Convolution Neural Network (CNN) is one of the powerful deep learning tools used in many computer vision tasks; however, it is still in a premature state while dealing with sensor data due to the unavailability of a large data set. Here in this paper, we present a CNN-based fall detection system. Deep convolution generative adversarial network (DCGAN) is used to synthesize more fall data before analysis. Sensor data along with synthesized data is prepossessed, and time-varying spectrograms are derived. We apply CNN on raw spectrogram images of different activities to categorize a fall from other activities. As a result, we successfully attain a classification accuracy of 97.2%, surpassing the conventional machine learning methods and previous works on the same data by a large margin.

The concept of granulized Z-numbers improves information utilization and manages the degree of uncertainty in decision-making. In this paper, a novel scoring method, namely, ordered weighted averaging-based expected granulized Z-number, a new aggregation operator, named, granulized Z-number-based ordered weighted averaging operator, and a novel fuzzy risk assessment scheme is proposed. The proposed scoring method is used to order the granulized Z-numbers and takes care of the possibilistic as well as probabilistic information contained in the granulized Z-numbers. The maximum entropy principle-based nonlinear optimization model is formulated to capture the aforesaid probabilistic information during the scoring process. Based on the proposed scoring and ordering method, the granulized Z-number-based ordered weighted averaging operator is developed, which judiciously integrates the benefits of the granulized Z-numbers and the ordered weighted averaging operator to provide an improved aggregation of the decision-making information collected from multiple sources (experts). The required properties are also proved. Finally, the novel fuzzy risk assessment scheme is developed using the granulized Z-number-based ordered weighted averaging operator, the average linkage-based ordered weighted averaged similarity measure between two granulized Z-numbers, and the basic operations of logical gates of a fault tree. This scheme provides the system-level failure probability in an easy-to-understand form with reliability. A case study is also presented to demonstrate the usability and feasibility of the proposed models and schemes.

In this work, we propose a dynamic decision query mapping mechanism, DQ-Map, for provisioning Safety-as-a-Service (Safe-aaS) (Royet al., 2018). A Safe-aaS infrastructure provides customized safety-related decisions simultaneously to multiple end-users. We consider road transportation as the application scenario of Safe-aaS and termed the safety-related decision to be delivered to the end-users as decision queries (DQs). These DQs are generated according to the decision parameters selected by the end-users. The primary aim of our proposed work is to reduce the total number of sensor nodes required to generate safety-related decisions, which minimizes both energy and time consumption. Further, the requested DQs are processed and a decision is generated in three different stages. First, the DQs are categorized asemergency decision query(EDQ) andnonemergency decision query(NEDQ), depending upon the type of vehicle from where the end-users have requested safety services. The EDQs and NEDQs are mapped with the stored decisions present in the database of the decision virtualization layer during the second level. In case of mismatch with the stored decisions in the database, EDQs are directly executed from the sensor nodes deployed at a particular geographical location or into the vehicles, in the device layer of the Safe-aaS infrastructure. In the third level, the similarity score of NEDQs, which do not match with the parameters of the stored decisions, is computed. Based on the number of similar decision parameters present in them, the similarity score is computed. Extensive simulation results of the proposed scheme, DQ-Map, depict that the amount of energy consumed and time required to generate a decision is reduced by 55.16% and 54.55%, respectively, compared to the traditional Safe-aaS architecture.

Abstract Classifying or predicting occupational incidents using both structured and unstructured (text) data are an unexplored area of research. Unstructured texts, i.e., incident narratives are often unutilized or underutilized. Besides the explicit information, there exist a large amount of hidden information present in a dataset, which cannot be explored by the traditional machine learning (ML) algorithms. There is a scarcity of studies that reveal the use of deep neural networks (DNNs) in the domain of incident prediction, and its parameter optimization for achieving better prediction power. To address these issues, initially, key terms are extracted from the unstructured texts using LDA-based topic modeling. Then, these key terms are added with the predictor categories to form the feature vector, which is further processed for noise reduction and fed to the adaptive moment estimation (ADAM)-based DNN (i.e., ADNN) for classification, as ADAM is superior to GD, SGD, and RMSProp. To evaluate the effectiveness of our proposed method, a comparative study has been conducted using some state-of-the-arts on five benchmark datasets. Moreover, a case study of an integrated steel plant in India has been demonstrated for the validation of the proposed model. Experimental results reveal that ADNN produces superior performance than others in terms of accuracy. Therefore, the present study offers a robust methodological guide that enables us to handle the issues of unstructured data and hidden information for developing a predictive model.

In this article, we have developed an improved failure mode and effect analysis (FMEA) model by leveraging the concepts of Z-number, rough number (RN), and probabilistic distance measure. Two new concepts, namely, double upper approximated rough number (DUARN) and granulized Z-number (gZN), a new scheme for measuring distance between two gZNs using weighted similarity and average linkage method, a new risk prioritization model, named, granulized Z-VIKOR, and a scheme for uncertainty assessment using box-plot are proposed. DUARN embodies the notion of double sided upper approximation of an ordinal decision class. gZN is developed using Z-number and DUARN. The distance between two gZNs is computed using the maximum entropy principle that captures the relationship between the A (opinion) and B (reliability of A) parts of gZN. The granulized Z-VIKOR involves synergistic integration of gZN and VIKOR. Both objective and subjective risk measures are computed and a combined risk measure is defined, which considers the interactions among the risk criteria using λ-Shapley index. Two case studies are conducted. Sensitivity and comparative analysis is carried out to demonstrate the applicability, effectiveness, and robustness of the proposed model, as well as its superiority to existing models.

