Paper accepted (RA-L): "Solving footstep planning as a feasibility problem using L1-norm minimization"

This paper is an extension to the SL1M conference paper presented at ICRA 2020 (see below). We propose to combine a low-dimensional sampling-based planner with SL1M to reduce the combinatorics of large footstep planning problems, which results in a significant improvement of the computation times with respect to a MIP solver. We also present an analysis of the behaviour of Mixed Integer solvers on footstep planning problems and demonstrate the framework on new examples, including ANYmal.

More information on the project page. The paper is available here.