Reasearch On Distributed Approximate Similarity Search Approach In Wireless Sensor Networks

With the rapid growth of the intelligent terminal and the applications of the Internet of Things(IOT), massive amounts of the sensed data have become diversified. The demands for the multimedia data such as images and video streams have increased data by day. How to process the massive amounts of high-dimensional multimedia data is still an open problem.For high-dimensional data, approximate similarity search is an effective means to improve the performance at the cost of a little loss in accuracy. Approximate similarity search has been widely applied in the field of the tradition database. However, considering the distributed nature of WSN, how to acquire feasible solutions to the high-dimensional similarity search in large-scale WSN is still an open problem.To address the above issues, a distributed approximate similarity search method in proposed in this article based on the locality sensitive hashing(LSH), which projects the high-dimensional sensed data to the low-dimensional fingerprint to achieve energy efficiency. At the same time, The distributed approximate similarity search method provides the comparison of similarity score with the specified threshold to filter out the irrelevant sensed data. Finally, extensive simulations are conducted using the real world data. Simulation results show that the proposed algorithm provides significant performance gain in terms of search accuracy and energy efficiency.