Robots that can adapt like animals
Open accessThe DOI matches its Crossref record; 39 of the 40 references it cites were found in a registry or an index.
Abstract
Robots have transformed many industries, most notably manufacturing, and have the power to deliver tremendous benefits to society, such as in search and rescue, disaster response, health care and transportation. They are also invaluable tools for scientific exploration in environments inaccessible to humans, from distant planets to deep oceans. A major obstacle to their widespread adoption in more complex environments outside factories is their fragility. Whereas animals can quickly adapt to injuries, current robots cannot 'think outside the box' to find a compensatory behaviour when they are damaged: they are limited to their pre-specified self-sensing abilities, can diagnose only anticipated failure modes, and require a pre-programmed contingency plan for every type of potential damage, an impracticality for complex robots. A promising approach to reducing robot fragility involves having robots learn appropriate behaviours in response to damage, but current techniques are slow even with small, constrained search spaces. Here we introduce an intelligent trial-and-error algorithm that allows robots to adapt to damage in less than two minutes in large search spaces without requiring self-diagnosis or pre-specified contingency plans. Before the robot is deployed, it uses a novel technique to create a detailed map of the space of high-performing behaviours. This map represents the robot's prior knowledge about what behaviours it can perform and their value. When the robot is damaged, it uses this prior knowledge to guide a trial-and-error learning algorithm that conducts intelligent experiments to rapidly discover a behaviour that compensates for the damage. Experiments reveal successful adaptations for a legged robot injured in five different ways, including damaged, broken, and missing legs, and for a robotic arm with joints broken in 14 different ways. This new algorithm will enable more robust, effective, autonomous robots, and may shed light on the principles that animals use to adapt to injury.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
The cited source doesn’t support the claim as written.
In the citing paper
“We render images using the MuJoCo Physics Engine's [45] built-in renderer. This renderer is not intended to be photo-realistic, and physically plausible choices of textures and lighting are not needed.”
From Domain randomization for transferring deep neural networks from simulation to the real world (2017) · cites this paper as [45]
Checked by a reasoning model — may be wrong.
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“under Quality-Diversity optimization (Pugh et al., 2016), the search space is divided into behavioral niches to evolve diverse solutions, a paradigm that challenged the dominant approach of training a single agent, as it proved more robust in real-world settings (Cully et al., 2015).”
From Dynamics of niche construction in adaptable populations evolving in diverse environments (2023) · cites this paper as Cully et al., 2015
Experimental — accuracy not yet measuredThe source backs part of the claim, not all of it — or backs it for a different population, condition or comparison.
In the citing paper
“Similarly, extensive reliance on simulation and offline training further distances learning from the full variability and constraint structure of real sensorimotor interaction (Jakobi et al., 1995; Cully et al., 2015).”
From Toward Enactive Artificial Intelligence (2026) · cites this paper as Cully et al., 2015
Experimental — accuracy not yet measured
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Gaussian Processes for Machine Learning
Rasmussen, C. E., & Williams, C. K. I. (2005) · doi:10.7551/mitpress/3206.001.0001 · Cited by 11,705
Corke, P., Roberts, J., Cunningham, J., & Hainsworth, D. (2008) · Springer Handbook of Robotics · doi:10.1007/978-3-540-30301-5_50 · Cited by 3,543
A survey of robot learning from demonstration
Argall, B. D., Chernova, S., Veloso, M., & Browning, B. (2009) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2008.10.024 · Cited by 3,227
Reinforcement learning in robotics: A survey
Kober, J., Bagnell, J. A., & Peters, J. (2013) · The International Journal of Robotics Research · doi:10.1177/0278364913495721 · Cited by 3,160 · cited 9×
“The key insight is that, whereas current learning algorithms either start with no knowledge of the search space 17 or with minimal knowledge from a few human demonstrations 17,18 , animals better understand the space of possible behaviors and their value from previous experience 19 , enabling…”
BackgroundMethodResultInfluentialDiagnosis and Fault-Tolerant Control
Blanke, M., Kinnaert, M., Lunze, J., & Staroswiecki, M. (2003) · Springer eBooks · doi:10.1007/978-3-662-05344-7 · Cited by 2,789
Nonlinear Programming: Sequential Unconstrained Minimization Techniques
Fiacco, A. V., & McCormick, G. P. (1968) · Defense Technical Information Center (DTIC) · Cited by 2,262
Stanley: The robot that won the DARPA Grand Challenge
Thrun, S., Montemerlo, M., Dahlkamp, H., Stavens, D., Aron, A., Diebel, J., Fong, P., Gale, J., Halpenny, M., Hoffmann, G., et al. (2006) · Journal of Field Robotics · doi:10.1002/rob.20147 · Cited by 2,237 · cited 2×
“3C, undam-aged: 24 [16; 41] seconds, 3 [2; 5] physical trials, repaired: 29 [16; 82] seconds, 3.5 [2; 10] trials).”
BackgroundInfluentialBrochu, E., Cora, V. M., & de Freitas, N. (2010) · arXiv (Cornell University) · doi:10.48550/arxiv.1012.2599 · Cited by 2,011
Bayesian integration in sensorimotor learning
Körding, K. P., & Wolpert, D. M. (2004) · Nature · doi:10.1038/nature02169 · Cited by 2,004 · cited 1×
“…optimization procedure followed by our robot appears similar to the technique employed by humans when they optimize an unknown function 23 , and there is strong evidence that animal brains learn probability distributions, combine them with prior knowledge, and act as Bayesian optimizers 25,26 .”
