International Journal of Innovative Research and Development

Soumen Saha ¹, Devadutta Sinha ¹, Utpal Roy ²

Affiliations
1 Department of CSE, University of Calcutta, Kolkata, West Bengal, IN
2 Department of Computer & System Sciences, Siksha Bhavana, Visva Bharati, West Bengal, IN

Abstract

A recent upcoming smart traffic is Vehicular ad hoc network (VANET) is also considered as a sub-set of mobile ad hoc network (MANET). It provides wireless ad-hoc communication in between vehicles and vehicle to roadside equipments. Based on this technology traffic  network is classified into two types 1. vehicle to vehicle interaction, 2. vehicle to infrastructure interaction. The objective of VANET is to provide safe, secure and automated traffic system. For this automation of traffic technique is different types of routing protocols have been developed. But routing protocols of MANET are not directly applicable to VANET.

VANET is designed on IEEE 802.11b wireless standard. This helps to communicate vehicle to vehicle and vehicle to trafic communications. According to Federal Communications Commission (FCC) suggests for VANET frequency spectrum of 75 MHz in the range of 5.850 GHz to 5.925 GHz. It communicate  from one vehicle (source) to another vehicle (destination) through different vehicles (intermediate nodes). A numbers of different routing protocols for communication, ie multimedia data, text data etc. from one vehicle (node) to another vehicle are existing. The Ad hoc On-Demand Distance Vector (AODV)[1] routing algorithm is one of the popular routing  protocols for ad-hoc mobile networks. AODV is used for both unicast and multicast routing. Earlier we have modified AODV with stack and dqueue, where we have find a considerable amount of betterment of result with respect of AODV.  In this paper, we propose and implement in the NCTUns-6.0 simulator neural network based Modified AODV on dqAODV(dqueue implemented AODV)[2] and sAODV (stack implemented AODV)[3] routing protocol considering Power, TTL, Node distance and Payload parameter to  find the optimal route from the source station (vehicle) to the destination station in VANET com-munications.  Further we compare both neural network optimized dqAODV (dqueue implemented AODV) and sAODV (stack implemented AODV) performance on a signal fading model (Rayleigh). This gives us a better result in ai-sAODV(Neural network optimized Stack implemented AODV) compared to queue implemented AODV (ai-dqAODV).

Keywords

VANET, Dqueue, Stack, AODV, Neural Network, Rayleigh Signal Fading, NCTUns-6.0.

Full Text

   

Soumen Roy ¹, Utpal Roy ², D. D. Sinha ¹

Affiliations
1 Department of Computer Science and Engineering, University of Calcutta, West Bengal, IN
2 Department of Computer & System Sciences, Visva-Bharati, Santiniketan, West Bengal, IN

Abstract

User authentication or identification is the big challenges in E-Business. In this paper, we have implemented a typing biometric technique which increases the security level up to 98.1% without changing existing authentication technique. Habitual typing speed pattern is a behavioural biometric characteristic in Biometric Science can be effectively implemented to classify the users. This typing speed pattern is promising as biometric characteristics which cannot be lost or stolen in addition with inexpensive to collect. Many statistical, distance-based and machine learning algorithms are proposed on habitual typing pattern and many have obtained impressive results, but in practice, the accuracy level is not much promising, it demands higher level of security and reliability. In our experiment, we have collected press and release time of 12096 keystrokes using Java Applet programming form 12 individuals during 12 months in 4 sessions for 1440 samples then we analysed that data using R statistical programming language and obtained average Equal Error Rate (EER) of 21 different data-mining and distance-based classification algorithms and compared their performance in accuracy to search the suitable algorithms on typing patterns. But in evaluation process, a classifier’s average Equal Error Rate (EER) widely jumped from 1.9% to 63%. The question may arise, which classifier is suitable on typing speed patterns, which pattern of string is suitable. To get the answer, we have started our experiment and created our own rhythmic keystroke dynamics database of different pattern of strings and executed various classification algorithms on it, so, we can compare their performance soundly. 

Keywords

Keystroke Dynamics, EER, Behavioral Biometric, Canberra, Chebyshev, Czekanowski, Gower, Intersection, Kulczynski, Lorentzian, Minkowski, Motyka, Ruzicka, Soergel, Sorensen, Wavehedges, Manhattan Distance, Euclidean Distance, Mahanobolis Distance, Z Score, KMean, SVM, NaiveBaysian, ROC Curve.

Full Text

   

Subhasis Banerjee ¹, Utpal Roy ¹

Affiliations
1 Department Computer & System Sciences, Visva-Bharati, Santiniketan, West Bengal, IN

Abstract

In this paper, we present a scheme to authenticate SMS communication. In the suggested scheme authentication text is generated depending upon a ticket. The ticket is generated randomly by the service provider. Our authentication scheme is meant to work on android platform. The Android platform has been dealt as a topic of mobile security because Android is an open platform whose sources can be observed by anyone. Unauthenticated message may create serious problem for many mobile based applications.. In this paper, we have developed a two way authentication system which can work for any mobile based applications in android. We have also used simple hashing technique to create the message digest.

Keywords

BTS, UE, SMSC, PKI.

Full Text