There are web mappings service applications like Google Maps, Apple Maps, and Waze that offer satellite imagery, street maps, real-time traffic conditions, departure arrival timings, and route planning for traveling b...
There are web mappings service applications like Google Maps, Apple Maps, and Waze that offer satellite imagery, street maps, real-time traffic conditions, departure arrival timings, and route planning for traveling by foot, car, bike, and public transportation. These applications enable users to view and hear traffic alerts, turn instructions, lane guidance, and available alternate route options. Road signs like speed breakers, sharp turns, accident-prone zone, steep slopes, etc. are placed beside or above streets to assist the drivers. These applications usually run on the display screen of the car or the commuter's mobile. The driver must keep an eye on the screen as well as on the road signs. At times the driver might miss the road sign due to high speed leading to a tense situation of varying severity. To overcome this problem, we propose an application in which the moving car is connected to a nearby Fog Node that stores road signs along with their proper location. Fog computing minimizes the amount of data transferred to the cloud, lowering bandwidth consumption and associated expenses.. Because the initial data processing occurs near the device, latency is reduced, and overall response time is improved. As the car moves closer to a road sign, The Fog Node evaluates the distance between the road sign and the car's geolocation. If the car is approaching a road sign, the Fog Node will generate an alert voice message. Fog Node being closer to the car will have minimum latency. Using Fog Computing we can send an early voice alert to a user within a certain time before he reaches the location. The solution proposed will provide timely alerts to users before reaching the road segment that requires the attention of the driver.
Predicting intrusion detection in a wireless sensor network (WSN) involves using machine learning algorithms to analyze network traffic data and predict whether an intrusion or security breach will occur. The goal of ...
Predicting intrusion detection in a wireless sensor network (WSN) involves using machine learning algorithms to analyze network traffic data and predict whether an intrusion or security breach will occur. The goal of intrusion prediction is to provide advanced warning of potential attacks and allow for proactive measures to be taken to prevent or mitigate the attack's impact. Several machines and deep learning algorithms can be used for intrusion prediction in WSNs. This research describes a practical deep-learning approach for Intrusion Detection in WSN based on artificial neural networks. Spyder IDE of python with version 3.7 is used for the simulation work. The results of the simulations demonstrate that the anticipated model is accurate 99.99% of the time.
This paper provides a practical view of implementing a real-time vision-based fire detection system that can accurately detect and alert the presence of fires in live CCTV footage. The proposed system combines motion-...
This paper provides a practical view of implementing a real-time vision-based fire detection system that can accurately detect and alert the presence of fires in live CCTV footage. The proposed system combines motion-flicker detection, which uses a background subtraction algorithm to extract the moving objects in a video and a flicker detection algorithm to verify flickering in suspected regions. Then using a YCbCr colour model, the colour of the flickering region is verified to match the characteristics of fire. Further, a YOLOv7 model trained on a large dataset of fire images is employed to verify the results and raises the alarm to alert the presence of fire. Several optimizations have been done to reduce false alarm rate and GPU memory usage times. The accuracy is 98.4%, and false alarm rate is as low as 2.1%.
One of the major natural disasters that take place on earth is landslide, which can happen naturally or as a result of human activity. In mountainous and hilly terrain, landslides, which can be fatal, are more frequen...
详细信息
One of the major natural disasters that take place on earth is landslide, which can happen naturally or as a result of human activity. In mountainous and hilly terrain, landslides, which can be fatal, are more frequent. To prevent the loss of life and minimize economic loss, the two main repercussions of the deadly calamity, susceptibility assessments and the identification of probable landslide zones become essential. The current study aims to provide a comprehensive performance assessment of a multivariate statistical technique and a bagging ensemble for landslide susceptibility mapping along the NH-10, which has an approximate expansion of 51.5 kms from Sivok Khola to Rangpoo. In total, 309 landslide sites were identified and divided randomly into training and testing data with 70:30 ratio. The prepared susceptibility maps were validated and substantiated through the receiver operating characteristics curve, which showed bagging ensemble performed better.
The Internet of Things (IoT) is a significant technology that enables us to develop many helpful online applications. The Internet of Things allows remote operating items using preexisting network infrastructure. Serv...
The Internet of Things (IoT) is a significant technology that enables us to develop many helpful online applications. The Internet of Things allows remote operating items using preexisting network infrastructure. Services that are based on information technology, computers, and other technologies are included in the cyber world. The cyber-world sees many protocols and technological advancements that lead to improvements. When it comes to cyber-based services, security is an extremely significant issue. Classification methods based on machine learning (ML) and deep learning (DL) can accurately forecast the most effective model for a dataset. A cyber assault on a wide scale carried out by devices infected with malware and managed remotely is known as a botnet attack. Under the context of IoT applications, an effective method for detecting botnet attacks is presented in this research.
The increasing electricity demand, combined with the continuous depletion of fossil fuel reserves, has intensified the search for alternative renewable energy sources. To bridge the gap between energy demand and suppl...
