Path Planning based on Probabilistic Foam for a Robotic Smart Walker in Dynamic Environments
Abstract
Human walking can be affected by several factors, such as aging, neurological or neuromuscular diseases, spinal cord or lower limb injuries, etc., which directly impact on quality of life. Different devices such as canes, crutches, orthoses and walkers are currently used to assist patients that suffer mobility problems. In the last decade, new robotic devices such as exoskeletons and smart walkers have been developed to intelligently assist the gait of this kind of patient. In particular, the incorporation of various sensors and the use of robotic perception and navigation techniques in smart walkers made it possible to improve mobility, providing walking assistance, preventing falls and improving safety in locomotion. Furthermore, intelligent robotic walkers made it possible to optimize and personalize rehabilitation therapies according to the specific characteristics of each patient. In this context, safe navigation in environments with static and dynamic obstacles is an important aspect of smart walkers. This functionality can be achieved using path planning techniques commonly used in mobile robotics applications. This paper presents an improved version of the Probabilistic Foam Method (PFM) applied to path planning of an intelligent robotic walker in dynamic environments. The Probabilistic Foam method allows the planning of paths with a safe region around them, which is suitable for applications where safe navigation is important. In this article, the probabilistic foam method, originally proposed for static environments, is adapted to take into account the presence of moving obstacles, such as people moving close to the intelligent robotic walker. The proposed version follows a deliberative approach, where the temporal dimension is added to the Configuration Space. This allows kinodynamic features to be incorporated naturally into path planning. In order to validate the proposed path planning method in dynamic environment, simulation results for the Ambulus robotic smart walker prototype are presented.
