Improving Fuzzy-Logic based Map-Matching Method with Trajectory Stay-Point Detection

Slide 1

Slide 1 text

Improving Fuzzy-Logic based Map-Matching Method with Trajectory Stay-Point Detection Authors: Minoo Jafarlou, Omid Mahdi Ebadati, and Hassan Naderi

Slide 2

Slide 2 text

Introduction   Moving object databases have some irregularity, we need to preprocess the trajectory   Vehicle GPS trajectory datasets include stay-points abnormality, which defines as noise in dataset   Stay-points make matching algorithms mismatch trajectories to irrelevant roads

Slide 3

Slide 3 text

The map-matching process

Slide 4

Slide 4 text

Fuzzy-logic system blocks

Slide 5

Slide 5 text

Initial map-matching

Slide 6

Slide 6 text

Subsequent map-matching 1 (SMP1) θ↑′  is the vehicle's direction at previous point. Δd is d−d1. Abs (θ−θ↑′ ) is heading increment.

Slide 7

Slide 7 text

Subsequent map-matching 2 (SMP2) d is the traveled distance by vehicle, d2= AO+OM, d3= AO+ON, d4= AO+OO’+OQ

Slide 8

Slide 8 text

Pseudo-code for fuzzy map-matching algorithm

Slide 9

Slide 9 text

Map-matching process

Slide 10

Slide 10 text

Different density levels

Slide 11

Slide 11 text

Stay-Point sample

Slide 12

Slide 12 text

Stay-Point •  Stay-point region misleads the map-matching process

Slide 13

Slide 13 text

Digital road network and tested GPS trajectory

Slide 14

Slide 14 text

Proposed method follow-chart

Slide 15

Slide 15 text

Three nearest neighbor distance curves

Slide 16

Slide 16 text

Number of clustered and non-clustered points for different epsilons

Slide 17

Slide 17 text

Clustered stay-points data with DBSCAN

Slide 18

Slide 18 text

Accuracy performance evaluation •  Number of correct links detected among 399 links

Slide 19

Slide 19 text

Time efficiency evaluation •  Overall processing time comparison between two approaches

Slide 20

Slide 20 text

Volume efficiency evaluation •  The proposed approach decreased the data volume

Slide 21

Slide 21 text

Speed efficiency evaluation •  Average execution time speed for each point is calculated

Slide 22

Slide 22 text

Map-matching sample on the dataset

Slide 23

Slide 23 text

References   [1] Jafarlou, Minoo, Omid Mahdi Ebadati E, and Hassan Naderi. "Improving Fuzzy-Logic based Map- Matching Method with Trajectory Stay-Point Detection." arXiv e-prints (2022): arXiv-2208.   [2] Schubert, Erich, et al. "DBSCAN revisited, revisited: why and how you should (still) use DBSCAN." ACM Transactions on Database Systems (TODS) 42.3 (2017): 1-21.   [3] Quddus, Mohammed A., Robert B. Noland, and Washington Y. Ochieng. "A high accuracy fuzzy logic based map matching algorithm for road transport." Journal of Intelligent Transportation Systems 10.3 (2006): 103-115.