MethodControl of mental activities by internal models in the cerebellum
Ito, M. (2008) · Nature Reviews Neuroscience · doi:10.1038/nrn2332 · Cited by 1,356 · cited 1×
“A final parallel is that the simulator and Gaussian process components of Intelligent Trial and Error are two forms of predictive models, which are known to exist in animals 29,12 .”
Research works printed but not linked: 2
- Siciliano, B. & Khatib, O. Springer Handbook of Robotics (Springer, 2008). · doi:10.1007/978-3-540-30301-5
- Kluger, J. & Lovell, J. Apollo 13 (Mariner Books, 2006).
Cited by
Items 601–700 of 1,007
Multi-objective Analysis of MAP-Elites PerformancePreprint
Samuelsen, E., & Glette, K. (2018) · arXiv (Cornell University) · doi:10.48550/arxiv.1803.05174 · Cited by 6
“QD algorithms have been particularly successful in the Evolutionary Robotics (ER) field [8], with applications such as soft robot evolution [4], damage recovery [9], and locomotion repertoire generation [10], [11].”
BackgroundExploring the Dynamic Properties of Interaction in Mixed-Initiative Procedural Content Generation
Alvarez, A. (2020) · doi:10.24834/isbn.9789178771400 · Cited by 6
Mix-ME: Quality-Diversity for Multi-Agent LearningPreprint
Ingvarsson, G., Samvelyan, M., Lim, B., Flageat, M., Cully, A., & Rocktäschel, T. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2311.01829 · Cited by 6
“It has also been used in previous studies [9] to allow robots to recover from mechanical damage.”
BackgroundMethodSelf-citationWang, Y., Wang, Y., Chen, X., Wang, Y., & Chang, Z. (2024) · Electronics · doi:10.3390/electronics13173545 · Cited by 5
Gait planning and adjustment for a hexapod walking robot with one leg failure
Zhang, Y., Zhang, H., Ma, W., Li, P., & Fang, D. (2024) · Australian Journal of Mechanical Engineering · doi:10.1080/14484846.2024.2395120 · Cited by 5
“In recent years, machine learning or reinforcement learning methods have been used to generate adaptive walking gaits for hexapod robots with leg damage (Chatzilygeroudis, Vassiliades, and Mouret 2018; Cully et al. 2015; Koos, Cully, and Mouret 2013; Mailer, Nitschke, and Raw 2021).”
Qin, H., Xiang, Y., Han, Y., Wang, Y., Li, J., & Pan, Q.-K. (2025) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2025.113352 · Cited by 5
MadEvolve: evolutionary optimization of cosmological algorithms with large language models
Li, T., 臧, S.-H. 诗. Z., & Münchmeyer, M. (2026) · Machine Learning Science and Technology · doi:10.1088/2632-2153/ae7e3b · Cited by 7
“In all cases, algorithms are trained on a single simulation and tested on held-out realizations.”
A Preliminary Study of Multi-task MAP-Elites with Knowledge Transfer for Robotic Arm Design
Liu, P., Guo, Z., Yu, H., Linghu, H., Li, Y., Hou, Y., Ge, H., & Zhang, Q. (2022) · IEEE Congress on Evolutionary Computation (CEC) · doi:10.1109/cec55065.2022.9870374 · Cited by 5
“MAP-Elites have been successfully used in a variety of scenarios, such as adapting robots to damage in minutes [32], designing robot arm morphologies [31], driving robots through mazes [33], generating images that fool deep neural networks [34].”
MethodA Quality-Diversity Approach to Evolving a Repertoire of Diverse Behaviour-Trees in Robot Swarms
Montague, K., Hart, E., Nitschke, G., & Paechter, B. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-30229-9_10 · Cited by 5
Human Variability and the Explore–Exploit Trade‐Off in Recommendation
Yang, S. C., Rank, C., Whritner, J. A., Nasraoui, O., & Shafto, P. (2023) · Cognitive Science · doi:10.1111/cogs.13279 · Cited by 6
Evolving Flying Machines in Minecraft Using Quality Diversity
Medina, A. L., Richey, M., Mueller, M. A., & Schrum, J. (2023) · Genetic and Evolutionary Computation Conference (GECCO) · doi:10.1145/3583131.3590352 · Cited by 5
“and The Legend of Zelda [16], multi-legged robot gaits [5], and ar X iv :2 30 2.”
BackgroundA systematic study of biological SE systems from complexity and design perspectives
Brooks, S., Roy, R., Dirks, J., & Taylor, D. M. (2023) · Journal of Engineering Design · doi:10.1080/09544828.2023.2266864 · Cited by 5
“…biology, such a vascular self-healing materials (Hone, Kelehan, and Taylor 2021; Norris et al.2011; Pang and Bond 2005), robotic self-adaptations (Cully et al. 2015; Homchanthanakul and Manoonpong 2022) and self-sealing materials (Bauer, Nellesen, and Speck 2010; Becker et al.2022) However,…”
BackgroundQin, H., Xiang, Y., Han, Y., & Yan, X. (2024) · International Conference on Data-driven Optimization of Complex Systems (DOCS) · doi:10.1109/docs63458.2024.10704469 · Cited by 8
“Indeed, the QD algorithm has garnered success across a myriad of domains, such as robotics [11], gaming [12].”
