focus on the nuances involved in handling the current wind speed. The proposed architecture gives higher importance to the currently recorded wind speeds and performs significantly better than the baseline models. It successfully obtained an RMSE of 6.31, MAE of 0.093 and MAPE of 4.53.
{"title":"Forecasting North Indian Ocean Tropical Cyclone Intensity","authors":"Akshath Mahajan, Deap Daru, Aditya Thaker, M. Narvekar, Debajyoti Mukhopadhyay","doi":"10.1109/STCR55312.2022.10009275","DOIUrl":"https://doi.org/10.1109/STCR55312.2022.10009275","url":null,"abstract":"India is prone to tropical cyclones annually, originating from the North Indian Ocean basin. Tropical cyclones are destructive and sudden natural occurrences that annually wreak havoc by taking a huge toll on human lives and property. This engenders a need for accurately forecasting the scale of such mass-destructive events, to provide us with enough time to take precautionary measures that can reduce the death toll and minimize costs. Using the CyINSAT dataset, which gives a multimodal and temporal resolution for TCs occurred from 2014 to 2022, this paper employs and compares multiple techniques to solve the wind speed forecasting issue. All models involve recurrent networks along with image feature extractors, which are used together to predict the next wind speeds from a sequence of images. The architectural differences between these models mainly focus on the nuances involved in handling the current wind speed. The proposed architecture gives higher importance to the currently recorded wind speeds and performs significantly better than the baseline models. It successfully obtained an RMSE of 6.31, MAE of 0.093 and MAPE of 4.53.","PeriodicalId":338691,"journal":{"name":"2022 Smart Technologies, Communication and Robotics (STCR)","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124427410","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：