A Trial of Vehicle Activated Signs to Prevent Road Vehicles Blocking Back Over Level Crossings Claire Turner 1 Marcus Beard 2 Dr Alastair Mckenzie-Kerr 1 John McMorrow 3 1 Human Engineering Limited, Bristol, UK 2 Arthur D. Little, London, UK 3 RSSB, London, UK.
What is Blocking Back? Blocking back - a recognised risk within the rail industry Included within the All Level Crossing Risk Model (ALCRM) following collisions at Pooley Green and Waterbeach As road vehicles traverse a level crossing they can encounter obstructions causing a queue to form. That queue can sometime extend back to the level crossing and even encroach on the crossing
Vehicle Activated Signs VAS: Vehicle activated detection loop positioned 50m downstream of crossing If sensor is occupied for more than 2 seconds, VAS is activated Displays Keep Crossing Clear Remains activated for 60 seconds, unless loop is occupied again
Boldon and Tile Shed Located in North East England Passenger, Metro and Freight traffic Both crossings are approximately 160m from the miniroundabout Video recording at Boldon and Tileshed: 14 months pre- VAS installation 25 months post- VAS installation
Project objectives 1. Have VAS reduced or prevented blocking back? 2. Under what conditions can VAS be successful in reducing blocking back SFAIRP at other AHBs? Project structure Phase 1 Preliminary Review Phase 2 Detailed Review Project team Multi-disciplinary project team: Human Factors Human Engineering Ltd Risk Analysis Arthur D Little Video analysis Nationwide Data Collection
Phase 1: Preliminary analysis Four activities: Analyse video footage to classify blocking back Analyse data to establish reduction or otherwise in blocking back incidents before and after VAS installation Estimate reduction in risk from blocking back Record salient environmental and physical aspects of the crossing that may contribute to blocking back Phase 2 detailed analysis of video footage and data has been planned
Classifying blocking back Systematic approach designed to ensure consistency Five classifications reflecting the levels of severity of blocking back events: Amber 1 Amber 2 Red 1 Red 2 Red 3 Queues classified by position of rearmost vehicle in queue
Amber 1 and 2
Red 1 and 2
Overview of results Focus of the analysis to date is on Boldon Headlines: Severe blocking back occurs infrequently VAS reduces the number of Reds, the increase in Amber 1s can be accounted for by a reduction in Amber 2s Events/day (average) 4 3.5 3 2.5 2 1.5 1 0.5 0 Pre-VAS Post-VAS Amber 1 Amber 2 Red 1 Red 2 Red 3 Blocking back classification But these figures do not accommodate variation in the opportunity for blocking back (e.g. rail and road traffic)
Overview of results Data normalised by Amber 2 = measure of opportunity for BB Reduction in Red 1 and Red 2 Red events (normalised by Amber 2) 0.5 0.4 0.3 0.2 0.1 0 Pre-VAS Post-VAS Red 1 Red 2 Red 3 Blocking back classification Red events (normalised by Amber 2) 0.5 0.4 0.3 0.2 0.1 0 Pre-VAS - Apr/May/Jun 2007 Post-VAS - Apr/May/Jun 2008 Post-VAS - Apr/May/Jun 2009 Red 1 Red 2 Red 3 Blocking back classification Reduction in Red 1 is sustained over long term Reduction in Red 2 over time is apparent, but appears to be short term
Human factors analysis Key error modes associated with blocking back: Incorrect decision making misjudging the space available downstream of the crossing Unintentional behaviour failing to notice crossing safety features Intentional behaviour deliberately choosing to violate safe crossing procedure Behavioural profiles: Tailgaters Follow the lead of the vehicle in front Cautious Slow approach to the crossing, quick traverse once the decision to cross is made Oblivious No obvious deceleration on approach to crossing Late decision makers Follow the lead of vehicle in front, then reassess position once stationary Predictors Predict when is safe to cross from the movement of queue ahead
Risk analysis Estimating the risk of collision arising from a blocking back event needs to consider other factors Vehicles foul running line (Red 2 or Red 3) Queue > than strike-in time Train arrives before the queue clears AND Train / road vehicle collision Vehicle driver unable to move to a safe position
Risk analysis Few Red 2 and 3 queues had longer duration than the crossing strike in times 80 70 60 50 40 30 20 10 0 Red 2 and 3 Duration Red2 Red3 Strike in times Freight 89s Slow passenger 40s Fast passenger 32s Duration (s) 11 Red 2s and 3 Red 3s in the whole project history had queue times longer than the crossing strike in times
Risk analysis A simple risk model driven by blocking back data estimates a collision once in about every 20 years at Boldon Vehicles foul running line (Red 2 or Red 3) Red 2: 0.3/day Red 3: 0.09/day Queue > strike-in time See previous Train arrives before the queue clears 214 trains/day assume random arrival Vehicle driver unable to move to a safe position Red 2: 10% Red 3: 50% AND Train / road vehicle collision 1 in 23 years
Conclusions and next steps VAS reduces the number of blocking back events Events per day Serious (Red) events normalised by opportunity for queuing Red 1 events - sustained reduction over time More data to analyse.. Next steps: Identify range of factors that contribute to blocking back Generalise findings to wider crossing population Enhance risk model to account for the above