Probabilistic nearest neighbor queries on uncertain moving object trajectories

Abstract

Nearest neighbor (NN) queries in trajectory databases have received significant attention in the past, due to their applications in spatiotemporal data analysis. More recent work has considered the realistic case where the trajectories are uncertain; however, only simple uncertainty models have been proposed, which do not allow for accurate probabilistic search. In this paper, we fill this gap by addressing probabilistic nearest neighbor queries in databases with uncertain trajectories modeled by stochastic processes, specifically the Markov chain model. We study three nearest neighbor query semantics that take as input a query state or trajectory q and a time interval, and theoretically evaluate their runtime complexity. Furthermore we propose a sampling approach which uses Bayesian inference to guarantee that sampled trajectories conform to the observation data stored in the database. This sampling approach can be used in Monte-Carlo based approximation solutions. We include an extensive experimental study to support our theoretical results.

PINOCCHIO: Probabilistic Influence-Based Location Selection over Moving Objects

“…NN queries over moving objects [28] find the closest trajectory to a query trajectory or point at a particular timestamp or in a certain interval. On the other hand, Continuous Aggregate NN [29] aims to continuously retrieve a moving object whose aggregate distance to all facilities is the smallest.…”

Dual Space-Time Representation

“…One research direction is to capture the uncertain nature of object/query trajectories (Cheng et al 2004;Niedermayer et al 2013). Probabilistic continuous querying is concerned with presenting a number of possible results along with probability assessments to a user over a period of time.…”

An extendable framework for managing uncertain spatio-temporal data

“…To solve this problem, the main challenge that we approached in our most recent publication ( [7]) is to enrich these general movement patterns given by the Markov model by knowledge of individual observations of an uncertain moving object. We pre-compute the probabilities that, at a given time, a given object performs a transition between two nodes of the network, given all observations of an object, at past, present and future times.…”

“…However, the location of objects at subsequent discrete points of time are highly correlated, as commanded for example by laws of physics and speed limits. Ignoring this positive correlation leads to systematic and significant errors as shown in [7].…”

“…The shortcomings of these models have been addressed in previous work ( [3,2,7]), by applying models from statistics, namely stochastic processes and Bayesian inference, to treat uncertain moving objects in a probabilistic way. In the following Section 2, the theoretic base of these works is sketched.…”