{ "cells": [ { "cell_type": "markdown", "id": "6d5b84b1-9b3b-41fe-aece-3fff0dd6404a", "metadata": {}, "source": [ "# Aggregated Trajectory: position errors and shape deviation\n", "\n", "In this tutorial, we will examine the aggregation of trajectories using two matching distances: the Fréchet distance and the DTW-L2 distance, and thus see its ability to reconstruct accurately the common path followed by all the individual sample trajectories by comparing the position errors and the shape deviation.\n", "\n", "Our experiment will be conducted in four steps:\n", "1. Step 1: creating a synthetically reference track considered as the ground truth track,\n", "2. Step 2: creating a set of simulated tracks from the reference track, \n", "3. Step 3: computing the aggregation track from the set previously constructed (step 2),\n", "4. Step 4: the error (position and shape) between the estimated and ground truth track is then evaluated.\n", "\n", "This tutorial is based on the work presented in [1 SIGSPATIAL 2024]\n", "\n", "