Ganapathy Sannasi

Bio: Ganapathy Sannasi is an academic researcher from VIT University. The author has contributed to research in topics: Wireless sensor network & Recommender system. The author has an hindex of 5, co-authored 15 publications receiving 86 citations. Previous affiliations of Ganapathy Sannasi include Anna University.

QoS Aware Trust Based Routing Algorithm for Wireless Sensor Networks

Abstract: In Wireless Sensor Network (WSN), the lifetime optimization based on minimal energy consumption and security are the crucial issues for the effective design of protocols to perform multi-hop secure routing. In order to address these issues, we propose a new routing protocol called Secured Quality of Service (QoS) aware Energy Efficient Routing Protocol in this paper which is designed based on trust and energy modelling for enhancing the security of WSN and also to optimize the energy utilization. In this proposed work, the trust modelling uses an authentication technique with a key based security mechanism for providing trust scores. Moreover, three types of trust scores namely direct, indirect and overall trust scores are calculated in this work for enhancing the security of communication. In addition, a cluster based secure routing algorithm is proposed in this work in which the cluster head has been selected based on QoS metrics and trust scores to perform cluster based secure routing. Finally, the final path has been selected based on path-trust, energy and hop count to efficiently carry out the secure routing process. The proposed work has been assessed by simulations carried out using NS2 simulator. The simulation results demonstrate that the proposed algorithm provides better performance in terms of increase in packet delivery ratio, network life time and security. Moreover, it provides reduction in delay and energy consumption when the proposed secure routing algorithm is compared to the other related secure routing algorithms.

An intelligent fuzzy rule-based e-learning recommendation system for dynamic user interests

Abstract: A relevant and suitable content recommendation is an important and challenging task in e-learning. Relevant terms are retrieved in a recommender system that should also cope with varying user preferences over time. This paper proposes a novel recommendation system which provides suitable contents by refining the final frequent item patterns evolving from frequent pattern mining technique and then classifying the final contents using fuzzy logic into three levels. This is achieved by generating frequent item patterns after consolidating the user interest changes with an extended error margin quotient. Moreover, fuzzy rules are used in this work to enable the rule mining constraints for accommodating all types of learners while applying rules on the pattern tables. This method aims at mining the data stream preferences into equal-sized windows and caters to the varying user interest ratings over time. Experiments prove its efficiency and accuracy over existing conventional methods.

Fuzzy family tree similarity based effective e-learning recommender system

Abstract: Recommender systems are nowadays used on a larger scale in e-learning field to populate the items of interest to the online users. These users, be it, teachers, students or researchers are in need of relevant recommendations list than a list containing mostly irrelevant or unordered recommendations. This paper aims at generating a list of recommendation alternatives with highest rankings among the anticipatory ratings of various key concepts ready to be read through by the online users. This novel approach makes use of fuzzy family tree similarity algorithm to select the key concepts that are of more interest to the online users. Empirical evaluations prove that the proposed technique is efficient and feasible in including the key concepts in the recommendation list, which would otherwise be left out in the conventional tree similarity technique. Anticipatory ratings are determined based on the recommendation alternatives, user key concept rate (UKCR) matrix and neighbors sorted in the order of semantic and content similarities.

Trust Based Ciphertext Policy Attribute Based Encryption Techniques for Decentralized Disruption Tolerant Networks

Abstract: In extreme wireless network environment, Mobile Nodes suffers from inter communication such as military networks and hostile environments. Disruption Tolerant Networks (DTNs) technologies provide challenging and successful solutions for end to end connection between nodes. In this network, the encrypted data is stored in the external storage node and retrieved it by decryption from it, as the confidential data to be retrieved securely it has to consider some security sch0065mes.Attribute Based Encryption (ABE) is a new approach to cryptography that full fills the requirements for secure data retrieval in DTNs. The existing system involves some challenging issues like intimacy, attribute update and trust management and in Cipher text Policy Attribute Based Encryption (CP-ABE) having key updating issue. To overcome this, we propose a new scheme in this paper for securing the data accessed which is based on the Trust based Cipher text Policy Attribute based Encryption (TCP ABE) scheme which provides a scalable way for confidential data retrieval and reduce complexity. In addition, the location of the node is also traced by using network protocols to make more flexible and for easy communication between surrounding nodes. Here various key authorities maintain their attributes independently. In addition, trust values are calculated for each node and updated in the trust table each and every time. This proposed method provides enhanced security using trust based encryption technique when compared with the other existing approaches.

QoS Aware Trust Based Routing Algorithm for Wireless Sensor Networks

Abstract: In Wireless Sensor Network (WSN), the lifetime optimization based on minimal energy consumption and security are the crucial issues for the effective design of protocols to perform multi-hop secure routing. In order to address these issues, we propose a new routing protocol called Secured Quality of Service (QoS) aware Energy Efficient Routing Protocol in this paper which is designed based on trust and energy modelling for enhancing the security of WSN and also to optimize the energy utilization. In this proposed work, the trust modelling uses an authentication technique with a key based security mechanism for providing trust scores. Moreover, three types of trust scores namely direct, indirect and overall trust scores are calculated in this work for enhancing the security of communication. In addition, a cluster based secure routing algorithm is proposed in this work in which the cluster head has been selected based on QoS metrics and trust scores to perform cluster based secure routing. Finally, the final path has been selected based on path-trust, energy and hop count to efficiently carry out the secure routing process. The proposed work has been assessed by simulations carried out using NS2 simulator. The simulation results demonstrate that the proposed algorithm provides better performance in terms of increase in packet delivery ratio, network life time and security. Moreover, it provides reduction in delay and energy consumption when the proposed secure routing algorithm is compared to the other related secure routing algorithms.

An ACO-Based Clustering Algorithm

Abstract: Data clustering is one of important research topics of data mining. In this paper, we propose a new clustering algorithm based on ant colony optimization, called Ant Colony Optimization for Clustering (ACOC). At the core of the algorithm we use both the accumulated pheromone and the heuristic information, the distances between data objects and cluster centers of ants, to guide artificial ants to group data objects into proper clusters. This allows the algorithm to perform the clustering process more effectively and efficiently. Due to the nature of stochastic and population-based search, the ACOC can overcome the drawbacks of traditional clustering methods that easily converge to local optima. Experimental results show that the ACOC can find relatively good solutions.

An Energy Efficient Clustered Gravitational and Fuzzy Based Routing Algorithm in WSNs

Abstract: Wireless sensor networks consist of many tiny sensor nodes which are deployed in various geographical locations for sensing the normal spectacles and also to transmit the collected information to the base station which is also named destination node through multiple nodes present in the network. Most of the existing heuristics algorithms used for finding the optimal routes have limitations in the provision of effective solutions for routing and clustering mechanisms in larger search spaces. Hence, when the search space increases exponentially, the chance of creating the optimal solution for clustering and routing is decreasing and ultimately an un-optimized process depletes the sensor node resources. In order to address the challenges and limitations present in the existing routing systems, two new heuristics algorithms namely gravitational approach based clustering method and a clustered gravitational routing algorithm have been proposed in this paper for providing an optimal solution for efficient clustering and effective routing. Moreover, a fuzzy logic based deductive inference system has been designed and used in this work for selecting the most appropriate nodes as cluster head nodes from the nodes present in each cluster. The simulation results obtained from this work show that the clustering accuracy and the network lifetime are increased and the energy consumption as well as delay are reduced with the application of these proposed algorithms.