BackgroundExploring the Performance-Reproducibility Trade-Off in Quality-Diversity
Flageat, M., Janmohamed, H., Lim, B., & Cully, A. (2025) · IEEE Transactions on Evolutionary Computation · doi:10.1109/tevc.2025.3548438 · Cited by 5
“A QD algorithm produces an archive of solutions that is then used to solve a downstream task (top).”
Self-citationSrisuchinnawong, A., & Manoonpong, P. (2025) · The International Journal of Robotics Research · doi:10.1177/02783649251336385 · Cited by 6
“…2023; Shafiee et al. 2024; Yu et al. 2023), providing task context (Ding and Zhu 2022; Smith et al. 2022a), designing leg encoding rules (Thor et al. 2020; Thor and Manoonpong 2022), and/or collecting sufficient fundamental skills/options (Zhang et al. 2023; Cully et al. 2015; Ding and Zhu 2022).”
Cui, J., Ren, Y., Xu, B., Yang, D., & Zeng, S. (2019) · Eksploatacja i Niezawodnosc - Maintenance and Reliability · doi:10.17531/ein.2020.1.6 · Cited by 5
“[5] proposed an intelligent self-adaption control method for a six-legged robot.”
MethodBootstrapped Neuro-Simulation for complex robots
Woodford, G. W., & du Plessis, M. C. (2020) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2020.103708 · Cited by 5
Evolutionary Computation and the Reinforcement Learning Problem
Kelly, S., & Schossau, J. (2023) · Genetic and evolutionary computation · doi:10.1007/978-981-99-3814-8_4 · Cited by 6
Bayesian Optimization of a Quadruped Robot During 3-Dimensional Locomotion
Zhu, J., Li, S., Wang, Z., & Rosendo, A. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-24741-6_26 · Cited by 5
Bayesian Optimization for Developmental Robotics with Meta-Learning by Parameters Bounds Reduction
Petit, M., Dellandréa, E., & Chen, L. (2020) · IEEE International Conference on Development and Learning (ICDL) · doi:10.1109/icdl-epirob48136.2020.9278071 · Cited by 5
“In particular, Cully et al implemented an extended version allowing a robot to quickly adjust its parametric gait after been damaged [11] by taking advantages of previous simulated experiences with damaged legs.”
BackgroundMorphological Development at the Evolutionary Timescale: Robotic Developmental Evolution
Benureau, F., & Tani, J. (2022) · Artificial Life · doi:10.1162/artl_a_00357 · Cited by 6
“…selective pressures (Doncieux & Artificial Life Volume 28, Number 1 17 Mouret, 2014; Risi et al., 2009), while others, like illumination algorithms (Cully et al., 2015; Mouret & Clune, 2015), modify the structure of the search to expose an ensemble of solutions across specific dimensions of…”
BackgroundEvolutionary Robotics Controllers with Proprioception Facilitated by Neural-Network-Based Simulators
Phillips, A. P., & du Plessis, M. C. (2020) · Artificial Life · doi:10.1162/artl_a_00331 · Cited by 7
WILDA: Wide Learning of Diverse Architectures for Classification of Large Datasets
Cardoso, R. P., Hart, E., Kurka, D. B., & Pitt, J. V. (2021) · Lecture notes in computer science · doi:10.1007/978-3-030-72699-7_41 · Cited by 6
Multi-objective multi-armed bandit with lexicographically ordered and satisficing objectives
Hüyük, A., & Tekin, C. (2021) · Machine Learning · doi:10.1007/s10994-021-05956-1 · Cited by 30
“For instance, Cully et al., (2015) uses a multi-armed bandit (MAB) framework to develop control policies for robots that seek to prevent damage to the parts of the robot.”
BackgroundDeformation and mechanics of a pulvinus-inspired material
Tadrist, L., Mammadi, Y., Diperi, J., & Linares, J. (2022) · Bioinspiration & Biomimetics · doi:10.1088/1748-3190/ac884f · Cited by 5
Damage resilience of manufactured and biological actuators
Perrier, R., Tadrist, L., & Linares, J. (2022) · Bioinspiration & Biomimetics · doi:10.1088/1748-3190/ac9fb6 · Cited by 5
Increasing Features in MAP-Elites Using an Age-Layered Population Structure
Pozzuoli, A., & Ross, B. J. (2023) · IEEE Congress on Evolutionary Computation (CEC) · doi:10.1109/cec53210.2023.10254093 · Cited by 5
“[15], design molecules [16], and navigate simulated robots through mazes [17].”
BackgroundXie, Y., Wang, X., Fumiya Iida, A., & Howard, D. (2024) · IEEE International Conference on Soft Robotics (RoboSoft) · doi:10.1109/robosoft60065.2024.10521959 · Cited by 5
“3: SOFA simulation setup algorithm combining the adaptation mechanisms of CMA-ES [23] with the archiving mechanism of MAP-Elites [24].”
