Indian Journal of Science and Technology

Abstract

Objectives: The main objectives of this work are to enable mobile nodes predict the switch off states of Hard Disk and other electronic components using APRIORI algorithm which has pre-mined list of timings of every request to be serviced by the node. Methods/Statistical Analysis: The application of this Algorithm is shown via Simulink (MATLAB). We use a Sequence generator of variable inputs (on/off) signals which controls the load circuit of the simulated node. The input of the sequence generator is defined by an example depicting a Mobile Node traversal and represented by a Table residing in the memory of the node. The table values are mined using preset values of Confidence and Support as in APRIORI algorithm. The data set used in the simulation represents all cases of the Algorithm in real time. Findings: Controlling the switch off states using pre-mined data values based on frequent item sets enables a Mobile Node to effectively decide the switch off states or power saving states for various electronic hardware, which includes Hard Disk. The Main power drawn in Hard Disk is by the Motor of the cylinder which if turned off and on (Spin UP/Spin DOWN) effectively will result in more efficient battery savings for each Mobile Node as compared to Adaptive and Threshold based Spin UP/ Spin DOWN based policies. Applications/Improvements: Since the switch timings decided by the algorithm entirely depend on mined values, application of better data mining techniques will result in more efficient battery operation. 

Keywords: Power Conservation, Prediction based Algorithm, Wireless Networks, APRIORI, MANETS