In-Vehicle Collision Avoidance Support Under Adverse Visibility Conditions
Wiel Janssen TNO Human Factors Research Institute, Soesterberg, The Netherlands
British Aerospace, Bristol, UK
Advanced sensor technology for road traffic promises a reduction of accident risk, which is to be achieved by means of so-called in-vehicle collision avoidance systems (CAS). The basis lies in the conviction that it should be possible to support human drivers by technical means in avoiding an impending collision that is unnoticed or perhaps not noticeable at all.
A CAS recognizes critical configurations just before they actually happen, after which it initiates some form of corrective action. The action may be a warning to the driver or (the beginning of) a specific evasive action itself. This chapter deals with so-called longitudinal collision avoidance systems, where the target is a preceding vehicle in the same lane. The sensor technology for this configuration, and therefore the potential for future application, is relatively well developed. Before considering CAS design issues in more detail, it seems wise to reflect on what it actually hopes to achieve.
The average Western automobile driver experiences a rear-end collision--the case to which the current generation of sensors has most to offer--about once every 25 to 30 years. If narrow escapes are also counted as events to be avoided, then support would, of course, be needed somewhat more frequently. Nevertheless, it is clear that the event a CAS should detect is so rare that serious doubts should be entertained regarding the possibility that detection of critical configurations could ever be performed flawlessly, let alone that it could be achieved without false alarms. Given that this is the case, it is probably better to take a modest stance on what