MethodRilkean Memories and the Self of a Robot
Chella, A. (2019) · Philosophies · doi:10.3390/philosophies4020020 · Cited by 5
“[16] and the hexapod robot by Mouret and collaborators [17].”
BackgroundRaihan, A. S., & Ahmed, I. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2302.13360 · Cited by 5
Braga, J., Regateiro, F. S., Stiubiener, I., & Braga, J. C. (2023) · OSF Preprints (OSF Preprints) · doi:10.31219/osf.io/wexjr · Cited by 6
AI-driven automated discovery tools reveal diverse behavioral competencies of biological networksPreprint
Etcheverry, M., Moulin-Frier, C., Oudeyer, P., & Levin, M. G. (2025) · eLife · doi:10.7554/elife.92683 · Cited by 15
Grillotti, L., Faldor, M., León, B. G., & Cully, A. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2403.09930 · Cited by 9
“Such descriptors can take the form of successor features (Zahavy et al., 2022; 2023) or of expected features obtained though entire episodes (Cully et al., 2015; Batra et al., 2023).”
BackgroundInfluentialSelf-citationTao, Y., Liu, H., Song, Y., Deng, C., Wang, B., & Zheng, P. (2024) · Journal of Computing and Information Science in Engineering · doi:10.1115/1.4067523 · Cited by 4
A Group-Based Many-Task Collaborative Optimization Framework for Evolutionary Robots Design
Hou, Y., Zhao-ping, Y., Guo, Z., Pei, W., Wu, Y., Ge, H., Xue, B., & Zhang, M. (2025) · IEEE Transactions on Systems Man and Cybernetics Systems · doi:10.1109/tsmc.2025.3541002 · Cited by 4
“…logic of each morphology, either encoding the robot morphological parameters together with the control parameters using gene regulatory networks [32], [33] and direct coding [29], [34], [35], or using more complex methods, such as co-evolution [36] and reinforcement learning [37], [38], [39], [40].”
A survey on Quality-Diversity optimization: Approaches, applications, and challenges
Qin, H., Xiang, Y., Zhang, H., Han, Y., Wang, Y., Tao, X.-R., & Liu, Y. (2025) · Swarm and Evolutionary Computation · doi:10.1016/j.swevo.2025.102240 · Cited by 4
Zou, R., Qin, H., Xiang, Y., & Wu, C. (2026) · Engineering Optimization · doi:10.1080/0305215x.2026.2637656 · Cited by 4
An origami-based anthropomorphic rotary robotic Arm
Wang, Z., Shi, X., Zhang, G., & Wang, W. (2026) · International Journal of Mechanical Sciences · doi:10.1016/j.ijmecsci.2026.111558 · Cited by 6
Creating Diverse Ensembles for Classification with Genetic Programming and Neuro-MAP-Elites
Nickerson, K., Kolokolova, A., & Hu, T. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-02056-8_14 · Cited by 6
Gavra, V., & Kampen, E.-J. V. (2024) · AIAA SciTech Forum · doi:10.2514/6.2024-0954 · Cited by 4
Flageat, M., Lim, B., & Cully, A. (2024) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v38i11.29090 · Cited by 7
Self-citationBody and Brain Quality-Diversity in Robot Swarms
Mkhatshwa, S., & Nitschke, G. (2024) · ACM Transactions on Evolutionary Learning and Optimization · doi:10.1145/3664656 · Cited by 6
García-Córdova, F., González, A. G., & Hidalgo-Castelo, F. (2024) · Biomimetics · doi:10.3390/biomimetics9060329 · Cited by 7
“These algorithms are based on the emulation of the principles and mechanisms of adaptation, learning, and robustness present in nature, with the aim of creating more resilient and autonomous robotic systems [15].”
MethodA Fault‐Tolerant Approach for Modular Robots through Self‐Reconfiguration
Qi, J., Lai, M., Yang, Z., Zhao, N., Han, K., Sui, X., Zhao, J., & Zhu, Y. (2024) · Advanced Intelligent Systems · doi:10.1002/aisy.202300774 · Cited by 4
Qu, S., Guo, R., Cao, Z., Liu, J., Su, B., & Liu, M. (2024) · Applied Sciences · doi:10.3390/app14188383 · Cited by 4
“Additionally, a new perspective [20] on policy search has been established by solely conducting architecture search while ignoring weights.”
Jaramillo, C. M., Squires, P., & Togelius, J. (2025) · Lecture notes in computer science · doi:10.1007/978-3-031-90062-4_15 · Cited by 4
“For example, if searching for robot gaits, Map-Elites can generate a map of gaits where the descriptors correspond to which legs are used, and a good robot gait that fits a specific leg configuration can easily be found [7].”
Use of Artificial Intelligence for Monitoring Algal Blooms in Aquatic Ecosystem
Dubey, D., & Dutta, V. (2025) · Microorganisms for sustainability · doi:10.1007/978-981-96-3448-4_6 · Cited by 6
Quality‐Diversity Methods for the Modern Data Scientist
Stock, M., Hauwermeiren, D. V., Baets, B. D., Taelman, S., Marzougui, D., & Haeverbeke, M. V. (2025) · Wiley Interdisciplinary Reviews Computational Statistics · doi:10.1002/wics.70047 · Cited by 4
“This pre-computed behavioral archive allowed a hexapod robot to rapidly adapt to unforeseen physical damage by identifying alternative, effective locomotion strategies from its archive, showcasing a form of resilience akin to biological adaptation (Cully et al. 2015).”