Existing traffic video summarization algorithms are capable of detecting one-class (i.e., collision) anomaly and cannot handle uncertainty issues arising between two-class anomalies, such as collision and near-miss. To address the issues, a new video summarization algorithm, namely Z-number s-based spatio-temporal rough fuzzy granulation (Z-STRFG) is developed. In Z-STRFG, various spatio-temporal features are computed over the video frames and used for obtaining the approximate anomaly-prone regions in terms of granules. In these regions, uncertainty (i.e., fuzziness) may arise among three scenarios, namely collision, near-miss, and normal traffic. Therefore, two types of rough fuzzy granules (RFGs) along with their roughness scores are computed to distinguish the aforesaid three scenarios. For each RFG, Z-number is computed based on the membership value of its roughness score to ensure a higher degree of reliability in the detection of anomaly class. Aforesaid characteristics of Z-STRFG improve its speed and accuracy for traffic anomaly detection. The efficacy of Z-STRFG has been demonstrated over 130 real-time traffic videos containing collisions, near-misses, and normal traffics. The superiority of Z-STRFG over some state-of-the-art is also proved through extensive experiments.

In this article, we introduce anenergy-efficient decision deliverymechanism,Safe-Serv, in the Safety-as-a-Service (Safe-aaS) infrastructure for the road transportation industry. A Safe-aaS architecture provides safety-related customized dynamic decisions to the registered end-users. Moreover, the concept ofdecision virtualizationenables to deliver the same decision to multiple end-users at the same time. The sensor nodes sense and transmit the data to the edge node/cloud, which is further processed to generate a decision. As the sensor nodes are energy-constrained in nature, energy efficiency is one of the important parameters to be considered for Safe-aaS infrastructure. Safe-Serv reduces energy consumption through the elimination of redundant data transmission from the sensor node to the edge node or cloud. We use the cooperative Nash bargaining approach among different homogeneous sensor nodes, which bargain among themselves to transmit data to the edge node/cloud. Based on the total dissipated energy, effective proportional distance, duty factor, nodal delay, and cost-efficient state, the appropriate sensor node is chosen. Thus, the selected sensor node transmits data to the edge layer or cloud. We incorporate the cost of data transmitted by the sensor node, which leads to cost-effective utilization of the resources. Through extensive simulation, we observe that the energy dissipated by the sensor nodes using the proposed scheme, Safe-Serv, is reduced by 85 and 78 percent approximately compared to the existing schemes – SASPENCE and manoeuvre-based trajectory planning – respectively.

In this article, we propose a new weighted Z-similarity measure between two Z-numbers, which is able to retain the original information provided by experts in linguistic terms. This measure takes care of the directional aspect of reliability of information by leveraging the Hausdorff distance. Probabilistic-approach-based statistical distance measure is used to characterize the internal relationship between two parts of a Z-number. Some properties of the Z-similarity measure are proved. The measure is capable of quantifying the probability of basic events in bow-tie analysis. It is experimentally shown that the Z-similarity measure is able to overcome the lacuna of the available techniques. Sensitivity analysis is done to verify its feasibility and applicability. We further propose another measure, called Z-similarity-based basic event contribution (Z-BEC), to quantify the contribution of basic events to the occurrence of different accidents. The performance of the Z-BEC measure is compared with that of Fussell-Vesely index and Birnbaum's structural index.

In this paper, we propose a dynamic edge node selection scheme, named as DENSE, for the Safety- as-a-Service (Safe-aaS) architecture [1]. A Safe- aaS infrastructure provisions customized safety- related decisions remotely to the registered end- users. Depending on the time-criticality of data, the static and mobile sensor nodes sense and transmit data to the edge nodes. The number of edge nodes present within the proximity of a mobile sensor node vary with the change in the locations of the vehicle. Moreover, the distance between the mobile sensor node and the edge nodes, within its proximity, change with the variation in the vehicle's location. Therefore, in such a situation, dynamic selection of the appropriate edge node for processing the time-critical data is necessary. To optimally select the edge node, we use cooperative coalition-based game theoretic approach. Further, we apply Karush-Kuhn-Tucker (KKT) conditions to find the existence of equilibrium. The analytical results of our proposed scheme, DENSE, shows that the average utility increases by 11.33% with respect to the available storage space of the edge nodes. Moreover, the average utility increases by 50.43% with respect to the average number of tasks executed per unit time by the edge node.

In this paper, we present solution for the development of a novel infrastructure, safety-as-a-service (Safe-aaS) for the road transportation industry. Safe-aaS provides safety related decisions to the registered end-users. The safety decisions are customized as per the end-user types and their requirements. Existing related research work on road safety focus on the development of the safety systems, which are able to assist the driver of the vehicle. However, none of the works serves as a common platform for providing customized decisions dynamically as per user requirements. As per our knowledge, Safe-aaS is one of the first attempts in its domain, where multiple end-users receive safety related decision dynamically. An end-user enjoys the pay-per-use service of Safe-aaS, without concerning about the back-end process. Safe-aaS is based on service oriented architecture, where different business entities such as vehicle owners, sensor owners, safety service provider, and end-users are involved. We introduce the term, decision virtualization, which enables multiple end-users to access the customized decisions remotely. We present possible cost analysis for the entities involved in the system. Analytical results show the cost and profit analysis of the different entities. We observe the profit gain by mobile sensor owner is 19.69% more as compared to static sensor owner. In the presence of 5, 10, and 15 end-users, payable rent varies between 15%-20%. Additionally, we present two case studies to depict a clear view of usage of Safe-aaS.