Reinforcement Learning for Robust Athletic Intelligence: Lessons from the 2nd 'AI Olympics with RealAIGym' Competition
On the map: Deep reinforcement learning methods → RL for robot locomotion and manipulation
Abstract (arXiv)
In the field of robotics many different approaches ranging from classical planning over optimal control to reinforcement learning (RL) are developed and borrowed from other fields to achieve reliable control in diverse tasks. In order to get a clear understanding of their individual strengths and weaknesses and their applicability in real world robotic scenarios is it important to benchmark and compare their performances not only in a simulation but also on real hardware. The '2nd AI Olympics with RealAIGym' competition was held at the IROS 2024 conference to contribute to this cause and evaluate different controllers according to their ability to solve a dynamic control problem on an underactuated double pendulum system with chaotic dynamics. This paper describes the four different RL methods submitted by the participating teams, presents their performance in the swing-up task on a real double pendulum, measured against various criteria, and discusses their transferability from simulation to real hardware and their robustness to external disturbances.
Key facts. Cited by 2 papers 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 13 of its 25 references that are in the map. First posted on arXiv 19 Mar 2025.
Builds on
13 of 25 references.
Closest ancestors · 1
Same direction, newest first
Other references · 12
- 2017Proximal Policy Optimization Algorithmscited by 15,215
- 2018Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actorcited by 4,885
- 2015High-Dimensional Continuous Control Using Generalized Advantage Estimationcited by 2,026
- 2019RLBench: The Robot Learning Benchmark & Learning Environmentcited by 689
- 2019An Open Torque-Controlled Modular Robot Architecture for Legged Locomotion Researchcited by 22
- 2020Average-reward model-free reinforcement learning: a systematic review and literature mappingcited by 22
- 2021Benchmarking Structured Policies and Policy Optimization for Real-World Dexterous Object Manipulationcited by 15
- 2021Model-Based Policy Search Using Monte Carlo Gradient Estimation with Real Systems Applicationcited by 10
- 2023AcroMonk: A Minimalist Underactuated Brachiating Robotcited by 4
- 2024Maximum Entropy On-Policy Actor-Critic via Entropy Advantage Estimationcited by 4
- 2024RicMonk: A Three-Link Brachiation Robot with Passive Grippers for Energy-Efficient Brachiationcited by 2
- 2024AI Olympics challenge with Evolutionary Soft Actor Criticcited by 2
Built on by
Citing papers per year: 2025: 2.
2 papers.
Similar, not cited
Close in wording, no citation either way
- 2024Average-Reward Maximum Entropy Reinforcement Learning for Underactuated Double Pendulum Tasks
- 2024Learning control of underactuated double pendulum with Model-Based Reinforcement Learning
- 2024Reinforcement Learning Within the Classical Robotics Stack: A Case Study in Robot Soccer
- 2024Technical Report on Reinforcement Learning Control on the Lucas-N"ulle Inverted Pendulum
- 2025Real-world Reinforcement Learning from Suboptimal Interventions
- 2024Velocity-History-Based Soft Actor-Critic Tackling IROS'24 Competition "AI Olympics with RealAIGym"
- 2025Gym4ReaL: A Suite for Benchmarking Real-World Reinforcement Learning
- 2023Real Robot Challenge 2022: Learning Dexterous Manipulation from Offline Data in the Real World