Random and directed walk-based top-(k) queries in wireless sensor networks

Area of Science:

• Computer Science
• Wireless Sensor Networks
• Data Mining

Background:

• Filter-based top-k query approaches are standard but sensitive to network parameters like size, sensor dynamics, and data range.
• Existing methods struggle with network variability, impacting query accuracy and efficiency.

Purpose of the Study:

• To propose novel robust and efficient top-k query approaches for wireless sensor networks.
• To address the limitations of existing filter-based methods in dynamic network environments.

Main Methods:

• Introduction of a random walk-based top-k query (RWTQ) approach.
• Development of a directed walk-based top-k query (DWTQ) approach utilizing spatial correlations.
• Tokens traverse the network, collecting and processing sensor readings randomly or via a directed strategy.

Main Results:

• Both RWTQ and DWTQ demonstrate significant robustness against network size, sensor dynamics, and data range variations.
• DWTQ significantly outperforms existing methods (TAG, FILA, EXTOK) in transmission cost, energy consumption, and network lifetime.

Conclusions:

• RWTQ and DWTQ provide robust solutions for top-k queries in wireless sensor networks.
• DWTQ offers superior performance metrics, making it a highly efficient approach for practical applications.