Reset-free Trial-and-Error Learning for Data-Efficient Robot Damage Recovery
Explore this paper's citation graph
Summary
This paper introduces a novel learning algorithm called “Reset-free Trial-and-Error” (RTE) that allows robots to recover from damage while completing their tasks and makes it possible to contemplate sending robots to places that are truly too dangerous for humans and in which robots cannot be rescued.
- Type
- article
- Published
- 2016-10-13
- Cited by
- 11
- References
- 44
- Access
- Open access
- OpenAlex
- https://openalex.org/W2536782980
- Semantic Scholar
- https://api.semanticscholar.org/CorpusID:266855
Keywords
Reset (finance), Computer science, Robot, Artificial intelligence, Economics
References
- Monte Carlo Tree Search for Continuous and Stochastic Sequential Decision Making Problems. (Monte Carlo Tree Search pour les problèmes de décision séquentielle en milieu continus et stochastiques)
- Illuminating search spaces by mapping elites
- Model learning for robot control: a survey
- Automatic Gait Optimization with Gaussian Process Regression
- A Unifying View of Sparse Approximate Gaussian Process Regression
- Efficient reinforcement learning for robots using informative simulated priors
- Robots that can adapt like animals
- Deep neural networks are easily fooled: High confidence predictions for unrecognizable images
- Sferesv2: Evolvin' in the multi-core world
- On building systems that will fail
- Reinforcement learning in robotics: A survey
- Fault-tolerant gait learning and morphology optimization of a polymorphic walking robot
- Generation of whole-body optimal dynamic multi-contact motions
- Resilient Machines Through Continuous Self-Modeling
- Active learning of inverse models with intrinsically motivated goal exploration in robots
- Evolving a Behavioral Repertoire for a Walking Robot
- Gaussian Processes for Data-Efficient Learning in Robotics and Control
- Fast damage recovery in robotics with the T-resilience algorithm
- Robot Skill Learning: From Reinforcement Learning to Evolution Strategies
- An experimental comparison of Bayesian optimization for bipedal locomotion
Cited by
- Scaling Up MAP-Elites Using Centroidal Voronoi Tessellations
- Black-box data-efficient policy search for robotics
- 20 years of reality gap: a few thoughts about simulators in evolutionary robotics
- Evolution of Repertoire-Based Control for Robots With Complex Locomotor Systems
- Chairs' welcome for GECCO'17 workshop "evolution in cognition"
- Using Centroidal Voronoi Tessellations to Scale Up the Multidimensional Archive of Phenotypic Elites Algorithm
- Map-based Multi-Policy Reinforcement Learning: Enhancing Adaptability of Robots by Deep Reinforcement Learning
- Technical Report on: Tripedal Dynamic Gaits for a Quadruped Robot
- Quality-Diversity Optimisation on a Physical Robot Through Dynamics-Aware and Reset-Free Learning
- Safety-Aware Robot Damage Recovery Using Constrained Bayesian Optimization and Simulated Priors
- A Learning Based Recovery for Damaged Snake-Like Robots
Related papers
- Soccer robot's navigating algorithm with predictive control and error correction
- Research into error recovery for sensory robots
- Visual feedback arm robot based on iterative learning control using repetitive control optimization
- Some Experiences of the use of Iterative Learning Control for Performance Improvement in Robot Control Systems
- Adaptive Iterative Learning Control for Robot Manipulators with Initial Resetting Errors
- Study on Delay and Accuracy of Control System for the Vascular Interventional Surgical Robot based on Fuzzy PID and Improved Smith Algorithms
- A Neural Network Based Learning Robot Manipulators Controller for
- An iterative learning scheme for motion control of robots using neural networks: A case study