| Item type |
SIG Technical Reports(1) |
| 公開日 |
2018-09-18 |
| タイトル |
|
|
タイトル |
Route graph: joint map-matching by maximizing posterior probability |
| タイトル |
|
|
言語 |
en |
|
タイトル |
Route graph: joint map-matching by maximizing posterior probability |
| 言語 |
|
|
言語 |
eng |
| 資源タイプ |
|
|
資源タイプ識別子 |
http://purl.org/coar/resource_type/c_18gh |
|
資源タイプ |
technical report |
| 著者所属 |
|
|
|
Artificial Intelligence Laboratory, Fujitsu Laboratories Ltd. |
| 著者所属 |
|
|
|
Artificial Intelligence Laboratory, Fujitsu Laboratories Ltd. |
| 著者所属 |
|
|
|
Artificial Intelligence Laboratory, Fujitsu Laboratories Ltd. |
| 著者所属 |
|
|
|
Graduate School of Information Science and Technology, Hokkaido University |
| 著者所属 |
|
|
|
Graduate School of Information Science and Technology, Hokkaido University |
| 著者所属(英) |
|
|
|
en |
|
|
Artificial Intelligence Laboratory, Fujitsu Laboratories Ltd. |
| 著者所属(英) |
|
|
|
en |
|
|
Artificial Intelligence Laboratory, Fujitsu Laboratories Ltd. |
| 著者所属(英) |
|
|
|
en |
|
|
Artificial Intelligence Laboratory, Fujitsu Laboratories Ltd. |
| 著者所属(英) |
|
|
|
en |
|
|
Graduate School of Information Science and Technology, Hokkaido University |
| 著者所属(英) |
|
|
|
en |
|
|
Graduate School of Information Science and Technology, Hokkaido University |
| 著者名 |
Hiroya, Inakoshi
Junichi, Shigezumi
Tatsuya, Asai
Takuya, Kida
Hiroki, Arimura
|
| 著者名(英) |
Hiroya, Inakoshi
Junichi, Shigezumi
Tatsuya, Asai
Takuya, Kida
Hiroki, Arimura
|
| 論文抄録 |
|
|
内容記述タイプ |
Other |
|
内容記述 |
We propose a joint map-matching for estimating unobservable paths from GPS traces. Our method is the first to maximize the posterior probability of stochastic generative model, in which traces are emitted as vehicles drive the roads. We employed the EM algorithm to find the parameters of the generative model, as well as to evaluate the expectations of the latent variable, which is indeed the estimated unobservable path. The EM algorithm is reduced to the exploratory search of the route graph, which is the geometric graph that is most likely emitting the traces and corresponds to the parameters of the model. Due to this stochastic formulation, our method works well with the presence of sampling noises in the traces. We report that the residual degradation of the estimated paths was no more than 7.0% even when they are sampled at a rate as low as 40%. |
| 論文抄録(英) |
|
|
内容記述タイプ |
Other |
|
内容記述 |
We propose a joint map-matching for estimating unobservable paths from GPS traces. Our method is the first to maximize the posterior probability of stochastic generative model, in which traces are emitted as vehicles drive the roads. We employed the EM algorithm to find the parameters of the generative model, as well as to evaluate the expectations of the latent variable, which is indeed the estimated unobservable path. The EM algorithm is reduced to the exploratory search of the route graph, which is the geometric graph that is most likely emitting the traces and corresponds to the parameters of the model. Due to this stochastic formulation, our method works well with the presence of sampling noises in the traces. We report that the residual degradation of the estimated paths was no more than 7.0% even when they are sampled at a rate as low as 40%. |
| 書誌レコードID |
|
|
収録物識別子タイプ |
NCID |
|
収録物識別子 |
AN10505667 |
| 書誌情報 |
研究報告数理モデル化と問題解決（MPS）
巻 2018-MPS-120,
号 3,
p. 1-6,
発行日 2018-09-18
|
| ISSN |
|
|
収録物識別子タイプ |
ISSN |
|
収録物識別子 |
2188-8833 |
| Notice |
|
|
|
SIG Technical Reports are nonrefereed and hence may later appear in any journals, conferences, symposia, etc. |
| 出版者 |
|
|
言語 |
ja |
|
出版者 |
情報処理学会 |
IPSJ-MPS18120003.pdf (743.0 kB)
