The establishment of future intelligent transport systems is dependable on the reliable and seamless function of Connected and Autonomous Vehicles (CAV). Reinforcement learning (RL), which allows ...autonomous vehicles (AVs) to learn an ideal driving strategy through constant contact with the environment, plays a significant part in the decision-making process of autonomous driving (AD). The networking of CAV is advantageous since it allows for the transmission of traffic-related data to vehicles via Vehicle-to-External (V2X) communication. Recognition and anticipation of driving behaviour are critical for avoiding collisions because they can provide useful information to other drivers and vehicles. The fundamental challenge in developing CAV is the construction of an autonomous controller that can effectively perform close real-time control selections, such as a fast acceleration while merging onto a highway and rapid speed adjustments in stop-and-go traffic congestion. CAV driving behaviours can be considerably improved by utilizing shared information, resulting in more accountable, intelligent, and efficient driving. In the present work, a deep reinforcement learning approach is proposed that integrates the information gathered through connectivity capabilities and sensing from neighbour automobiles in the vicinity of CAV. The fused information is used for providing safe and cooperative lane-changing behaviour. The deployment of an algorithm in CAV is expected to improve the transportation safety of CAV driving behaviours.
Dimensionality Reduction (DR) is the pre-processing step to remove redundant features, noisy and irrelevant data, in order to improve learning feature accuracy and reduce the training time. ...Dimensionality reductions techniques have been proposed and implemented by using feature selection and extraction method. Principal Component Analysis (PCA) one of the Dimensions reduction techniques which give reduced computation time for the learning process. In this paper presents most widely used feature extraction techniques such as EMD, PCA, and feature selection techniques such as correlation, LDA, forward selection have been analyzed based on high performance and accuracy. These techniques are highly applied in Deep Neural Network for medical image diagnosis and used to improve the classification accuracy. Further, we discussed how dimension reduction is made in deep learning.
Sign language is the most complex language to understand by the end-user without knowing the meaning of the sign and it depends on the special gesture motion. The gesture marks are characterized by ...hands with aid by facial appearance and body position. In this article, Gesture recognition is proposed for static sign language using Deep Learning with image processing. The involvement contains dual solutions to the problem. One is resized with Bicubic static American Sign Language binary images. Besides that, good recognition results in of detection the borderline hand using the various edge detection techniques. Another solution is to classify the alpha characters of sign language using a Convolution Neural Network (CNN). Owing to the anticipated method, test accuracy of 96% and F1-Score of 99% have achieved for 10 different hand gesture classifications.
Nowadays to deliver any enterprise application into the market the functionality of the application is tested over the automated environment. The application is developed using different software ...development life cycle approaches like waterfall and agile. The traditional development tool needs a lot of manual intervention and manpower which is a time-consuming process. In the case of agile, the trail deployment build happens once which is difficult to fix and rectify the bugs. To improve the quality and the speed of development DevOps came into a picture that can build and deploy the application at stages. To do the continuous integration and delivery the automation tools like Jenkins, bamboo, and Travis CI are available in the market. The proposed approach utilizes the azure cloud DevOps along with the feature of Jenkins for continuous integration and deployment. The application is built and can be auto-configured in a cloud platform which greatly improves the build and deployment speed. The parallel and continuous integration adopts the dynamic changes effectively using the automation tools.
GPS technology is the major component of fisherman for location identity and communication among the boats. Communicating with the other boats is possible only through intermediary access points. ...Identifying the location of the fisherman or boat during the rescue operation is still challenging. Sometimes the fisherman loses their lives because of the delay in finding the exact location whenever the GPS gets failure. The proposed approach makes use of the Identity Spreader System (ISS) for buffer location among the neighboring peer node to locate the identity. The concept of an ad-hoc cluster connects the group of nodes for easy access management. The routing table contains the identity of the neighboring node all the time for current communication and identity management. The node that is registered and left the cluster will send the identity along with the position periodically to the root access point. If the root access point already contains the identity with the time of catch is recent will be updated. The proposed approach is experimented with and simulated in NS2.35.
InceptionResNetV2 for Plant Leaf Disease Classification Naveenkumar, M; Srithar, S; Rajesh Kumar, B ...
2021 Fifth International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC),
2021-Nov.-11
Conference Proceeding
Diseases that assault the leaves of a plant are fundamentally spread by twists and turns. However, this can also happen for plants if there is shortage or excess of water from the downpour or water ...system and over pesticides. Because of different infection, diseases in vegetable plant life should be anticipated and cured early to reduce the harvest misfortune. A large portion of the sicknesses assaulted in the plants are found in their stems. By the advancement in deep learning techniques, infection order utilizing leaf picture is feasible. For this experiment, potato plant leaves are used. For image preprocessing, a public dataset is utilized likewise the reports are enlarged into products of pictures, and for the testing ongoing information was gathered from the farming field. Framework is planned and tried with various kinds of Convolutional Neural Networks to be like InceptionV3, ResNet and InceptionResNetV2. The general presentation of the framework is anticipated to be 95% to the most extreme and our analysis ends up with the same.
The deaf and dumb peoples mostly communicate with normal people through the standard nonverbal American gesture. But the sign languages are difficult to recognize and understandable by the common ...people. The existing computerized sign recognition methodologies suffer from the sign object classification. The false-positive test result reduces the prediction accuracy or increases the training overhead. The proposed approach named Improved Haar Feature (IHF) locate the skeleton of the hand pose and recognize the sign by the Angle Support Vector (ASV). The skeleton is estimated by the series of cyclic connected points of the box model. The object's centroid is estimated, and the cut-vertex is defined through the support vector. The number of vertices and the distance between the point-pin are also considered for region mapping. The number of segregation classes will differ based on the nature of the sign skeleton. The angle of the centroid and the connection points are examined to extract the right feature. The custom gesture is trained using the deep convolutional neural network models. The proposed methodology will observe the hand sign and produce an optimal output in the form of verbal speech or text. The algorithm produces a good prediction accuracy of 96.2 percent.
PDF
