“…The problem of line-of-sight indoor localisation was irst resolved through the matching of cloud point data (obtained from a line-ofsight sensor such as a Li-Dar) to retrieve tuple (x, y, θ ) describing the location and orientation of a robot in a known environment. This was irst achieved by algorithms such as the Iterative Closest Point (ICP) algorithm Besl and McKay [4], and a long line of alternative heuristic algorithms Lu and Milios [23], Diosi and Kleeman [11] [29], Biber and Strasser [5], Donoso-Aguirre et al [13], Konecny et al [19] and various improvements on the ICP's convergence speed [12] [32], dataset optimisation [33] [25] or precision metrics [14]. Performing indoor localisation without a priori knowledge of the robot's pose increases the diiculty to this problem, as a global search for the position must now be performed, rather than simply a pose reinement.…”