{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,17]],"date-time":"2026-04-17T20:38:17Z","timestamp":1776458297531,"version":"3.51.2"},"reference-count":90,"publisher":"Annual Reviews","issue":"1","license":[{"start":{"date-parts":[[2024,7,10]],"date-time":"2024-07-10T00:00:00Z","timestamp":1720569600000},"content-version":"unspecified","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2024,7,10]]},"abstract":"<jats:p>In recent years, the increasing complexity and safety-critical nature of robotic tasks have highlighted the importance of accurate and reliable robot models. This trend has led to a growing belief that, given enough data, traditional physics-based robot models can be replaced by appropriately trained deep networks or their variants. Simultaneously, there has been a renewed interest in physics-based simulation, fueled by the widespread use of simulators to train reinforcement learning algorithms in the sim-to-real paradigm. The primary objective of this review is to present a unified perspective on the process of determining robot models from data, commonly known as system identification or model learning in different subfields. The review aims to illuminate the key challenges encountered and highlight recent advancements in system identification for robotics. Specifically, we focus on recent breakthroughs that leverage the geometry of the identification problem and incorporate physics-based knowledge beyond mere first-principles model parameterizations. Through these efforts, we strive to provide a contemporary outlook on this problem, bridging classical findings with the latest progress in the field.<\/jats:p>","DOI":"10.1146\/annurev-control-061523-102310","type":"journal-article","created":{"date-parts":[[2023,10,20]],"date-time":"2023-10-20T18:48:01Z","timestamp":1697827681000},"page":"311-334","source":"Crossref","is-referenced-by-count":22,"title":["Robot Model Identification and Learning: A Modern Perspective"],"prefix":"10.1146","volume":"7","author":[{"given":"Taeyoon","family":"Lee","sequence":"first","affiliation":[{"name":"1Naver Labs, Seongnam, South Korea; email: ty-lee@naverlabs.com, jaewoon.kwon@naverlabs.com"}]},{"given":"Jaewoon","family":"Kwon","sequence":"additional","affiliation":[{"name":"1Naver Labs, Seongnam, South Korea; email: ty-lee@naverlabs.com, jaewoon.kwon@naverlabs.com"}]},{"given":"Patrick M.","family":"Wensing","sequence":"additional","affiliation":[{"name":"2Department of Aerospace and Mechanical Engineering, University of Notre Dame, Notre Dame, Indiana, USA; email: pwensing@nd.edu"}]},{"given":"Frank C.","family":"Park","sequence":"additional","affiliation":[{"name":"3Department of Mechanical Engineering, Seoul National University, Seoul, South Korea; email: fcp@snu.ac.kr"}]}],"member":"22","reference":[{"issue":"2","key":"B1","doi-asserted-by":"crossref","first-page":"225","DOI":"10.1016\/j.automatica.2003.10.001","article-title":"Parameter estimation in stochastic grey-box models","volume":"40","year":"2004","journal-title":"Automatica"},{"issue":"6","key":"B2","doi-asserted-by":"crossref","first-page":"28","DOI":"10.1109\/MCS.2019.2938121","article-title":"Nonlinear system identification: a user-oriented road map","volume":"39","year":"2019","journal-title":"IEEE Control Syst. Mag."},{"key":"B3","first-page":"2226","article-title":"DayDreamer: world models for physical robot learning","volume-title":"Proceedings of the 6th Conference on Robot Learning","year":"2023"},{"issue":"3","key":"B4","article-title":"Output error methods for robot identification","volume":"142","year":"2020","journal-title":"J. Dyn. Syst. Meas. Control"},{"issue":"5","key":"B5","doi-asserted-by":"crossref","first-page":"730","DOI":"10.1109\/70.631234","article-title":"Optimal robot excitation and identification","volume":"13","year":"1997","journal-title":"IEEE Trans. Robot. Autom."