The field of robotics is moving towards more autonomous and task-execution capable systems. Recent developments focus on integrating large language models, knowledge graphs, and vision-language-action frameworks to improve robot reasoning, planning, and natural language interaction. Noteworthy papers include Learn from What We HAVE, which introduces a novel History-Aware VErifier to disambiguate uncertain scenarios online, and ConceptBot, which combines Large Language Models and Knowledge Graphs to generate feasible and risk-aware plans. Additionally, papers like Robix and FPC-VLA propose unified models for robot interaction, reasoning, and planning, and frameworks with supervisors for failure prediction and correction, respectively. These advancements aim to improve the reliability and generalization of robotic systems in unstructured environments.
Advancements in Robot Autonomy and Task Execution
Sources
Learn from What We HAVE: History-Aware VErifier that Reasons about Past Interactions Online
ConceptBot: Enhancing Robot's Autonomy through Task Decomposition with Large Language Models and Knowledge Graph
Galaxea Open-World Dataset and G0 Dual-System VLA Model
Robix: A Unified Model for Robot Interaction, Reasoning and Planning
Fail2Progress: Learning from Real-World Robot Failures with Stein Variational Inference
Generalizable Skill Learning for Construction Robots with Crowdsourced Natural Language Instructions, Composable Skills Standardization, and Large Language Model
FPC-VLA: A Vision-Language-Action Framework with a Supervisor for Failure Prediction and Correction
RoboChemist: Long-Horizon and Safety-Compliant Robotic Chemical Experimentation