eCall, the pan-European automatic crash notification system, will facilitate road vehicles to contact emergency services autonomously when a potentially injurious crash has been detected by vehicle sensors. Type-approval requirements will set out conditions for assessing systems under which automatic triggering of eCalls will be mandatory. Research is needed to specify the accident typologies and severities represented by these conditions.
This paper analyses what definition of accident conditions would ensure that a high number of casualties benefit from automatic eCalls. The conditions should also allow cost-effective type-approval testing, avoid excessive numbers of superfluous eCalls, and not restrict manufacturer’s design freedom.
Two conditions were considered as being particularly suitable for the European type-approval system:
In-depth accident data from the Road Accident In-Depth Studies (RAIDS) database, collected between 2000 and 2010 for the Co-operative Crash Injury Study (CCIS), was analysed to produce an estimate of the proportion of car occupant casualties captured by each of these conditions and subsequently scaled to a national level for Great Britain.
The analysis found that Condition A captured only 34.7% of fatally and seriously injured casualties whereas Condition B would apply to 81.0%. For Great Britain, with about 9,335 fatally or seriously injured car occupants annually, this is a difference of 4,330 fatal or serious casualties which could benefit from automatic eCall triggering each year. However, if Condition B was applied, automatic eCalls would be triggered for 74,390 slight casualties per annum in GB (and for an additional unknown number of damage-only accidents).
The sensitivity of Condition B, i.e. the proportion of casualties successfully selected, is considerably higher compared to Condition A. Nevertheless, accident types where airbags are deliberately not deployed would not be captured. Condition B exhibited an almost unvaryingly high sensitivity in selecting fatal casualties and serious casualties respectively.
The higher sensitivity of Condition B is achieved at the expense of specificity in selecting fatal or severe casualties, i.e. more of the collisions for which an eCall is triggered would be slight casualties. There are indications, however, that the negative consequences of superfluous eCalls could be mitigated.
The results are based on British data and cannot be transferred in detail to other countries. The general trends identified would be expected to also appear in reproductions of this analysis using data from other European countries.
The most preferable mandatory automatic eCall triggering condition for type-approval legislation appears to be triggering in conjunction with deployment of any airbag. Nevertheless, up to 19.0% of fatal and serious car occupant casualties might not be captured by this condition. To allow this problem to be overcome using advanced triggering algorithms, a non-restrictive approach could be taken with regard to the triggering requirement, i.e. require triggering in the presence of the condition yet not prohibit triggering in its absence.