Man vs. Machine: The Battle for the Soul of Data Science
Reid, D. (2016) · Palgrave Macmillan UK eBooks · doi:10.1057/978-1-349-94885-7_2 · Cited by 6
Design choices for adapting bio-hybrid systems with evolutionary computation
Mariano, P., Salem, Z., Mills, R., Zahadat, P., Correia, L. M. P., & Schmickl, T. (2017) · Genetic and Evolutionary Computation Conference Companion (GECCO Companion) · doi:10.1145/3067695.3076044 · Cited by 4
“…[6, 8] we aim to get rid of the need for such an a priori model and use computational meth-ods to develop the needed proximate mechanisms of robots in a model-free way during runtime, following ideas that were recently suggested for robotics in a different context (without animal interaction) [2].”
BackgroundInfluentialOnline Learning with Stochastic Recurrent Neural Networks using Intrinsic Motivation Signals
Tanneberg, D., Peters, J., & Rueckert, E. (2022) · Technischen Universität Darmstadt · doi:10.26083/tuprints-00020580 · Cited by 4
Comparing Approaches for Evolving High-Level Robot Control Based on Behaviour Repertoires
Gomes, J., & Christensen, A. L. (2018) · IEEE Congress on Evolutionary Computation (CEC) · doi:10.1109/cec.2018.8477699 · Cited by 4
“Noveltydriven evolutionary algorithms have been used to evolve repertoires of robot behaviours in a number of different domains, including: virtual walking creatures [12]; morphological designs for walking soft robots [11]; locomotion behaviours for legged robots [13]–[16] and four-wheeled steering robots [17]; roboti…”
BackgroundGait Generation for Damaged Hexapods using a Genetic Algorithm
Kon, J., & Sahin, F. (2020) · International Conference on System of Systems Engineering · doi:10.1109/sose50414.2020.9130561 · Cited by 4
“While physics simulation environments are by no means new, few examples of simulations applied to hexapod gait generation were found [6] [3].”
BackgroundComplexity of self-engineering systems across the life cycle – Biological and engineering systems
Brooks, S., & Roy, R. (2021) · Procedia CIRP · doi:10.1016/j.procir.2021.01.016 · Cited by 4
Omni-directional gait of a passive-spine hexapod
Tang, Y., Zhang, G., Ge, D., & Ma, S. (2019) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2019.06.001 · Cited by 4
Complex Morphology Neural Network Simulation in Evolutionary Robotics
Woodford, G. W., & du Plessis, M. C. (2019) · Robotica · doi:10.1017/s0263574719001140 · Cited by 5
“14 A disadvantage of all these approaches is that they require physics-based simulators.”
BackgroundLim, G. H., Pedrosa, E., Amaral, F., Pereira, A., Lau, N., Azevedo, J. L., Cunha, B., & Badini, S. (2019) · Robotics and Autonomous Systems · doi:10.1016/j.robot.2019.06.006 · Cited by 4
Blocky Net: A New NeuroEvolution Method
Watts, T., Xue, B., & Zhang, M. (2019) · IEEE Congress on Evolutionary Computation (CEC) · doi:10.1109/cec.2019.8790302 · Cited by 4
“Similar unsupervised EC methods have been recently used to evolve highly adaptive creatures in digital environments, which are then transferred into real-world domains [22].”
BackgroundToward Faster Reinforcement Learning for Robotics: Using Gaussian Processes
Younes, A., & Panov, A. I. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-33274-7_11 · Cited by 4
Evolutionary Algorithms in Web Security: Exploring Untapped Potential
Attwood, S., Li, W., & Kharel, R. (2020) · International Symposium on Communication Systems, Networks and Digital Signal Processing (CSNDSP) · doi:10.1109/csndsp49049.2020.9249521 · Cited by 4
“Other quality diversity algorithms, such as the Multi Dimensional Archive of Phenotypic Elites algorithm (MAP-Elites) [8], operate in different ways but similarly hope to achieve a diversity of high quality solutions in a single run.”
MethodEmergent adaptive gait generation through Hebbian sensor-motor maps by morphological probing
Dujany, M., Hauser, S., Mutlu, M., van der Sar, M., Arreguit, J., Kano, T., Ishiguro, A., & Ijspeert, A. J. (2020) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros45743.2020.9341211 · Cited by 7
“In [2], researchers built a model of a hexapod in simulation whose leg configurations could easily be modified.”
BackgroundBayesian Optimization Approach to Input Shaper Design for Flexible Beam Vibration Suppression
Pásztori, Z., Ruggiero, F., Lippiello, V., & Castro, M. D. (2020) · IFAC-PapersOnLine · doi:10.1016/j.ifacol.2020.12.2159 · Cited by 4
“Dadios et al. (2006) has synthesised a controller based on a simulated system model with genetic algorithms.”
