Morphing-Enabled Path Planning for Flying Tensegrity Robots as a Semidefinite Program

Area of Science:

• Robotics
• Control Systems
• Computational Geometry

Background:

• Deformable drones, particularly those based on tensegrity structures, offer unique advantages but pose significant challenges in path planning.
• Efficiently navigating complex configurations is crucial for the practical application of these advanced robotic systems.

Purpose of the Study:

• To develop an efficient configuration-space path planning method for tensegrity-based deformable drones.
• To address the computational complexity associated with planning for robots with continuously variable shapes.

Main Methods:

• A simplified shape encoding is proposed to represent the drone's configuration space.
• Path planning is transformed into a sequence of semidefinite programs.
• A data-driven method, utilizing pre-computed stable configurations and their Löwner-John ellipsoids, facilitates rapid collision and containment checks.

Main Results:

• The proposed method significantly speeds up path planning by offloading intensive computations to an offline dataset generation phase.
• Containment checks exhibit a computational cost that scales linearly with the size of the pre-computed dataset.
• The approach effectively bridges the gap between simplified shape descriptions and complex tensegrity configurations.

Conclusions:

• The developed method offers an efficient solution for path planning in deformable tensegrity drones.
• This approach enhances the practical viability of tensegrity robots by addressing key computational challenges in motion planning.
• The integration of data-driven techniques with semidefinite programming provides a scalable framework for complex robotic systems.