},{"issue":"2","key":"B6","first-page":"428","article-title":"A new closed-loop output error method for parameter identification of robot dynamics","volume":"21","year":"2012","journal-title":"IEEE Trans. Control Syst. Technol."},{"key":"B7","doi-asserted-by":"crossref","first-page":"102398","DOI":"10.1016\/j.mechatronics.2020.102398","article-title":"A hybrid dynamic model for the ambidex tendon-driven manipulator","volume":"69","year":"2020","journal-title":"Mechatronics"},{"issue":"26","key":"B8","doi-asserted-by":"crossref","first-page":"eaau5872","DOI":"10.1126\/scirobotics.aau5872","article-title":"Learning agile and dynamic motor skills for legged robots","volume":"4","year":"2019","journal-title":"Sci. Robot."},{"issue":"4","key":"B9","doi-asserted-by":"crossref","first-page":"1100","DOI":"10.1109\/TRO.2020.3047515","article-title":"Kinodynamic model identification: a unified geometric approach","volume":"37","year":"2021","journal-title":"IEEE Trans. Robot."},{"key":"B10","volume-title":"Optimal Design of Experiments","year":"2006"},{"issue":"2","key":"B11","doi-asserted-by":"crossref","first-page":"303","DOI":"10.1016\/j.automatica.2007.05.016","article-title":"Optimal experimental design and some related control problems","volume":"44","year":"2008","journal-title":"Automatica"},{"issue":"4","key":"B12","doi-asserted-by":"crossref","first-page":"823","DOI":"10.1109\/TRO.2016.2583062","article-title":"Optimal exciting dance for identifying inertial parameters of an anthropomorphic structure","volume":"32","year":"2016","journal-title":"IEEE Trans. Robot."},{"key":"B13","doi-asserted-by":"crossref","first-page":"411","DOI":"10.1146\/annurev-control-042920-020211","article-title":"Safe learning in robotics: from learning-based control to safe reinforcement learning","volume":"5","year":"2022","journal-title":"Annu. Rev. Control Robot. Auton. Syst."},{"key":"B14","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1146\/annurev-control-062322-100607","article-title":"An overview of soft robotics","volume":"6","year":"2023","journal-title":"Annu. Rev. Control Robot. Auton. Syst."},{"issue":"3","key":"B15","doi-asserted-by":"crossref","first-page":"101","DOI":"10.1177\/027836498600500306","article-title":"Estimation of inertial parameters of manipulator loads and links","volume":"5","year":"1986","journal-title":"Int. J. Robot. Res."},{"key":"B16","article-title":"A geometric characterization of observability in inertial parameter identification","year":"2023"},{"key":"B17","first-page":"1367","article-title":"Base parameters of manipulator dynamic models","volume-title":"1988 IEEE International Conference on Robotics and Automation","volume":"3","year":"1988"},{"key":"B18","first-page":"1682","article-title":"A direct determination of minimum inertial parameters of robots","volume-title":"1988 IEEE International Conference on Robotics and Automation","volume":"3","year":"1988"},{"issue":"6","key":"B19","doi-asserted-by":"crossref","first-page":"517","DOI":"10.1109\/JRA.1987.1087145","article-title":"Minimum operations and minimum parameters of the dynamic models of tree structure robots","volume":"3","year":"1987","journal-title":"IEEE J. Robot. Autom."},{"key":"B20","doi-asserted-by":"crossref","first-page":"284","DOI":"10.1016\/j.mechmachtheory.2011.09.006","article-title":"3D inertia transfer concept and symbolic determination of the base inertial parameters","volume":"49","year":"2012","journal-title":"Mech. Mach. Theory"},{"issue":"4","key":"B21","doi-asserted-by":"crossref","first-page":"485","DOI":"10.1002\/rob.4620080405","article-title":"Numerical calculation of the base inertial parameters of robots","volume":"8","year":"1991","journal-title":"J. Robot. Syst."