Abstract Recent studies have suggested ways to enhance user perception of remote workspaces and robot autonomy, often at the expense of system suitability and efficiency. This research introduces a novel method, Distributed Supervisory Control (DSC), leveraging Virtual Reality (VR) to enhance teleoperation. The DSC method intelligently distributes tasks between the robot and the human operator, minimizing shared autonomy conflicts and sensory data transfer. It uses Deep Reinforcement Learning (DRL) through a Twin Delayed Deep Deterministic Policy Gradient (TD3) method for tasks like obstacle avoidance. Real-time experimentation validated the method’s effectiveness through performance metrics, including NASA-TLX, task execution time, and obstacle collision frequency. Kruskal-Wallis tests identified significant differences in task execution times and collision frequency across DSC, Direct Control (DC), and Assistive Direct Control (ADC). Dunn’s post-hoc tests indicated DSC significantly outperformed DC and ADC in both execution time and collision frequency. Similarly, NASA-TLX scores for effort, mental demand, performance, frustration, physical demand, and temporal demand also showed significant differences ( p < 0.05), supporting DSC’s lower task load. User case studies revealed enhanced user experience, as measured by the System Usability Scale (SUS) and Igroup Presence (IPQ) questionnaires for immersive experience. The non-parametric Kruskal-Wallis test was utilized to verify significant differences between group medians, followed by a Dunn’s test that revealed significant performance improvements with our Distributed Supervisory Control (DSC) method relative to Direct Control (DC) and Assistive Direct Control (ADC), thereby enhancing task efficiency and overall interface suitability. In conclusion, Distributed Supervisory Control significantly enhances task efficiency and user experience in robot teleoperation environments, demonstrating its potential as a new standard for remote operations.

Abstract Several countries are most reliant on agriculture either in terms of employment opportunities, national income, availability of a raw material, food production, to name but a few. However, it faces a big challenge such as climate changes, diseases, pets, weeds etc. Therefore, last decade has provided a machine learning‐based solution to the agricultural community, which helped farmers to identify the diseases at the early stages. In this article, our focus is on grape diseases, and proposes a novel framework to identify and classify the selected diseases at the early stages. A deep learning‐based solution is embedded into a conventional architecture for optimal performance. Three primary steps are involved; (a) feature extraction after applying transfer learning on pre‐trained deep models, AlexNet and ResNet101, (b) selection of best features using proposed Yager Entropy along with Kurtosis (YEaK) technique, (c) fusion of strong features using proposed parallel approach and later subject to classification step using least squared support vector machine (LS‐SVM). The simulations are performed on infected grape leaves obtained from the plant village dataset to achieving an accuracy of 99%. From the simulation results, we sincerely believe that our proposed approach performed exceptionally compared to several existing methods.

Abstract Effective recognition of fruit leaf diseases has a substantial impact on agro‐based economies. Several fruit diseases exist that badly impact the yield and quality of fruits. A naked‐eye inspection of an infected region is a difficult and tedious process; therefore, it is required to have an automated system for accurate recognition of the disease. It is widely understood that low contrast images affect identification and classification accuracy. Here a parallel framework for real‐time apple leaf disease identification and classification is proposed. Initially, a hybrid contrast stretching method to increase the visual impact of an image is proposed and then the MASK RCNN is configured to detect the infected regions. In parallel, the enhanced images are utilized for training a pre‐trained CNN model for features extraction. The Kapur's entropy along MSVM (EaMSVM) approach‐based selection method is developed to select strong features for the final classification. The Plant Village dataset is employed for the experimental process and achieve the best accuracy of 96.6% on the ensemble subspace discriminant analysis (ESDA) classifier. A comparison with the previous techniques illustrates the superiority of the proposed framework.

Data and knowledge management systems employ feature selection algorithms for removing irrelevant, redundant, and noisy information from the data. There are two well-known approaches to feature selection, feature ranking (FR) and feature subset selection (FSS). In this paper, we propose a new FR algorithm, termed as class-dependent density-based feature elimination (CDFE), for binary data sets. Our theoretical analysis shows that CDFE computes the weights, used for feature ranking, more efficiently as compared to the mutual information measure. Effectively, rankings obtained from both the two criteria approximate each other. CDFE uses a filtrapper approach to select a final subset. For data sets having hundreds of thousands of features, feature selection with FR algorithms is simple and computationally efficient but redundant information may not be removed. On the other hand, FSS algorithms analyze the data for redundancies but may become computationally impractical on high-dimensional data sets. We address these problems by combining FR and FSS methods in the form of a two-stage feature selection algorithm. When introduced as a preprocessing step to the FSS algorithms, CDFE not only presents them with a feature subset, good in terms of classification, but also relieves them from heavy computations. Two FSS algorithms are employed in the second stage to test the two-stage feature selection idea. We carry out experiments with two different classifiers (naive Bayes' and kernel ridge regression) on three different real-life data sets (NOVA, HIVA, and GINA) of the”Agnostic Learning versus Prior Knowledge” challenge. As a stand-alone method, CDFE shows up to about 92 percent reduction in the feature set size. When combined with the FSS algorithms in two-stages, CDFE significantly improves their classification accuracy and exhibits up to 97 percent reduction in the feature set size. We also compared CDFE against the winning entries of the challenge and found that it outperforms the best results on NOVA and HIVA while obtaining a third position in case of GINA.