LEARNING MODEL GENERATION METHOD, INFERENCE DEVICE, AND WIRELESS TRAIN CONTROL SYSTEM

Title:

LEARNING MODEL GENERATION METHOD, INFERENCE DEVICE, AND WIRELESS TRAIN CONTROL SYSTEM

Document Type and Number:

WIPO Patent Application WO/2021/106508

Kind Code:

A1

Abstract:

A communication status inference device (40) acquires, as data indicative of the radio wave environment of a wireless base station (10) during cognitive wireless communication between an on-vehicle device (52) and the wireless base station (10) disposed alongside a railway, a wireless signal that has been received by a reception device (20) for monitoring located in the vicinity of the wireless base station (10). The communication status inference device (40) inputs the acquired data indicative of the radio wave environment into a communication status inference model, and executes an inference process for outputting per-frequency communication status information based on the signal to noise ratio (SNR) in the wireless communication performed by the wireless base station (10) under that radio wave environment. The communication status inference model is a machine learning model that has been trained with training data, the input of which is data indicative of the radio wave environment of the wireless base station (10), and the output of which is per-frequency communication status information based on the SNR in the wireless communication performed by the wireless base station (10) under that radio wave environment.

Inventors:

SRISOOKSAI TOSSAPORN (JP)
NISHIDA SATOSHI (JP)
NAMBU SHUJI (JP)

Application Number:

PCT/JP2020/041095

Publication Date:

June 03, 2021

Filing Date:

November 02, 2020

Export Citation:

Click for automatic bibliography generation Help

Assignee:

KYOSAN ELECTRIC MFG (JP)

International Classes:

H04W16/18; B61L3/12; G06N3/04; H04B17/309; H04B17/391; H04W4/42; H04W16/14

Foreign References:

JP2018142957A	2018-09-13
JP2015518356A	2015-06-25
CN108712748A	2018-10-26
JP2013239935A	2013-11-28

Other References: