Computing a Subtrajectory Cluster from c-packed Trajectories

TL;DR

This paper introduces a near-linear time approximation algorithm for identifying clusters of similar subtrajectories within c-packed trajectories, significantly improving computational efficiency over previous cubic-time algorithms.

Contribution

The paper presents the first near-linear time algorithm for the subtrajectory cluster problem on c-packed trajectories, surpassing prior cubic-time solutions under certain conditions.

Findings

Achieves near-linear time complexity for the problem

Provides a constant-factor approximation within (1 + ε)

Breaks the conditional lower bound for general trajectories

Abstract

We present a near-linear time approximation algorithm for the subtrajectory cluster problem of $c$-packed trajectories. The problem involves finding $m$ subtrajectories within a given trajectory $T$ such that their Fr\'echet distances are at most $(1 + \varepsilon)d$, and at least one subtrajectory must be of length~$l$ or longer. A trajectory $T$ is $c$-packed if the intersection of $T$ and any ball $B$ with radius $r$ is at most $c \cdot r$ in length. Previous results by Gudmundsson and Wong \cite{GudmundssonWong2022Cubicupperlower} established an $\Omega(n^3)$ lower bound unless the Strong Exponential Time Hypothesis fails, and they presented an $O(n^3 \log^2 n)$ time algorithm. We circumvent this conditional lower bound by studying subtrajectory cluster on $c$-packed trajectories, resulting in an algorithm with an $O((c^2 n/\varepsilon^2)\log(c/\varepsilon)\log(n/\varepsilon))$ time complexity.