The vision of ground robots seamlessly operating among humans requires novel approaches which enhance classic robot motion planning by accounting for interaction effects and social preferences. Within our group we address social motion planning using learning from demonstrations, game theory and optimal control.