We review the use of factor graphs for the modeling and solving of large-scale inference problems in robotics. Factor graphs are a family of probabilistic graphical models, other examples of which are Bayesian networks and Markov random fields, well known from the statistical modeling and machine le