Search

Wireless Sensor Network (WSN) is one of the fundamental technologies used in the Internet of Things (IoT) which is deployed for diverse applications to carry out precise real-time observations. The limited resources of WSN with massive volume of fast-flowing IoT data make the aggregation and analytics of data more challenging. Recently, data mining-based solutions have been proposed to effectively handle the data being generated by the sensors and to analyze the data patterns for deducing the required information from it. The increasing need of these techniques motivated us to propose a distributed and efficient data mining technique that not only handles the massive and rapidly generated data by the nodes, but also increases the life span of the network. In this paper, we propose a novel scheme for the IoT based WSN that mines the sensor data using association rule without moving it to any Cluster Head (CH) or Base Station (BS). The new proposed scheme enables sensors to perform computations locally and only the minimum higher-level statistical summaries of the data at Cluster Members (CMs) are exchanged with their CH. This considerably reduces the communication overhead which ultimately prolongs the network lifetime. The proposed scheme is evaluated via extensive simulations and the results obtained demonstrate that the integration of the proposed scheme in the existing protocols significantly reduces the communication overhead which ultimately prolongs the network lifetime and stability.

Pneumonia is an infectious and deadly disease. According to the World Health Organization (WHO), every third person dies due to this disease. It can be cured if detected accurately and on time. Chest X-rays are used to diagnose this disease, but it requires expert radiotherapists and a very time-consuming process. So, it is the need of the hour to develop an automatic system to detect pneumonia that could perform better and produce faster results. However, traditional handcrafted machine learning techniques show low accuracy and are expensive in terms of complexity. Deep convolutional neural networks (D-CNNs) show better performance in this regard and are simple and easy to use as compared to machine learning algorithms. In this paper, a novel algorithm based on AlexNet and SVM is proposed to detect pneumonia. We also compared the results of AlexNet with other D-CNNs to check which one is performing better. Experimental results prove that AlexNet integrated with SVM outperforms all other techniques.

Emergency Service Facilities Site Selection (ESFSS) seeks to determine the most suitable locations for additional stations in a given area. Its objective is to make the site reachable in a minimal time following a service request. Currently, potential new locations for emergency stations are manually selected. While this has been successful in the past, a more rigorous mathematical model and algorithm are needed. Choosing the optimal site for an emergency station is attractive, as it maximizes the facility’s benefits and reduces citizen service response time. This research paper conducted a systematic literature review on 51 research articles addressing four main Research Questions (RQs) about current localization approaches with regard to use of resources. In that context, current approaches had been used with different variables to select the specific area and to determine the distance between emergency response stations. This review investigates relevant studies published between 2010 and 2020 related to reputable conferences and journals. Early works determined the optimal location of emergency stations by using optimization models like maximum coverage, K-means, linear programming, genetic algorithms and a heuristic algorithm. We summarized and analyzed the existing schemes used in ESFSS. A large amount of detailed information is required in order to introducing a new facility. This information is collected and analyzed in order to conduct a proper evaluation based on specific factors. These factors such as crime levels, population and geological features may affect the station’s position. This research study provides an overview of different methods for selecting an emergency station’s position.

Wildlife plays a vital role in balancing the environment. It also provides stability to different natural processes of nature. In recent year, there are many animals which are facing the danger of extinction. The reason for animal extinction is natural occurrences such as climatic heating, cooling, or changes in sea levels. In literate, many techniques are proposed to detect and classify animals, but each technique has a limitation. In this paper, we propose a novel framework using deep convolutional neural networks (D-CNN) and k Nearest Neighbors (kNN) to detect animals. The dataset contains four class snow leopard, Marco polo sheep, Himalayan bear, and other animals. Many D-CNN like AlexNet, ResNet-50, VGG-19, and inception v3 are used to extract features. The experimental results verify that inception v3 integrated with kNN outperforms other D-CNNs. It also has more accuracy of 98.3% with a classification error of 2%, which is quite negligible.

In recent years, breast cancer is causing the death of many women. The major reasons for breast cancer are an increase in age, dense breast tissue, alcoholism, body weight, and radiation exposure. If it is detected at an early stage, then the chances of survival increase. Machine Learning and deep learning can help us to detect breast cancer at initial stage. In this paper, we propose a novel deep neural network CanNet to detect cancer with high accuracy. CanNet has accuracy of 99% with 98.8% specmcity and 99.1% sensitivity. We also compared our experimental results with other state-of-the-art methods. The comparison shows that CanNet is more efficient than all other methods.

Green technology, a technology that has motivated everyone towards itself. The reason behind it is the advancement of the modern era. Likewise, with the advancement in communication researchers have the main focus of making this communication as green as possible. In cellular technology, the increase in the number of users demands high data rates, and the evolution of 5G is the best solution to fulfill this problem. As demand increases the energy computation increases and evolution of CO2 also increases. Increment in the level of CO2 in the environment is a health hazard. In this paper, we briefly discuss the techniques and challenges of green communication. The techniques include D2D communication, mMIMO systems, heterogeneous networks (HetNets) and G-IoT. The challenges include cost, spectral efficiency, and bandwidth.

