Photo credit: science.nasa.gov

Innovative Robotic Projects Addressing Advanced Autonomy

Introduction

Recent advancements in robotics focus on enhancing autonomy and performance in challenging environments. Various institutions are leading projects aimed at improving robotic capabilities for scientific exploration. This article highlights key initiatives and their objectives, demonstrating how researchers are leveraging cutting-edge techniques to tackle complex problems in robotic operations.

ARROW Projects Overview

Jonathan Bohren – Honeybee Robotics

Project: Stochastic PLEXIL (SPLEXIL)
Testbed Used: OceanWATERS
Purpose: This project aims to extend the capabilities of PLEXIL (Plan Execution Interchange Language) by integrating stochastic decision-making through reinforcement learning methods.

Pooyan Jamshidi – University of South Carolina

Project: Resource Adaptive Software Purpose-Built for Extraordinary Robotic Research Yields (RASPBERRY SI)
Testbed Used: OceanWATERS & OWLAT
Purpose: The focus is on developing sophisticated software tools and algorithms that aid in identifying fault causes, enabling causal debugging, causal optimization, and verification processes for robotic systems.

COLDTech Projects Overview

Eric Dixon – Lockheed Martin

Project: Causal And Reinforcement Learning (CARL) for COLDTech
Testbed Used: OceanWATERS
Purpose: This initiative integrates a model of the Cold Operable Lunar Deployable Arm (COLDarm) and employs advanced image analysis along with machine learning to address and correct operational faults, such as issues caused by ice accumulation.

Jay McMahon – University of Colorado

Project: Robust Exploration with Autonomous Science On-board, Ranked Evaluation of Contingent Opportunities for Uninterrupted Remote Science Exploration (REASON-RECOURSE)
Testbed Used: OceanWATERS
Purpose: The project focuses on leveraging automated planning and formal methods to optimize scientific output from landers while reducing dependency on communication with the ground team.

Melkior Ornik – University of Illinois, Urbana-Champaign

Project: aDaptive, ResIlient Learning-enabLed oceAn World AutonomY (DRILLAWAY)
Testbed Used: OceanWATERS & OWLAT
Purpose: This research emphasizes autonomous learning and adaptation in unfamiliar terrains, utilizing previously acquired knowledge to perform efficient scooping actions despite limited experience.

Joel Burdick – Caltech

Project: Robust, Explainable Autonomy for Scientific Icy Moon Operations (REASIMO)
Testbed Used: OceanWATERS & OWLAT
Purpose: This initiative focuses on developing autonomous systems capable of detecting off-nominal conditions and implementing recovery procedures while also selecting optimal samples for scientific analysis.

Conclusion

The projects outlined here illustrate significant strides in robotics, highlighting the integration of machine learning, causal reasoning, and autonomous decision-making. As researchers continue to innovate, these advancements hold the potential to revolutionize robotic operations in various settings, particularly in extraterrestrial environments where traditional methods are not as effective. The emphasis on resilience and adaptability will likely play a critical role in the future of robotic exploration and research.

Source
science.nasa.gov