Collision Probability Estimation For Small Unmanned Aircraft Systems

With the continuous expansion of the small unmanned aircraft market, the collision risk problem for small unmanned aircraft systems (sUAS) has been increasingly highlighted. In this paper, the rapid calculating methods of collision probability estimation for sUAS are proposed supposing that the predicted position error follows a certain Gaussian distribution. There are three types of collision zones established for collision modelling including cuboid, ellipsoid and cylinder. For each type of collision zones, a corresponding algorithm is derived for collision probability estimation based on the univariate conditioning or Laguerre polynomials. Randomized tests are conducted to validate the effectiveness of the proposed methods. The test results indicate that the average computation times of these three methods are approximately two orders of magnitude faster than the corresponding exact solutions, while the average computation errors are all less than 1%. Numerical simulations are carried out to analyze the differences in the collision probabilities when using different types of collision zones. The simulation results testify that the selection of the collision zones affects the collision probability estimation apparently, especially when the crossing angle of the two sUAS is between 0 degrees and 40 degrees.