MethodInfluentialYu, W., Liu, C. K., & Turk, G. (2021) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra48506.2021.9562051 · Cited by 4
“2] Joint Stiffness [70, 130]% Torque limit [100, 300]Nm [60,150]% [15,25]Nm Position gain - [60, 150]%”
BackgroundAdaptive Walking Pattern Generation with Damaged Legs for the six-legged Robot LAURON V
Plasberg, C., Kögel, D., Roennau, A., & Dillmann, R. (2021) · International Conference on Advanced Robotics (ICAR) · doi:10.1109/icar53236.2021.9659374 · Cited by 4
Robots and Machine Consciousness
Chella, A. (2022) · The MIT Press eBooks · doi:10.7551/mitpress/13780.003.0029 · Cited by 5
Advanced Turning Maneuver of a Many-Legged Robot Using Pitchfork Bifurcation
Aoi, S., Tomatsu, R., Yabuuchi, Y., Morozumi, D., Okamoto, K., Fujiki, S., Senda, K., & Tsuchiya, K. (2022) · IEEE Transactions on Robotics · doi:10.1109/tro.2022.3158194 · Cited by 4
“Furthermore, damage to even one leg greatly degrades their locomotive performance [12].”
BackgroundTactile Resilience of Sensory Whisker by Adaptive Morphology
Nguyen, N. H., Häuser, H., Maiolino, P., & Ho, V. A. (2022) · IEEE Access · doi:10.1109/access.2022.3208883 · Cited by 4
“Short expression for above method is as follows : 2) The second approach is inspired by searching adap-441 tive behaviors of natural creatures throughout a ‘‘trial-442 and-error’’ process [30], [31].”
MethodAboufazeli, M., Filsoofi, A. S., Gurney, J., Meek, S. G., & Mathews, V. J. (2023) · Frontiers in Robotics and AI · doi:10.3389/frobt.2023.1127898 · Cited by 4
“, 2019) and intelligent trial and error algorithm for robots to adapt to damage presented in (Cully et al., 2015).”
BackgroundActive Sensing for Data Quality Improvement in Model Learning
Napolitano, O., Cognetti, M., Pallottino, L., Kanoulas, D., Salaris, P., & Modugno, V. (2024) · IEEE Control Systems Letters · doi:10.1109/lcsys.2024.3410152 · Cited by 6
“Therefore, to best predict the model, during the learning process, it is of paramount importance to identify the few data points that contain the largest amount of information [3].”
BackgroundEfficient online adaptation with stochastic recurrent neural networks
Tanneberg, D., Peters, J., & Rueckert, E. (2017) · IEEE-RAS International Conference on Humanoid Robots (Humanoids) · doi:10.1109/humanoids.2017.8246875 · Cited by 5
“In future therefore, we want to combine this approach with the factorized population coding from [24] to scale to higher dimensional problems and apply the framework to recover from failure tasks with broken joints [31], [32].”
MethodDevelopmental Bayesian Optimization of Black-Box with Visual Similarity-Based Transfer Learning
Petit, M., Depierre, A., Wang, X., Dellandréa, E., & Chen, L. (2018) · IEEE International Conference on Development and Learning (ICDL) · doi:10.1109/devlrn.2018.8761037 · Cited by 5
“developed a walking robot that can quickly adapt its gait after been damaged using an intelligent trial and error algorithm [10].”
Zhou, Q., Liu, M., Wang, W., Xu, L., Ao, T., & Zhang, H. (2018) · Chinese Control Conference (CCC) · doi:10.23919/chicc.2018.8483976 · Cited by 4
Brych, S., & Cully, A. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2009.08438 · Cited by 4
“This is illustrated by the diversity of approaches that have been explored in the literature, for instance: Policy Gradient Methods [15, 16, 17, 10, 18], Evolutionary Algorithms [3, 1, 14], Particle Swarm Optimization [19] or Bayesian Optimization [20].”
BackgroundMethodInfluentialSelf-citationEvolutionary Gait Transfer of Multi-Legged Robots in Complex TerrainsPreprint
Jiang, M., Chi, G., Pan, G., Guo, S., & Tan, K. C. (2020) · arXiv (Cornell University) · doi:10.48550/arxiv.2012.13320 · Cited by 4
“Recently, studies have shown that quality diversity [45] can generate a variety of motion behaviors for complex robots in each evolutionary operation [46]–[48].”
BackgroundOCCAM: Online Continuous Controller Adaptation with Meta-Learned ModelsPreprint
Sanghvi, H., Folk, S., & Taylor, C. J. (2024) · arXiv (Cornell University) · doi:10.48550/arxiv.2406.17620 · Cited by 8
“[14, 15] add prior datapoints to a GP to find control policies for damaged robots online; [16, 17] instead use offline data to learn fixed kernel functions that make BO more efficient.”
BackgroundMultimodal methods for managing deontic rights in interdisciplinary trials
Nicolaisen, E. M., & Rasmussen, G. (2023) · Multimodality & Society · doi:10.1177/26349795231188727 · Cited by 3
Enhancing Robot Learning Through Cognitive Reasoning Trajectory Optimization Under Unknown Dynamics
Dong, Q., Wu, T., Zeng, P., Zang, C., Wan, G., & Cui, S. (2025) · IEEE Robotics and Automation Letters · doi:10.1109/lra.2025.3558648 · Cited by 4
“Most existing meth-odsemploytask-specificpriors,includingthosefordynamicsys-tems[11],[12],policyparametersandstructures[13],[14],[15], expectedreturns[16],[17],[18],andthelearningofdynamic systemmodels[19],[20],toendowrobotswithinnatefounda-tionalknowledge.”
