What Is The Collision Probability And How To Compute It

We revisit the computation of probability of collision in the context of automotive collision avoidance (the estimation of a potential collision is also referred to as conflict detection in other contexts). After reviewing existing approaches to the definition and computation of collision probability we argue that the fundamental quantity that is required to answer a question like ”What is the probability of collision within the next three seconds?” is a collision probability rate. We derive a general expression for the distribution of the collision probability rate and demonstrate that it exactly reproduces distributions obtained by large-scale Monte-Carlo simulations. This expression is valid for arbitrary prediction models including process noise. We derive an approximation for the distribution of the collision probability rate that can be computed on an embedded platform. In order to efficiently sample this probability rate distribution for determination of its characteristic shape an adaptive method to obtain the sampling points is proposed. The probability of collision is then obtained by one-dimensional numerical integration over the time period of interest. We also argue that temporal collision measures such as time-to-collision should not be calculated as separate or even prerequisite quantities but that they are properties of the distribution of the collision probability rate.