Feasibility study: The Train Brain for bus
|Coordinator||COMMUTER COMPUTING AB|
|Funding from Vinnova||SEK 490 000|
|Project duration||June 2018 - September 2018|
Purpose and goal
This project is aimed at investigating the possibility of adapting The Train Brain´s automated delayed forecasts, for passenger train traffic, to bus traffic.
Expected results and effects
The study has found that the model can be adapted to bus traffic.
Planned approach and implementation
1. We received data for all bus services 2017-10-18 to 2018-09-19 within SL´s traffic area 2. The data was washed for analysis 5. In parallel, we found that we have data access to real-time data that allows the Train Brain model to work in real time 6. Data is then cleaned and structured in such a way that we can test The Train Brain 7. Data is also structured for pattern identification for certain stretches 8. Test and analysis are conducted in R, run on a 40-core production server with 256GB RAM 9. The results are summarized in a final report