Development of Simulation Platform for Human-Robot-Environment Interface in the UFES CloudWalker

Techniques for Human-Robot-Environment Interaction allow sharing control between assistive devices, such as Smart Walkers, and its users by taking into account dynamic environments composed of objects and people. This work presents a simulation platform to accelerate the development and evaluation of such interaction strategies. To enhance the physiotherapists’ control during lower-limb rehabilitation therapies, we employ a map-based strategy to map and allow the customization of pre-defined paths. Inspired by hospital layouts, a simulation environment was created to test and evaluate our system, and two experiments were conducted. In the first one, an accuracy analysis of generated maps was performed in comparison with the actual hospital environment. In the second experiment, we evaluated the avoidance obstacles strategy by positioning unknown objects along the Smart Walker targets. In the first experiment we obtained percentual errors around 2.658 and 8.356% between the original and mapped segments. In the second experiment, besides reached all the targets without collisions the position errors were between 0.062 and 0.118 m and the orientation errors, 0.024 and 0.044 radians. These results indicating the feasibility of employing such methodology for the development of the proposed interaction strategy.