Insect‐Inspired Resilient Machines
Chuthong, T., Büscher, T. H., Gorb, S. N., & Manoonpong, P. (2025) · Advanced Intelligent Systems · doi:10.1002/aisy.202500270 · Cited by 3
Jin, S., Ke, K., Gao, B., Fu, L., & Huang, X. (2025) · Chinese Journal of Mechanical Engineering · doi:10.1186/s10033-025-01356-x · Cited by 4
“Meanwhile, studies have drawn inspiration from the adaptation mechanisms of animals to multimodal information and applied self-diagnosis mechanisms to robots to improve adaptive ability [27].”
The Future of Research in Cognitive Robotics: Foundation Models or Developmental Cognitive Models?
Vernon, D. (2025) · Advanced Robotics Research · doi:10.1002/adrr.202500066 · Cited by 3
Learning to Tune a Class of Controllers with Deep Reinforcement Learning
Shipman, W. J. (2021) · Minerals · doi:10.3390/min11090989 · Cited by 3
“Number of hidden layers in the policy function [1; 4] IntUniform Number of hidden layers in the value function [1; 4] IntUniform Number of neurons in a hidden layer [8; 256] IntUniform Hidden layer activation function {tan h; relu; leakyrelu} -”
BackgroundSoft Robots as a Platform to Understand Embodied Intelligence
Nanayakkara, T., Mulvey, B. W., Perera, S., Ge, Y., Yu, Z., & Sunilkumar, P. (2024) · Springer eBooks · doi:10.1007/978-3-031-68620-7_3 · Cited by 5
Medvet, E., Lippolis, S., & Nadizar, G. (2025) · Genetic Programming and Evolvable Machines · doi:10.1007/s10710-025-09520-x · Cited by 4
“This diversity is especially valuable in dynamic or uncertain environments, where having a range of robust solutions supports adaptation to changing conditions [21] or resilience against hardware failures [8].”
Enhancing Adaptability in Embodied Agents: A Multi-Quality-Diversity Approach
Nadizar, G., Medvet, E., & Wilson, D. G. (2025) · IEEE Transactions on Evolutionary Computation · doi:10.1109/tevc.2025.3596746 · Cited by 7
“A promising direction would be to predict in advance which robots are better suited for transfer to a given task, as multiple reassembly is unfeasible in real-world scenarios [8].”
Full Domain Analysis in Fluid Dynamics
Hagg, A., Gaier, A., Wilde, D., Asteroth, A., Foysi, H., & Reith, D. (2025) · Machine Learning and Knowledge Extraction · doi:10.3390/make7030086 · Cited by 3
Gesture reconfiguration from joint failure using genetic algorithm
Jutharee, W., & Maneewarn, T. (2016) · International Conference on Control, Automation and Systems (ICCAS) · doi:10.1109/iccas.2016.7832455 · Cited by 3
Autonomous Unmanned Aerial Vehicle (UAV) landing in windy conditions with MAP-Elites
Adibi, S. A., Forer, S., Fries, J., & Yliniemi, L. (2017) · The Knowledge Engineering Review · doi:10.1017/s0269888917000121 · Cited by 3
“MAP-Elites has been successfully used in the past for: re-training robots to recover performance after damage to limbs (Cully et al., 2015), manipulating objects (Ecarlat et al.”
MethodDistributed Pressure Sensing for Enabling Self-Aware Autonomous Aerial Vehicles
Cellucci, D., Cramer, N., & Swei, S. S.-M. (2018) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2018.8593664 · Cited by 3
“gov The ability of a fully-autonomous aerial vehicle to have a robust estimation of itself and its environment is critical for the vehicle to be able to adapt its behavior, and current research has made progress on this capability with terrestrial robots [7].”
BackgroundBioGD: Bio-inspired robust gradient descent
Kulikovskikh, I., Prokhorov, S., Lipić, T., Legović, T., & Šmuc, T. (2019) · PLoS ONE · doi:10.1371/journal.pone.0219004 · Cited by 3
Task Feasibility Maximization Using Model-Free Policy Search and Model-Based Whole-Body Control
Lober, R., Sigaud, O., & Padois, V. (2020) · Frontiers in Robotics and AI · doi:10.3389/frobt.2020.00061 · Cited by 3
“BO solvers usually require fewer trials to obtain an optimal solution and have become a popular choice in robotics because of this efficacy (Calandra et al., 2014; Cully et al., 2015; Antonova et al., 2016; Englert and Toussaint, 2016).”
Fitriyani, S., Liu, C., Hung, Y.-H., Zhang, Y., & Liu, J. H. (2020) · eXPRESS Polymer Letters · doi:10.3144/expresspolymlett.2020.46 · Cited by 3
“For thousands of years, from self-assembly to morphological control of different dimensions, scales, and functions to sophisticated mechanical locomotion and autonomous sensing strains, humans have continuously learned from biomimicry research, thus creating modern civilization [2–6].”
