iDRM: Finding Stance Poses and Collision Free Configurations for Humanoid Robots in Complex Environments

Finding valid stance pose and collision–free configuration is a crucial problem for many humanoid robot applications such as providing goal states for walking planners and motion planners. However, finding such stance configurations is non-trivial in cluttered and complex environments, where some standing locations and reaching postures could be blocked by obstacles. We introduce a new approach, namely the inverse Dynamic Reachability Map (iDRM) which allows a robot to find valid stance poses and collision–free configurations in complex and changing environments in real time. We extend the inverse reachability map (IRM) approach by introducing information about how the robot occupies the workspace. This allows efficient online reconstruction of the IRM in different environments. We have evaluated the performance of our method in a variety of reaching tasks using the 38 degree-of-freedom (DOF) NASA Valkyrie humanoid robot. Our results show that the approach is capable of finding valid stance poses and collision–free configurations in a fraction of a second.