PMID- 34149057
OWN - NLM
STAT- PubMed-not-MEDLINE
LR  - 20220615
IS  - 1958-9395 (Electronic)
IS  - 0003-4347 (Print)
IS  - 0003-4347 (Linking)
VI  - 77
IP  - 5-6
DP  - 2022
TI  - Generalization aspect of accurate machine learning models for CSI-based 
      localization.
PG  - 345-357
LID - 10.1007/s12243-021-00853-z [doi]
AB  - Localization is the process of determining the position of an entity in a given 
      coordinate system. Due to its wide range of applications (e.g. autonomous 
      driving, Internet-of-Things), it has gained much focus from the industry and 
      academia. Channel State Information (CSI) has overtaken Received Signal Strength 
      Indicator (RSSI) to achieve localization given its temporal stability and rich 
      information. In this paper, we extend our previous work by combining classical 
      and deep learning methods in an attempt to improve the localization accuracy 
      using CSI. We then test the generalization aspect of both approaches in different 
      environments by splitting the training and test sets such that their intersection 
      is reduced when compared with uniform random splitting. The deep learning 
      approach is a Multi Layer Perceptron Neural Network (MLP NN) and the classical 
      machine learning method is based on K-nearest neighbors (KNN). The estimation 
      results of both approaches outperform state-of-the-art performance on the same 
      dataset. We illustrate that while the accuracy of both approaches deteriorates 
      when tested for generalization, deep learning exhibits a higher potential to 
      perform better beyond the training set. This conclusion supports recent 
      state-of-the-art attempts to understand the behaviour of deep learning models.
CI  - (c) Institut Mines-Telecom and Springer Nature Switzerland AG 2021.
FAU - Sobehy, Abdallah
AU  - Sobehy A
AUID- ORCID: 0000-0003-1455-5165
AD  - Samovar, CNRS, Telecom SudParis, 9 Rue Charles Fourier, 91000 Evry, France. GRID: 
      grid.4444.0. ISNI: 0000 0001 2112 9282
FAU - Renault, Eric
AU  - Renault E
AD  - EISEE Paris, 2 Boulevard Blaise Pascal, 93160 Noisy-le-Grand, France.
FAU - Muhlethaler, Paul
AU  - Muhlethaler P
AD  - Inria, 2 Rue Simone IFF, 75012 Paris, France. GRID: grid.5328.c. ISNI: 0000 0001 
      2186 3954
LA  - eng
PT  - Journal Article
DEP - 20210614
PL  - France
TA  - Ann Telecommun
JT  - Annales des telecommunications
JID - 9875870
PMC - PMC8200796
OTO - NOTNLM
OT  - Channel state information
OT  - Deep learning
OT  - Ensemble learning
OT  - Generalization
OT  - Indoor localization
OT  - KNN
OT  - MIMO
OT  - Neural networks
EDAT- 2021/06/22 06:00
MHDA- 2021/06/22 06:01
PMCR- 2021/06/14
CRDT- 2021/06/21 05:43
PHST- 2020/10/14 00:00 [received]
PHST- 2021/05/04 00:00 [accepted]
PHST- 2021/06/22 06:00 [pubmed]
PHST- 2021/06/22 06:01 [medline]
PHST- 2021/06/21 05:43 [entrez]
PHST- 2021/06/14 00:00 [pmc-release]
AID - 853 [pii]
AID - 10.1007/s12243-021-00853-z [doi]
PST - ppublish
SO  - Ann Telecommun. 2022;77(5-6):345-357. doi: 10.1007/s12243-021-00853-z. Epub 2021 
      Jun 14.