The applications and protocols conceived for mobile ad hoc networks rely on the assumption of cooperation amongst the mobile nodes because of lacking infrastructure. All nodes have to spend their precious resources (e.g. battery power, memory, computational power, and network bandwidth) for routing and packet forwarding operations for other nodes, in a cooperative way in the network. However, there are some nodes that may intentionally turn themselves to behave selfishly in order to conserve their valuable resources. The selfish behaviour of such nodes drastically reduces the desired degree of cooperation amongst the mobile nodes. Over the course of time, the non-cooperative activities of, such selfish nodes would paralyze the normal functioning of the whole network. Therefore, these types of nodes should be detected and isolated from the network, as soon as they begin to exhibit their selfish behaviour. In this paper, a dynamic trust based intrusion detection technique is presented to detect and isolate the selfish nodes from the network, where the direct trust degree based on direct communication interactions and indirect (recommended) trust degree based on the neighbours’ recommendations are taking into account to accurately judge the selfishness nature of the nodes. The results obtained throughout the simulation experiments clearly show the feasibility and effectiveness of the proposed intrusion detection technique.

Trust Based Intrusion Detection Technique to Detect Selfish Nodes in Mobile Ad Hoc Networks

In recent years Delay/Disruption Tolerant Network (DTN) has been identified as the most promising and challenging areas of wireless communication where sparseness and delay are particularly high. It is emerging as a promising network in the areas of disaster response, inter-planetary, vehicular, military/tactical, underwater, satellite networks. DTN is characterized by the large communication latency and lack of end to end path from source to destination. Due to intermittent connectivity in DTN, data forwarding is possible using multiple hops in "store-carry-forward" manner. So, the entire data communication is solely dependent on the cooperation of the intermediate nodes or hops. However, these intermediate nodes sometimes may become selfish due to preserve its limited resources or due to some malicious intention [1][2]. In order to increase data delivery, the detection and avoidance of such selfish nodes are foremost required. Here we are using reputation based strategy where reputation can be used as a measure to detect the cooperation pattern of a node with other nodes in the network during data communication. In this paper, we propose a dynamic two hops reputation technique for detection and avoidance of selfish nodes where neighbor nodes will monitor the behavior of a particular node as forwarder and receiver. These one hop and two hops neighbor nodes will give their feedback after data communication takes place at certain interval. However, the universal reputation matrix will be published time to time based on the cooperation pattern of nodes by the Trusted Authority (TA). The performance of the proposed scheme was evaluated on ONE simulator [3].

A dynamic two hops reputation assignment scheme for selfish node detection and avoidance in delay tolerant network

Scheduling plays an important role in providing quality of service (QoS) support to multimedia communications in various kinds of wireless networks, including cellular networks, mobile ad hoc networks, and wireless sensor networks. The authors propose a scheduling algorithm at the medium access control (MAC) layer for multiple connections with diverse QoS requirements, where each connection employs adaptive modulation and coding (AMC) scheme at the physical (PHY) layer over wireless fading channels. Each connection is assigned a priority, which is updated dynamically based on its channel and service status; the connection with the highest priority is scheduled each time. The authors' scheduler provides diverse QoS guarantees, uses the wireless bandwidth efficiently, and enjoys flexibility, scalability, and low implementation complexity. Its performance is evaluated via simulations.

A cross-layer scheduling algorithm with QoS support in wireless networks

An ad hoc network is a suburbanized types of wireless network. The ad hoc which means “for this purpose” formed network instantly when nodes are connected in network for communication. Ad hoc network is a temporary network built for specific purpose. It does not require any pre-established infrastructure. The dynamic nature of ad hoc networks makes them more utilisable. On the other hand, selfish nodes in the network may degrade the network performance in an attempt to save their resources. The side effects of such selfish nodes are lower data accessibility, break-down in communication, and increase in communication cost for legitimate nodes and so on. In past years, various solutions for extenuating selfishness are proposed by many researchers. In this paper, we survey notable work done by various researchers to detect selfish nodes in the network. Moreover, we also mention future scope of these existing schemes which may be beneficial to researchers for further research in this direction.

Various schemes to detect selfishness in wireless ad-hoc networks: A survey

Due to the limited transmission capability of mobile nodes in mobile ad hoc networks (MANETs), the intermediate nodes are used in multi-hop fashion for forwarding the packets for other nodes. However, the multi-hop communication causes a serious problem that a node may act as selfish by refraining from forwarding the packets for other nodes in order to save its energy and computing resources. A selfish node attempts to utilize the network resources for its own benefits, but reluctant to spend its resources for others. Therefore, it is really important to scrutinize and isolate the selfish nodes from the network in order to trim down the risks from such nodes, and improve the security of the network. This paper presents a comparative study of most prominent selfish node detection and mitigation techniques proposed by the researchers in the literature.