The increasing electricity demand, combined with the continuous depletion of fossil fuel reserves, has intensified the search for alternative renewable energy sources. To bridge the gap between energy demand and supply, the development of an affordable electrical infrastructure is imperative. Such an infrastructure should encourage energy conservation and discourage using energy-intensive devices during peak hours. Achieving these goals relies heavily on the accuracy and time complexity of algorithms used for electricity consumption prediction. Additionally, factors such as weather conditions and static characteristics must be taken into account, as they significantly influence user acceptance and technology adoption. Recent research has shown that integrating energy-saving knowledge and environmental awareness into the second generation of the UTAUT model can positively impact the acceptance of smart meters among residential consumers. Moreover, the implementation of Demand Response (DR) programs can effectively reduce system peaks and generate billions of dollars in annual savings. Notably, studies have indicated that a small percentage of customers (20%) account for a substantial majority (80%) of price responses, emphasizing the potential impact of a targeted approach. Researchers have demonstrated the effectiveness of such strategies by training a model on data from over 5567 households and validating it with another algorithm. The implementation of a smart electrical grid holds promise in meeting the increasing demand for electricity, promoting energy conservation, and mitigating peak-hour energy usage. The success of this endeavor relies on the deployment of accurate and efficient algorithms, the widespread acceptance of smart technologies, and the utilization of big data analytics. By embracing these factors, we can pave the way toward a sustainable and efficient energy future.
One of the important areas of research that needs a global attention is the agribusiness and its network of supply chains. Internet of Things (IoT) based smart models of this system facilitates all the participants in...
One of the important areas of research that needs a global attention is the agribusiness and its network of supply chains. Internet of Things (IoT) based smart models of this system facilitates all the participants in several means to make the life easy. Blockchain can play an inevitable role to enhance the security, to protect the end users from the potential health risks, and to avoid the financial frauds. Therefore, some blockchain based food supply systems have been developed to provide the transparency and immutability. However, the security of such existing systems can be optimized with improved integrity, confidentiality, latency, authentication delay, throughput, and computational time. In this direction, a blockchain and IoT enabled secure and smart agricultural food supply chain system has been devised and presented in this paper. The suggested model performs better than the current plans in terms of trust, authentication delay, computational time, and throughput.
When training convolutional neural networks, noise distribution in data can affect segmentation accuracy. Therefore, many neural networks will strengthen the segmentation ability of the model by artificially adding no...
When training convolutional neural networks, noise distribution in data can affect segmentation accuracy. Therefore, many neural networks will strengthen the segmentation ability of the model by artificially adding noise. However, because the noise distribution of the input data is not known, the model training process can only generate random noise, e.g., gaussian noise, but cannot simulate real-world noise. In the medical imaging domain, more attention is paid to the image contours, which can also be used for noising simulation. Meanwhile, considering the existence of rich unlabeled data sets in medical images, this paper proposes a semi-supervised framework for uncertainty estimation by using Monte Carlo Dropout and generates adaptive noise using the uncertainty map. Specifically, the framework consists of three parts, including a noise generation model, a teacher model, and a student model. The noise generation part processes the data input to the other two parts and uses the contour feature identified by the posteriori uncertainty map to simulate priori noise to guide the process of training segmentation model. Meanwhile, the student model updates its parameters based on the teacher model output and the ground truth annotations by minimizing a joint loss. Segmentation experiments on left atrium 3D data show that our method has competitive results with other SOTA semi-supervised methods.
The number of vehicles being used for daily commute within the cities have increased exponentially since 2015. This has caused a major concern regarding parking spaces especially in a rapidly developing cities. More t...
The number of vehicles being used for daily commute within the cities have increased exponentially since 2015. This has caused a major concern regarding parking spaces especially in a rapidly developing cities. More than 25% of vehicles driving around the cities are looking for a parking space. Numerous issues arise as a result of poor parking infrastructure such as significant wastage of time for drivers, traffic congestion and increase in car emissions. Implementing smart technology to facilitate this task will solve this problem, improving operational efficiency, simplifying the flow of urban traffic and offering drivers a more enjoyable and time-saving experience. The proposed IoT-based smart parking system allows real-time data about parking availability to be obtained via a mobile application. An IoT gadget, which includes sensors and microcontrollers is placed in each parking space. The data from the gadget will be pushed to a cloud-based system which analyses this data to calculate the availability of free slots in parking spaces. The users receive real-time information on the availability of all parking spaces giving them the choice to choose the best one.
An Intrusion Detection System (IDS) is often used to keep safe and efficiently utilise communication network. And generally this is achieved by monitoring the network regularly for possible intrusions. Nowadays in net...
详细信息
An Intrusion Detection System (IDS) is often used to keep safe and efficiently utilise communication network. And generally this is achieved by monitoring the network regularly for possible intrusions. Nowadays in networks, zero-day attacks are common type of intrusions. So it became necessary to use IDS, which is capable of detecting zero-day attacks. Machine Learning (ML) algorithms are suitable in designing such IDS. Deep Learning (DL), being part of ML, have promising approaches that can be used in designing IDS, which is capable of detecting intrusions of new and old types. This paper proposes DL based IDS which create model of normal network traffic. This model classifies attacks based on the various features present in the dataset. The proposed model uses a Convolution Neural Network, bi-directional Long Short Term Memory (LSTM), and a stack of encoders to handle spatial and temporal features more effectively. The model achieves a low False Positive Rate and high Detection Rate compared to existing models.
暂无评论