Due to the lack of centralized identity management and the broadcast nature, wireless ad hoc networks are always a palatable target for masquerading attacks. The attackers can spoof identities of privileged legitimate users for various malicious reasons, such as to launch DoS or DDoS attacks, to access unauthorized information, and to evade the detection and accountability. In the current and limited literature, masquerading attacks are mostly counteracted by signal-strength-based detection systems. However, these schemes are mostly proposed to work for infrastructure-based IEEE 802.11 wireless networks using fixed access points, air monitors, or fixed anchor nodes, which are not suitable for the ad hoc architecture. In this paper, we propose a detection system for masquerading attacks without using fixed anchor nodes or air monitors. We develop an anomaly detection model based on the statistical significant testing for our masquerading detection system that takes into consideration the signal strength fluctuation. We conduct a test bed of Samsung Galaxy-based smartphones in order to analyze the real-world signal strength variation. We also plug the real-world signal strength variation in our model for the evaluation of the detection accuracy. We propose the received signal-strength-based masquerading attack detection scheme, which is carried out first by each node in its one-hop vicinity and then extended to five-hop proximity for broader detection scope and improved accuracy. The proposed scheme is evaluated using an NS-2 network simulator for detection accuracy in different environments. The results obtained indicate that our proposed scheme produces more than 90% true positives.

Internet of Things (IoT) devices industry is rapidly growing, with an accelerated increase in the list of manufacturers offering a wide range of smart devices selected to enhance end-users’ standard of living. Security remains an after-thought in these devices resulting in vulnerabilities. While there exists a cryptographic protocol designed to solve such authentication problem, the computational complexity of cryptographic protocols and scalability problems make almost all cryptography-based authentication protocols impractical for IoT. Wireless RFF (Radio Frequency Fingerprinting) comes as a physical layer-based security authentication method that improves wireless security authentication, which is especially useful for the power and computing limited devices. As a proof-of-concept, this paper proposes a universal SDR (software defined Radio)-based inexpensive implementation intended to sense emitted wireless signals from IoT devices. Our approach is validated by extracting mobile phone signal bursts under different user-dedicated modes. The proposed setup is well adapted to accurately capture signals from different telecommunication standards. To ensure a unique identification of IoT devices, this paper also provides an optimum set of features useful to generate the device identity fingerprint.

Steganography is the practice of hiding information and is well known for concealing the existence of the communication. This technique deals with the insertion of the secret message in an appropriate carrier like writing, audio. Out of these carriers, digital writing (text file) message are mostly preferred because of their wide usage on the Internet. The LSB modification technique allows the secret information to be embedded in the least significant bit of the cover message but this technique can be broken very easily. MATLAB software has been used for the execution of steganography. This paper is used Least Significant Bit (LSB) modification technique. Two techniques are being proposed which work on the basis works on the current sample bits. These techniques spread the secret message on different bits of the sample. The methodology has also involved an AES encryption scheme as a backup in case the steganography algorithm has been broken. This will not compromise the confidentiality of the secret information and implemented Point to Point link between two pc.

Mobile Ad Hoc Networks (MANETs) are self-arranging, dynamic, infrastructure-less and multi-hop networks formed by a group(s) of mobile nodes (MNs), which transmit information via wireless channels. The broadcast nature of the wireless medium makes MANETs vulnerable to various types of attacks (e.g., masquerading attack, node replication attack, Sybil attack and so on). A critical challenge that allows replicas to defeat the well-known detection algorithms is known as abnormal or selected silence. This problem arises when replicas intentionally become silent at specific time periods or stop broadcasting detection packets. As a result, the attackers become invisible and successfully evade the detection process. Most of the existing detection techniques are vulnerable to these malicious activities. In this paper, we propose a new lightweight technique for mitigating the above issue in MANETs. Our solution is lightweight and does not need the use of extra communication or geographical positioning system (GPS). Simulation results show excellent detection accuracy of the proposed scheme.

Mobile ad hoc Networks (MANETs) are created by a collection of mobile nodes (MNs), which communicate and exchange information with each other via wireless links. These networks are widely known by their unique characteristics such as mobility, changing topology, multi-hop communication and lack of infrastructure. In the last few years, various routing protocols have been proposed for these networks. It is worthy to note that mobility brings many challenges when designing protocols for these networks. In this paper, we assessed the influence of Reference Point Group mobility model (RPGM) on the performance of both reactive and proactive protocols. Simulation results indicate that the hop-by-hop AODV performs better in high stressful environments, which makes AODV the robust protocol for RPGM model. We also observed that DSR is suitable for less tense environments. Furthermore, we found interesting facts about proactive routing protocols. These protocols are evaluated in terms of packet delivery Ratio (PDR), dropped packets, routing overhead and end-to-end delay.

The interactions between Active Queue Management (AQM) algorithms and TCP have been extensively investigated in the last few years. However, majority of the studies are conducted without considering the consequences of different attacks on TCP flows. This research work aims to investigate the performance of TCP flows under Low-Rate Denial-of-Service (LDoS) attacks. In particular, we have performed diverse LDoS experiments by varying critical experimental parameters in order to evaluate their effects on the performance of TCP under droptail and four other AQM algorithms. Simulation results indicate that an LDoS attack with short attack duration of ≈ 0.5 seconds produces more effective outcomes for the attacker as compared to the finding previously reported. Furthermore, the results reveal that droptail and PI are highly robust compared to the other three AQM schemes. RED was designed to bring major improvements over the simple droptail algorithm. However, simulation results show that the simple droptail algorithm outperforms RED in all of the three experiments. On the other hand, the Adaptive Virtual Queue (AVQ) algorithm shows the worst performance in the presence of LDoS attack.

Mobile Ad hoc Networks (MANETs) represent complex distributed systems that do not rely on any infrastructure or centralized administration. These networks consist of wireless mobile devices which communicate with each other over the wireless channel. Because each mobile node (MN) has a limited transmission range, multi-hop communication is necessary to deliver data across the entire network. Such type of communication requires routing protocols which are capable of discovering multi-hop routes between the sending and receiving nodes. In addition, various mobility models are used to mimic the walking patterns of MNs in real-world environments. These models are generally categorized into two types, i.e., (i) entity and (ii) group-based mobility models depending upon their functionalities. In this paper, we assessed the impacts of different networking contexts on the performance of different mobility models. We selected Random Waypoint (RWP) and Gauss-Markov (GM) models from the first category and Reference Point group (RPGM) and Nomadic Community (NCMM) mobility models from the second. We performed extensive simulations using Network Simulator (NS-2). Simulation results reveal that there is a decline in cost as the speed of MNs is increased. In addition, node density differently impacted the performance of both types of models. The recommendation of this research will provide better understanding of the different mobility models in MANETs and their uses in real-world environments such as search and rescue operations, battlefield situations, tracking operations, people-centric networking and many more.

In multihop networks such as mobile ad hoc networks selfish or misbehaving nodes can degrade network performance. Reputation based models have been developed to enforce cooperation and to discourage node misbehaviour. These schemes detect and isolate selfish nodes and maintain network throughput by enabling nodes to construct paths that only include good nodes. In order to punish selfish nodes and restrict their services, selfish nodes are not provided with packet forwarding services anymore by the good nodes in the network; however they can still gain throughput or utility by interacting directly with the destination nodes. The commonly adopted simulation parameters used by the research community for the evaluation of reputation based schemes in mobile environments usually produces high direct interactions. These direct interactions result in noticeably higher throughput and node utility for the network overall causing confusion: these metrics are increased or decreased due to the direct interactions or due to the reputation based scheme. We implemented the popular CONFIDANT reputation-based scheme to confirm how much throughput is the result of these direct interactions. Our simulation studies show that in the presence of 20% selfish nodes and in a continuous mobile environment, 59% of the evil throughput and 29% of the good throughput was gained through direct interactions among nodes. These direct interactions accordingly affect delay and routing overhead in the network.

In multihop networks such as mobile ad hoc networks (MANETs), a node can misbehave by selfishly dropping others' packets to save battery life. This selfishness or misbehaviour can disrupt the whole network functionality and can severely degrade network performance. Reputation based models have been developed to discourage node misbehaviour. However, these models are vulnerable to whitewashing attacks. Whitewashing occurs when a node having poor reputation/trust changes its identity to start afresh and escape from the consequences of its bad actions. This attack can significantly affect the performance of reputation based models. We propose a reputation based scheme for MANETs that acts as a deterrent for whitewashing attacks. Rather than trying to detect whitewashing attacks, we approach the problem in a novel way by removing the advantages that whitewashing can provide. In our proposed scheme, each node must pay an entry fee to consume network services. As monetary fees are not suitable for MANETs due to fee management complications, instead of a monetary fee we use a fee in the form of cooperation. A node will receive services from the network after it cooperates until its reputation is increased to a certain level Y. For a normal selfish node, it is no longer beneficial to perform a whitewash because it will be required to pay the entry fee each time it enters into the network. Simulation results show that our scheme performs well in reducing evil throughput and evil nodes' utility as compared to the CONFIDANT scheme in the presence of whitewashing nodes.

Signal Strength based position verification is more promising as compared to other Sybil attack detection techniques because it is lightweight and does not require centralisation. Nodes can locally determine their locations through received signal strength variations and inform neighbours about location updates accordingly. In position verification, network nodes verify the position of each node and further ensure that each physical location is bounded by only one identity at any particular time. In this paper we will survey Sybil attack detection techniques and their related issues. Finally we will present our ideas about the Sybil attack detection via one-time localization (only at the entrance into the network) which significantly reduces the communication overhead unlike other schemes where per packet location tracking or periodic dissemination of location information is used to detect Sybil attacks.