A detailed survey on selfish node detection techniques for mobile ad hoc networks

A mobile ad hoc network (MANET) is a collection of wireless mobile nodes that form a dynamic network without the need for any infrastructure. Due to the dynamic nature of MANET, it is prone to different kinds of malicious attacks. In order to pursue secure communication in such networks, there are many research solutions proposed for detecting and avoiding such malicious activities. As we know, MANET works properly if participating nodes cooperate in routing and forwarding. However, a node may decide not to cooperate just to save its resources but still use network to relay its traffic. In this scenario, we propose a reputation-based strategy to detect non-cooperative or selfish nodes and to select proper forwarder node for improving overall packet delivery of the network. Moreover, our proposed solution ensures data integrity. We have also done survey and performance analysis on some existing malicious attacks detection and prevention techniques available. The entire simulation has been done using the Network Simulator (NS-2) (http://www.isi.edu/nsnam/ns/) and simulation results show better delivery compared to some of other existing techniques discussed in the papers by Khamayseh et al. (J Netw 7(1):116–125, 2012), Marti et al. (Proceedings of International Conference on Mobile Computing and Networking (MOBICOM’OO), 255–265, 2000), Woungang et al. (Proceedings of IEEE Conference, 2012) and Hu et al. (IEEE J Sel Areas Commun 24(2):307–380, 2006).

A Novel Approach for Non-cooperative Node Detection and Avoidance Using Reputation-Based Scheme in Mobile Ad hoc Network

Future generation of wireless mobile communication aims to provide high speed Internet access along with support for high mobility to the end users. Mobile WiMAX (IEEE 802.16e) has been developed in this direction in 2005 which provides broadband wireless Internet access up to 70Mbit/s. Its high speed enables various multimedia applications along with the conventional telephony service. These applications are categorized into classes like Platinum, Gold, Silver and Bronze according to their priority. In this paper we investigate the behavior of different scheduling algorithms for OFDMA/MIMO based WiMAX (IEEE 802.16m) architecture and analyze the quality of service (QoS) issues like throughput, end to end delay, interference, signal to noise ratio, call blocking probability and call dropping probability. For simulation, we used OPNET Modeler. The simulation results show high values of the QoS parameters although there is lack of consistency in some of those values that calls for further improvement in design, architecture and deployment strategy.

Performance analysis of service classes for IEEE 802.16m QoS optimization

Malaria is caused when a person is bitten by a female Anopheles mosquito. It is a fatal disease. Lots of work has been done in the detection of this disease all over the world. Every year millions of people all over the world are affected by malaria. It is a contagious disease. Traditionally, malaria is diagnosed by distinguishing healthy red blood cells from the infected ones. It is also very time consuming and it may produce inaccurate reports due to human error. Compared with the traditional methods used to extract the engineering feature of the hand, the proposed method uses an in-depth study of end-to-end design that makes both the extraction features and direct separation from the dots separated by red blood bands. The data, used in this study, were taken from the National Institute of Health called the NIH malaria dataset. A pre-trained model and a custom model have been compared here to analyze the better model. To maximize performance, standard preprocessing methods have been used to extract various features of the data. Besides, more complex structures have been used and tested to select the most efficient model. Machine learning and deep learning have played a great role in the detection and prediction of numerous diseases all over the world. Machine learning algorithms also help people to analyze large and complex medical databases and analyze them with clinical understanding. This paper uses these advanced techniques for prediction of malaria. In this paper, convolutional neural network (CNN) has been used to build a model and to train the model to detect the parasitized from non-parasitized samples. The dataset, used here, contains stained red blood cell images. The accuracy of the custom model is 97.50%. 

Malaria Detection Using VGG19 and Deep Convolutional Neural Network

Due to lack of features of traditional network characteristics, DTN becomes one of the promising research paradigms of wireless network. As there are intermittent connectivity and store-carry-forward features, end-to-end acknowledgement is not possible. DTN is also named as ‘opportunistic network’. But interestingly, intermediate nodes may always not behave fairly. Sometimes, due to limited resources (bandwidth and memory) intermediate nodes may behave ‘selfish’. Intermediate nodes may also behave unfairly to next hop node intentionally. This behaviour is known as ‘misbehaving behaviour’. Recent studies on ‘selfish node’ and ‘misbehaving node’ show us the negative impact on network performance (delivery ratio, delivery cost, delivery latency). So we design a slightly new concept inspired by real-life application in our paper showing the positive aspect of the involved nodes in the network. Apart from participating in data forwarding, node may behave as ‘traffic node’ trying to show the right path towards destination.

A New Approach to Trace the Behaviour Pattern of Nodes in the Delay Tolerant Network

Disaster is an undesirable and unfortunate extreme event. It causes great damages to human beings, livestock, infrastructures & communications, socio-economic status and many more. During and after math of disaster, controlling authority at the national and international levels mainly depend on rapidly acquired information to evaluate impacts, set priorities, identify needs, plan early recovery, organize resources. In this paper, we focus on post disaster needs assessment and damage analysis by surveying different post disaster needs assessment prepared by different relief organizations and Governments.

/pdf/survey-on-analysis-of-damage-and-needassessment-in-post-4n9i01gy8h.pdf

Angana Chakraborty

Papers

A Novel Approach for Non-cooperative Node Detection and Avoidance Using Reputation-Based Scheme in Mobile Ad hoc Network

Performance analysis of service classes for IEEE 802.16m QoS optimization

Malaria Detection Using VGG19 and Deep Convolutional Neural Network

A New Approach to Trace the Behaviour Pattern of Nodes in the Delay Tolerant Network

Survey on Analysis of Damage and NeedAssessment in Post Disaster Environment