Simultaneous Adaptive Path planning system for the real world application

In this paper, we propose a path planning system named SAP (simultaneous adaptive path planning) which can make a plan adapt to kinematic and dynamic constraints, and dynamically changing environment simultaneously. SAP has three key issues: condition layers, a dynamic space, and a search tree. Each condition layer involves a factor which influences running of the robot such as coefficient of friction and elevation of the field. Dynamic space is an area in velocity domain where the robot could reach in the next time step. The limitation (means outer shape) of the dynamic space is defined as a function of all conditions and a robot model, velocities that can be achieved within a short time interval. Search tree that consider the dynamic space expands like tree to search a path from start to goal. By these mechanisms, SAP can generate a path that robot can achieve in real environment.