| I&T Solution |
Forecasting allowable time of train operation using open weather data from HKO and collected data from SACEM datalogger using reinforcement learning model
(REF: S-1662) |
| Trial Project |
|
| Solution Feature |
- Collect railway safety system (SACEM) collected Realtime railway operation status from our 2XX SACEM datalogger train and historical information since 2020
- Collect open data from HKO that including the weather information (like wind speed) that including the realtime
- From historical information of SACEM train data and HKO data, find out the threshold condition that train operation against environmental condition under different lines
- Compare the collected data, use reinforcement model to predict the time limit that could achieve the limited threshold condition
- Provide with API to enquiry the predict result of 3rd parties and a cloud based web app for public to enquiry the estimated time
|
| Trial Application and Expected Outcome |
- Trial system could be import into our existing developed SACEM monitoring system in MTR that could direct evaluate by OCC and design team
- After internal testing, we could port the result to the web app in public cloud system for public to enquiry
- The expected outcome will benefit the OCC team have one more reference for the operation time and the forecast for shut down planning
- The accuracy shall achieve 80-90% in term of +/-15 minutes
|
| Info on I&T Solution Provider |
| Solution Provider | : | GNS TECHNOLOGY LIMITED | | Address | : | Unit 707, 7/F, Kai Fuk Industrial Centre, 1 Wang Tung Street, Kowloon Bay, Kowloon, HK | | Contact Person | : | Anson Ng |
| Position | : | Managing Director | | Tel | : | 97425753 | | Email | : |
gns@gnstec.com.hk | | Webpage | : | www.gnstec.com.hk |
|