},{"key":"B22","volume-title":"The Total Least Squares Problem: Computational Aspects and Analysis","year":"1991"},{"issue":"1","key":"B23","first-page":"132","article-title":"A generic instrumental variable approach for industrial robot identification","volume":"22","year":"2013","journal-title":"IEEE Trans. Control Syst. Technol."},{"key":"B24","doi-asserted-by":"crossref","first-page":"1962","DOI":"10.1109\/IROS.2009.5354469","article-title":"Using robust regressions and residual analysis to verify the reliability of LS estimation: application in robotics","volume-title":"2009 IEEE\/RSJ International Conference on Intelligent Robots and Systems","year":"2009"},{"issue":"2","key":"B25","doi-asserted-by":"crossref","first-page":"253","DOI":"10.1109\/TMECH.2005.844703","article-title":"Robust dynamic experimental identification of robots with set membership uncertainty","volume":"10","year":"2005","journal-title":"IEEE\/ASME Trans. Mechatron."},{"issue":"4\u20135","key":"B26","first-page":"335","article-title":"Identification for control: from the early achievements to the revival of experiment design","volume":"11","year":"2005","journal-title":"Eur. J. Control"},{"issue":"9","key":"B27","doi-asserted-by":"crossref","first-page":"4303","DOI":"10.3390\/app11094303","article-title":"Inertial parameter identification in robotics: a survey","volume":"11","year":"2021","journal-title":"Appl. Sci."},{"issue":"3","key":"B28","doi-asserted-by":"crossref","first-page":"446","DOI":"10.1177\/0278364913495932","article-title":"Identifiability and identification of inertial parameters using the underactuated base-link dynamics for legged multibody systems","volume":"33","year":"2014","journal-title":"Int. J. Robot. Res."},{"key":"B29","doi-asserted-by":"crossref","first-page":"4575","DOI":"10.1109\/IEMBS.2008.4650231","article-title":"Motion capture based identification of the human body inertial parameters","volume-title":"2008 30th Annual International Conference of the IEEE Engineering in Medicine and Biology Society","year":"2008"},{"key":"B30","doi-asserted-by":"crossref","first-page":"457","DOI":"10.1109\/HUMANOIDS.2014.7041401","article-title":"Dynamic parameters identification of a humanoid robot using joint torque sensors and\/or contact forces","volume-title":"2014 IEEE-RAS International Conference on Humanoid Robots","year":"2014"},{"issue":"5","key":"B31","doi-asserted-by":"crossref","first-page":"1091","DOI":"10.1109\/TRO.2016.2593027","article-title":"Robustness to joint-torque-tracking errors in task-space inverse dynamics","volume":"32","year":"2016","journal-title":"IEEE Trans. Robot."},{"issue":"6","key":"B32","doi-asserted-by":"crossref","first-page":"931","DOI":"10.1177\/0278364913514870","article-title":"Physical feasibility of robot base inertial parameter identification: a linear matrix inequality approach","volume":"33","year":"2014","journal-title":"Int. J. Robot. Res."},{"key":"B33","doi-asserted-by":"crossref","first-page":"5446","DOI":"10.1109\/IROS.2016.7759801","article-title":"Identification of fully physical consistent inertial parameters using optimization on manifolds","volume-title":"2016 IEEE\/RSJ International Conference on Intelligent Robots and Systems (IROS)","year":"2016"},{"issue":"1","key":"B34","first-page":"60","article-title":"Linear matrix inequalities for physically consistent inertial parameter identification: a statistical perspective on the mass distribution","volume":"3","year":"2017","journal-title":"IEEE Robot. Autom. Lett."},{"issue":"3","key":"B35","doi-asserted-by":"crossref","first-page":"726","DOI":"10.1109\/TRO.2016.2558190","article-title":"Humanoid and human inertia parameter identification using hierarchical optimization","volume":"32","year":"2016","journal-title":"IEEE Trans. Robot."