Modelling the UK?s Busiest Road - Using Detailed Data to Inform Design at Junction 15 on the M25
E Seaman, SIAS, UK
This paper details the extensive data analysis process undertaken to ensure the robust development of a microsimulation model, which included Active Traffic Management (ATM) to test merge layouts and compare option designs at Junction 15 of the M25.
The UK?s busiest road is London?s M25 orbital motorway, carrying traffic flows of up to 250,000 vehicles per day. The task of keeping traffic moving on the M25 is seen as vital to the UK?s economy, and the responsibility for this falls to the UK government?s Highways Agency. The M25 lies within the Agency?s Area 5 region, and SIAS Limited was commissioned in 2009 to support Area 5?s Managing Agent, Mouchel, to analyse merge layouts at Junction 15. The M25 close to Junction 15 is unusually wide, and spans across 12 lanes immediately south of the junction. The junction itself is located directly west of London, and connects the M25 and M4 motorway, which links London to the west of England and into Wales.
The study involved extensive data analysis of detailed traffic data, and required the development of a microsimulation model, which included Active Traffic Management (ATM), to test merge layouts and compare design options. The principal section of study involved the merge layouts for the links from the M4, both eastbound and westbound, to the M25 northbound.
Conditions on the M25 can vary considerably day by day, so the historical approach of attempting to model an ?average day? was considered inappropriate. Rather than make assumptions as to the likely traffic flows on certain days of the week, daily data records were analysed.. Detailed loop detector data was made available from the M25?s Motorway Incident Detection and Signalling system (MIDAS) for the past two years. This contained detailed one-minute information on flows and speeds by lane and vehicle type on both the M25 and M4 motorways. Using the MIDAS data, each day was treated as an individual entity and placed within a group containing days exhibiting similar flow and speed characteristics. This helped to identify the most frequently occurring flow conditions. This key aspect of the study avoided grouping flows into the chronological arrangement of days of the week or months more normally considered to exhibit similar traffic conditions.
This paper details the extensive data analysis process, how the most frequently occurring flow conditions were identified, and how the data was applied in the modelling context. A description of the inclusion of ATM within the microsimulation model will be presented. The paper will also discuss the model calibration and validation, and the wide range of validation outputs considered. The paper emphasises the importance of preliminary data analysis to the proper representation of the commonly occurring flow conditions and the determination of robust design solutions.
Association for European Transport