mlsanityData to 14 Oct 2026

AgenticRL: Agentic Reinforcement Learning with Self-Refinement for Complex UAV Navigation

arXiv 2606.03963 ↗First posted 2 Jun 2026cs.RO cs.AIcited by 0

On the map: LLM agents and tool use → Reinforcement learning for LLM agents

Abstract (arXiv)

Deep reinforcement learning enables autonomous robots to learn complex navigation tasks, but still relies heavily on time consuming manual reward design and fine tuning. Existing automated reward generation and refinement methods reduce this effort, yet often lack task-level behavioral diagnosis for directing subsequent reward revisions. We introduce AgenticRL, a multimodal closed loop framework in which role-specialized agents generate executable rewards, diagnose failures of the resulting policies, formulate targeted refinement instructions, and regenerate improved rewards. Before training, a task grounding stage automatically selects a compatible action profile, together with its observation and reward interfaces. Each generated reward is used to train a policy using Proximal Policy Optimization (PPO), which is subsequently evaluated under randomized conditions. Task-level behavioral, geometric, and safety measurements are organized into a structured diagnosis packet and jointly analyzed with the current reward code, task specification, behavioral summary, and visual scene context. Unlike one-shot reward generation, human-guided refinement, or broad candidate search, AgenticRL uses automated diagnosis of the behavior induced by a reward to direct its next revision. We evaluate the framework across eight UAV tasks covering navigation, obstacle interaction, trajectory tracking, agile manoeuvres, and cluttered flight. Under the reported comparative evaluation, AgenticRL achieves success rates of 100% in racing and 88% in cluttered navigation, exceeding the strongest Eureka and Text2Reward baselines, respectively. Reward refinement increases mean simulation success from 37.2% to 96.4%, while the resulting policies achieve a collective real-world success rate of 90.0% and a sim-to-real accuracy of 93.4%.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Placed in Reinforcement learning for LLM agents, one of 16 directions in LLM agents and tool use. Builds on 19 of its 28 references that are in the map. First posted on arXiv 2 Jun 2026.

Builds on

19 of 28 references.

Closest ancestors · 6

Same direction, newest first

  1. 2026Exploring Reasoning Reward Model for Agents29 Jan
  2. 2025OpenReward: Learning to Reward Long-form Agentic Tasks via Reinforcement Learning28 Oct
  3. 2025AgentRL: Scaling Agentic Reinforcement Learning with a Multi-Turn, Multi-Task Framework5 Oct
  4. 2025Agentic Reinforcement Learning with Implicit Step Rewards23 Sep
  5. 2025The Landscape of Agentic Reinforcement Learning for LLMs: A Survey2 Sep
  6. 2025Agent-RLVR: Training Software Engineering Agents via Guidance and Environment Rewards13 Jun

Other references · 13

  1. 2017Proximal Policy Optimization Algorithmscited by 15,215
  2. 2017Deep reinforcement learning from human preferencescited by 3,912
  3. 2022Code as Policies: Language Model Programs for Embodied Controlcited by 1,210
  4. 2023Reward Design with Language Modelscited by 261
  5. 2023Text2Reward: Reward Shaping with Language Models for Reinforcement Learningcited by 160
  6. 2021Learning to Fly -- a Gym Environment with PyBullet Physics for Reinforcement Learning of Multi-agent Quadcopter Controlcited by 77
  7. 2020Active Preference-Based Gaussian Process Regression for Reward Learningcited by 71
  8. 2025ReWiND: Language-Guided Rewards Teach Robot Policies without New Demonstrationscited by 60
  9. 2026RoboReward: General-Purpose Vision-Language Reward Models for Roboticscited by 53
  10. 2024Code as Reward: Empowering Reinforcement Learning with VLMscited by 39
  11. 2026LiteResearcher: A Scalable Agentic RL Training Framework for Deep Research Agentcited by 10
  12. 2023Sim-to-Real Deep Reinforcement Learning based Obstacle Avoidance for UAVs under Measurement Uncertaintycited by 5
1 more
  1. 2022Reinforcement Learning for UAV control with Policy and Reward Shapingcited by 1

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2026FARL: A Fully Automated Pipeline of RL Interface and Solver Design
  2. 2026Beyond Scripted Search: Sample-Efficient Reward Discovery via Agentic Black-box Optimization
  3. 2026HARBOR: A Harness Framework for Agentic Robot Reinforcement Learning
  4. 2026ARS: Agentic Reward System for Robot Learning
  5. 2026Learning to Explore: Scaling Agentic Reasoning via Exploration-Aware Policy Optimization
  6. 2026Decoupling Task and Behavior: A Two-Stage Reward Curriculum in Reinforcement Learning for Robotics
  7. 2026GzDRL: Reproducible and Scalable Deep Reinforcement Learning with Gazebo
  8. 2026AgentV-RL: Scaling Reward Modeling with Agentic Verifier