110 likes | 126 Views
Efficient Valid Scope Computation for Location-Dependent Spatial Queries in Mobile and Wireless Environments. Ken C.K. Lee, Wang-Chien Lee, Hong Va Leong, Brandon Unger and Baihua Zheng, ICUIMC’09 Manu Shukla August 2, 2009. Introduction.
E N D
Efficient Valid Scope Computation for Location-Dependent Spatial Queries in Mobile and Wireless Environments Ken C.K. Lee, Wang-Chien Lee, Hong Va Leong, Brandon Unger and Baihua Zheng, ICUIMC’09 Manu Shukla August 2, 2009
Introduction • Mobile clients access information with respect to their location by submitting Location Dependent Spatial Queries (LDSQ) to Location-Based service servers • Due to limited mobile bandwidth and limited client battery life, frequent LSDQ submission must be avoided • Authors define the concept of valid-scope, a spatial area in which a set of LSDQs will retrieve exactly the same query results • Propose algorithms for valid scope computation for common LDSQs like range query and NN query
Valid Scope • Determine a small and representative subset of non-result objects called “complementary objects” together with result objects • Use R-tree and best first search
Valid Scope for NN Query • Objects that contribute the bisector as voronoi cell perimeter are added to complementary objects • Exploit the largest empty circle property
NN Query • Two steps to determine the NN query result and the valid scope of the result and one index lookup • First retrieve the nearest object • Examine non-result objects to refine valid scope
Valid Scope for Range Query • Minkowski region is defined as • Valid Scope is computed as an area equal to the intersection of the Minkowski regions of all result objects
Range Query • Formula for valid scope can be defined as
Experiments • Compared with time-parameterized query method by Zhang et.al • Three data sets, real and two synthetic based on Gaussian and Uniform distribution
Experiments • Overhead incurred by valid scope computations
Conclusions • Valid for spatial queries in mobile environment where bandwidth and battery power in mobile devices are limited • Most calculations done on server and hence scalable