Researchers at the BioRobotics Institute and the National University of Singapore have developed plant-inspired controllers for improved robotic arm performance in real-world environments.
Robotic systems often mimic nature, emulating biological processes, structures, and behaviors to improve performance beyond lab environments. Drawing inspiration from animals and plants, these systems leverage innate abilities for survival, enhancing their effectiveness in real-world applications.
Researchers at the Brain-Inspired Robotics (BRAIR) Lab, BioRobotics Institute of Sant’Anna School of Advanced Study in Italy and the National University of Singapore have created a plant-inspired controller to enhance robotic arm performance in real-world environments. Soft robot arms mimic boneless organisms like octopus tentacles, elephant trunks, and plants. They use flexible materials for compliant and dexterous motion, conforming to surfaces and ensuring human safety at a low cost. Soft robot arms are valuable for automating inaccessible tasks, and researchers aim to develop effective controllers for these flexible arms.
The team have developed a plant-inspired controller for soft robot arms to excel in real-world settings, going beyond the limitations of lab-focused controllers. Plants utilize purposeful movement strategies based on growth, allowing them to colonize diverse habitats. Unlike animals, plant movement doesn’t rely on a central nervous system but on decentralized computing mechanisms. The researchers’ controller emulates decentralized plant movements using behavior-based artificial intelligence (AI) tools with bottom-up structures. The arm consists of redundant soft modules activated by radially arranged actuators capable of generating six principle bending directions.
The controller’s computing agents exploit actuator configuration to replicate plant movements, specifically circumnutation (plant oscillations) and phototropism (movement towards light). The controller enables sequential control of robotic arms in two stages: exploration and reaching. Moreover, the controller can be applied to any soft robot arm with a similar actuation arrangement, paving the way for embedded sensing and distributed control strategies in continuum and soft robots. The researchers tested their controller on a lightweight, modular, cable-driven soft robotic arm with 9 degrees of freedom (9-DoF). Promising results showed improved exploration and reaching compared to previous control strategies.
The controller may be applied to other soft robotic arms for testing in various settings. The team aims to expand its capabilities for complex behaviors like target tracking and whole-arm twining.
Reference: Enrico Donato et al, Plant-inspired behavior-based controller to enable reaching in redundant continuum robot arms, 2023 IEEE International Conference on Soft Robotics (RoboSoft) (2023). DOI: 10.1109/RoboSoft55895.2023.10122017.
