Abstract
The need for combined task and motion planning (CTAMP) in robotics is well known as robotic technologies become more mature. The goal of CTAMP is to determine a proper sequence of a robot’s actions based on symbolic and geometric reasoning. Because of the fundamental difference in symbolic and geometric reasoning, a CTAMP system often requires an interface module between the two reasoning modules. We propose a CTAMP system in which a symbolic action sequence is generated in task planning, and each action is verified geometrically in motion planning using the off-the-shelf planners and reasoners. The approach is that a set of action models is defined with PDDL in the interface module (action library) and the required information to each planner is automatically provided by the interface module. The proposed method was successfully implemented in three simulated experiments that involve manipulation tasks. According to our findings, the proposed method is effective in responding to changes in the environment and uncertainty with errors in recognition of the environment and the robot motion control.
Highlights
A service robot requires a system to manipulate objects indoors
By ignoring the feasibility of the actions and considering only the causality between them, the task planner automatically generates a sequence of actions that can reach the goal state from the initial state of the task
Srivastava et al (2014) used geometric parameters such as grasp postures or object positions in the action model in advance to construct the predicates of the action precondition so that the motion planning can interfere with the task planning
Summary
A service robot requires a system to manipulate objects indoors (e.g., pour drinks). This system must allow the robot to plan several actions in order, without collisions with other objects: 1) move the robot base near the object to be manipulated; 2) move the robot arm around the object; and 3) grasp the object. Srivastava et al (2014) used geometric parameters such as grasp postures or object positions in the action model in advance to construct the predicates of the action precondition so that the motion planning can interfere with the task planning They did task planning first and implemented an interface layer to call the motion planner for each action. They take a long time to solve to relocate obstacles because they did not consider linking with additional geometric reasoning modules to make more efficient plans (Lee and Kim, 2019). In the proposed CTAMP system, when the sequence of primitive action was first determined by the task planner, we proposed an interference method, which automatically calls the motion planners and geometric reasoners required for the execution of each primitive action.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have