Lifetime policy reuse and the importance of task capacity
Bossens, D. M., & Sobey, A. J. (2023) · AI Communications · doi:10.3233/aic-230040 · Cited by 5
“…and which are based on Markov Decision Processes, or have used a large amount of policies developed in an off-line stage of learning [13, 35, 76], this paper demonstrates the viability of adaptive policy selection with continual refinement based on the lifetime average performance to solve…”
Adaptive Navigation Strategies in Evolutionary Robotics Controllers with Location Perception
Phillips, A. P., & du Plessis, M. C. (2021) · IEEE Congress on Evolutionary Computation (CEC) · doi:10.1109/cec45853.2021.9504759 · Cited by 5
OPEn: An Open-ended Physics Environment for Learning Without a Task
Gan, C., Bhandwaldar, A., Torralba, A. B., Tenenbaum, J. B., & Isola, P. (2021) · IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros51168.2021.9636830 · Cited by 7
“, [43], [14], [36], [37], [21], [28], [29], [38], [9]).”
The Walking Robot Control System that is Adaptive to Changes in the Kinematics
Сергеевна, С. А., & Vorotnikov, S. A. (2021) · Mekhatronika Avtomatizatsiya Upravlenie · doi:10.17587/mau.22.601-609 · Cited by 3
HEXAPOD MOVEMENT ALGORITHMS TO AVOID INTERFERENCE. ANGULAR MOVEMENT
Platov, I. M., Pavlovskyi, O., & Pavlovska, Y. (2021) · Bulletin of Kyiv Polytechnic Institute Series Instrument Making · doi:10.20535/1970.62(2).2021.249214 · Cited by 3
“The work [8], carried out by developers from France led by Jean-Baptiste Moret, deserves special attention.”
BackgroundFast Adaptation Dynamics Model for Robot’s Damage Recovery
Chen, C., Wang, D., Yu, J., Xiang, P., Lu, H., Wang, Y., & Xiong, R. (2022) · IEEE International Conference on Real-time Computing and Robotics (RCAR) · doi:10.1109/rcar54675.2022.9872230 · Cited by 3
“[7] proposed an intelligence trial-and-error method, which enables the robots to store previous experience knowledge in the form of a behavior-representation space map, and uses the Bayesian optimization method combined with the map to guide the damaged robot to try different types of behaviors.”
MethodKim, T., Yadav, P., Suk, H., & Kim, S. (2022) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2022.105367 · Cited by 3
An Improved Flocking Model for UAVs in Constrained Environments
Zhang, X., Yuan, Y., & Zhang, F. (2022) · International Conference on Big Data and Information Analytics (BigDIA) · doi:10.1109/bigdia56350.2022.9874099 · Cited by 3
“[9] proposed a sophisticated trial-and-error algorithm allows the robot to react to damage in less than two minutes.”
BackgroundA scalable species-based genetic algorithm for reinforcement learning problems
Seth, A., Nikou, A., & Daoutis, M. (2022) · The Knowledge Engineering Review · doi:10.1017/s0269888922000042 · Cited by 3
“In a subsequent paper (Conti et al ., 2018), the same researchers incorporated quality diversity (QD) (Cully et al ., 2015; Pugh et al ., 2016) and novelty search (NS) (Lehman & Stanley, 2008) to ES, and proposed three different algorithms, which improved performance and also directed the…”
Background
Metadata
- Authors
- Cully, A., Clune, J., Tarapore, D., & Mouret, J.-B.
- Affiliations
- Centre National de la Recherche Scientifique (FR)
- University of Wyoming (US)
- Corresponding author
- Mouret, Jean-Baptiste
- Venue
- Nature
- Published
- 2015-05-27
- Volume
- 521
- Issue
- 7553
- Pages
- 503-507
- ISSN
- 0028-0836 (ISSN-L)
- 1476-4687
- OpenAlex
- W1738827650
- Cited by
- 979
- References
- 40
- Citations by year
- 2026: 51 · 2025: 76 · 2024: 88 · 2023: 90 · 2022: 100 · 2021: 115 · 2020: 129 · 2019: 112 · 2018: 78 · 2017: 78 · 2016: 48 · 2015: 13
- FWCI
- 70.47
- Type
- journal article
- Language
- en
- Open access
- green
- arxiv.org
- Licence
- www.springer.com/tdm (tdm)
- www.springer.com/tdm (vor)
- Locations
- Nature (publishedVersion)
- PubMed (publishedVersion)
- ePrints Soton (University of Southampton) (acceptedVersion)
- arXiv (Cornell University) (submittedVersion)
- HAL (submittedVersion, other-oa)
- RePEc: Research Papers in Economics (submittedVersion)
- arXiv (Cornell University)
- Funders
- Agence Nationale de la Recherche (ANR-12-JS03-0009)
- Topics
- Physical Sciences › Computer Science › Artificial Intelligence › Reinforcement Learning in Robotics
- Physical Sciences › Engineering › Mechanical Engineering › Modular Robots and Swarm Intelligence
- Physical Sciences › Computer Science › Artificial Intelligence › Evolutionary Algorithms and Applications
- Keywords
- compensatory behavior, trial-and-error learning, autonomous mobile robots, legged robots, robot manipulators, prior knowledge, robot robustness, search and rescue robots, disaster response
- MeSH terms
- Adaptation, Physiological, Algorithms, Animals, Artificial Intelligence, Behavior, Animal, Dogs, Extremities, Motor Skills, Robotics, Time Factors, Biomimetics
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.