Skip to content

SH51 / Napier Road Roundabout Performance Modelling

Our team was engaged to assess the efficiency and safety performance of this 5 legged dual lane roundabout. An understanding of entry and departure lanes, lane discipline, compliance with lane markings, near misses and testing of different lane configurations were key aspects of the investigation. 

We quickly identified that due to the requirements for logging entry and departure lanes that conventional manual traffic surveys would require significant number of surveyors. It was decided that utilising a drone would provide the most efficient and reliable data. We utilised AI software to undertake the traffic counts and provide speed data. This was UCL’s 1st project utilising this technology, we therefore decided to undertake some manual quality audits to understand the accuracy of the AI analysis. This work determined that the AI analysis provided a very high level of accuracy, higher than could be expected from manual surveyors.

Preparation of a SIDRA model to assess the performance of the roundabout and alternative configurations was completed utilising the survey data.

Speed Heatmap

This approach to traffic surveys provides several benefits:

  • A higher level of accuracy compared to manual surveys;
  • Faster processing of traffic count data;
  • Increased richness of data collected with speed data;
  • The ability to develop AI crash models based on conflict angles;
  • Faster calibration of traffic models;
  • A visual record of surveys provides a robust evidence base if ever challenged; and
  • The visual record allows technical staff and stakeholders to observe behaviours which may not be noted during site visits.