On the Fly Adaptation of Behavior Tree-Based Policies through Reinforcement Learning
On the map: Deep reinforcement learning methods → RL for robot locomotion and manipulation
Abstract (arXiv)
With the rising demand for flexible manufacturing, robots are increasingly expected to operate in dynamic environments where local -- such as slight offsets or size differences in workpieces -- are common. We propose to address the problem of adapting robot behaviors to these task variations with a sample-efficient hierarchical reinforcement learning approach adapting Behavior Tree (BT)-based policies. We maintain the core BT properties as an interpretable, modular framework for structuring reactive behaviors, but extend their use beyond static tasks by inherently accommodating local task variations. To show the efficiency and effectiveness of our approach, we conduct experiments both in simulation and on a Franka Emika Panda 7-DoF, with the manipulator adapting to different obstacle avoidance and pivoting tasks.
Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Placed in RL for robot locomotion and manipulation, one of 15 directions in Deep reinforcement learning methods. Builds on 18 of its 28 references that are in the map. First posted on arXiv 8 Mar 2025.
Builds on
18 of 28 references.
References · 18
- 2018Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actorcited by 4,885
- 2015Contextual Markov Decision Processescited by 220
- 2021Multi-Task Reinforcement Learning with Context-based Representationscited by 174
- 2017Behavior Trees in Robotics and AI: An Introductioncited by 107
- 2017Markov Decision Processes with Continuous Side Informationcited by 56
- 2020A Survey of Behavior Trees in Robotics and AIcited by 55
- 2022Contextualize Me -- The Case for Context in Reinforcement Learningcited by 46
- 2021Augmented World Models Facilitate Zero-Shot Dynamics Generalization From a Single Offline Environmentcited by 42
- 2018A survey on policy search algorithms for learning robot controllers in a handful of trialscited by 35
- 2022Learning to Grasp the Ungraspable with Emergent Extrinsic Dexteritycited by 28
- 2023Dynamics Generalisation in Reinforcement Learning via Adaptive Context-Aware Policiescited by 20
- 2017Sampling-based Planning of In-Hand Manipulation with External Pushescited by 8
6 more
- 2023BeBOP -- Combining Reactive Planning and Bayesian Optimization to Solve Robotic Manipulation Taskscited by 6
- 2023Learning to Adapt the Parameters of Behavior Trees and Motion Generators (BTMGs) to Task Variationscited by 5
- 2023Learning Extrinsic Dexterity with Parameterized Manipulation Primitivescited by 5
- 2024Inferring Behavior-Specific Context Improves Zero-Shot Generalization in Reinforcement Learningcited by 5
- 2021Learning of Parameters in Behavior Trees for Movement Skillscited by 4
- 2022A Stack-of-Tasks Approach Combined with Behavior Trees: a New Framework for Robot Controlcited by 2
Built on by
Citing papers per year: 2026: 1.
1 paper.
Similar, not cited
Close in wording, no citation either way
- 2024Adaptable Recovery Behaviors in Robotics: A Behavior Trees and Motion Generators(BTMG) Approach for Failure Management
- 2026Learning from Imperfect Demonstrations via Temporal Behavior Tree-Guided Trajectory Repair
- 2024HBTP: Heuristic Behavior Tree Planning with Large Language Model Reasoning
- 2025Video-to-BT: Generating Reactive Behavior Trees from Human Demonstration Videos for Robotic Assembly
- 2020Supervised Learning and Reinforcement Learning of Feedback Models for Reactive Behaviors: Tactile Feedback Testbed
- 2024Guided Reinforcement Learning for Robust Multi-Contact Loco-Manipulation
- 2026Progress Constraints for Reinforcement Learning in Behavior Trees
- 2026Safety-aware Skill Adaptation for Reinforcement Learning in Dynamic Environments