Planning in constraint space for multi-body manipulation tasks

Robotics;Structure design;Kinodynamic limitations

Erdogan, Can ; Christensen, Henrik Dellaert, Frank Computer Science Egerstedt, Magnus Bobick, Aaron Lozano-Perez, Tomas Kuffner, James ; Christensen, Henrik

University:Georgia Institute of Technology

Department:Computer Science

关键词: Robotics;    Structure design;    Kinodynamic limitations;   

Others  :  https://smartech.gatech.edu/bitstream/1853/54978/1/ERDOGAN-DISSERTATION-2016.pdf

美国|英语

来源: SMARTech Repository

【 摘 要 】

Robots are inherently limited by physical constraints on their link lengths, motor torques, batterypower and structural rigidity. To thrive in circumstances that push these limits, such as in searchand rescue scenarios, intelligent agents can use the available objects in their environment astools. Reasoning about arbitrary objects and how they can be placed together to create usefulstructures such as ramps, bridges or simple machines is critical to push beyond one's physicallimitations. Unfortunately, the solution space is combinatorial in the number of available objectsand the configuration space of the chosen objects and the robot that uses the structure is highdimensional.To address these challenges, we propose using constraint satisfaction as a means to test thefeasibility of candidate structures and adopt search algorithms in the classical planning literatureto find sufficient designs. The key idea is that the interactions between the components of astructure can be encoded as equality and inequality constraints on the configuration spaces of therespective objects. Furthermore, constraints that are induced by a broadly defined action, such asplacing an object on another, can be grouped together using logical representations such as PlanningDomain Definition Language (PDDL). Then, a classical planning search algorithm can reason aboutwhich set of constraints to impose on the available objects, iteratively creating a structure thatsatisfies the task goals and the robot constraints. To demonstrate the effectiveness of thisframework, we present both simulation and real robot results with static structures such as ramps,bridges and stairs, and quasi-static structures such as lever-fulcrum simple machines.