Constructing informative Bayesian map priors: a multi-objective optimisation approach applied to indoor occupancy grid mapping

The problem of simultaneous localisation and mapping (SLAM) has been addressed in numerous ways with different approaches aiming to produce faster, more robust solutions that yield consistent maps. This focus, however, has resulted in a number of solutions that perform poorly in challenging real life scenarios. In order to achieve improved performance and map quality this article proposes a novel method to construct informative Bayesian mapping priors through a multi-objective optimisation of prior map design variables defined using a source of prior information. This concept is explored for 2D occupancy grid SLAM, constructing such priors by extracting structural information from architectural drawings and identifying optimised prior values to assign to detected walls and empty space. Using the proposed method a contextual optimised prior can be constructed. This prior is found to yield better quantitative and qualitative performance than the commonly used non-informative prior, yielding an increase of over 20% in the F2 metric. This is achieved without adding to the computational complexity of the SLAM algorithm, making it a good fit for time critical real life applications such as search and rescue missions.