},{"key":"B36","doi-asserted-by":"crossref","first-page":"6282","DOI":"10.1109\/ICRA.2011.5979903","article-title":"Real-time implementation of physically consistent identification of human body segments","volume-title":"2011 IEEE International Conference on Robotics and Automation","year":"2011"},{"issue":"13\u201314","key":"B37","first-page":"1437","article-title":"Parameter and contact force estimation of planar rigid-bodies undergoing frictional contact","volume":"36","year":"2017","journal-title":"Int. J. Robot. Res."},{"key":"B38","first-page":"1036","article-title":"Physically consistent state estimation and system identification for contacts","volume-title":"2015 IEEE-RAS 15th International Conference on Humanoid Robots","year":"2015"},{"key":"B39","article-title":"gradSim: differentiable simulation for system identification and visuomotor control","year":"2021","journal-title":"In The Ninth International Conference on Learning Representations"},{"key":"B40","article-title":"DiffTaichi: differentiable programming for physical simulation","year":"2020","journal-title":"In The Eighth International Conference on Learning Representations"},{"key":"B41","article-title":"An end-to-end differentiable framework for contact-aware robot design","year":"2021","journal-title":"In Robotics: Science and Systems XVII"},{"key":"B42","doi-asserted-by":"crossref","first-page":"6","DOI":"10.3389\/fnbot.2019.00006","article-title":"A differentiable physics engine for deep learning in robotics","volume":"13","year":"2019","journal-title":"Front. Neurorobot."},{"issue":"6","key":"B43","first-page":"236","article-title":"Real2Sim: visco-elastic parameter estimation from dynamic motion","volume":"38","year":"2019","journal-title":"ACM Trans. Graph."},{"issue":"6","key":"B44","first-page":"190","article-title":"ADD: analytically differentiable dynamics for multi-body systems with frictional contact","volume":"39","year":"2020","journal-title":"ACM Trans. Graph."},{"key":"B45","article-title":"Fast and feature-complete differentiable physics engine for articulated rigid bodies with contact constraints","volume-title":"Robotics: Science and Systems XVII","year":"2021"},{"key":"B46","first-page":"20668","article-title":"Do differentiable simulators give better policy gradients?","volume-title":"Proceedings of the 39th International Conference on Machine Learning","year":"2022"},{"key":"B47","first-page":"276","article-title":"Rethinking optimization with differentiable simulation from a global perspective","volume-title":"Proceedings of the 6th Conference on Robot Learning","year":"2023"},{"key":"B48","article-title":"Contact models in robotics: a comparative analysis","year":"2023"},{"issue":"3","key":"B49","doi-asserted-by":"crossref","first-page":"259","DOI":"10.1016\/0167-6911(89)90058-3","article-title":"An indirect adaptive robot controller","volume":"12","year":"1989","journal-title":"Syst. Control Lett."},{"issue":"1","key":"B50","doi-asserted-by":"crossref","first-page":"122","DOI":"10.7210\/jrsj.14.122","article-title":"When is the set of base-parameter values physically impossible?","volume":"14","year":"1996","journal-title":"J. Robot. Soc. Jpn."},{"issue":"5","key":"B51","doi-asserted-by":"crossref","first-page":"498","DOI":"10.1177\/02783640022066996","article-title":"Verification of the positive definiteness of the inertial matrix of manipulators using base inertial parameters","volume":"19","year":"2000","journal-title":"Int. J. Robot. Res."},{"key":"B52","volume-title":"Modern Robotics","year":"2017"},{"key":"B53","volume-title":"Dynamics of Multibody Systems","year":"2008"},{"key":"B54","doi-asserted-by":"crossref","first-page":"5026","DOI":"10.1109\/IROS.2012.6386109","article-title":"MuJoCo: a physics engine for model-based control","volume-title":"2012 IEEE\/RSJ international conference on intelligent robots and systems","year":"2012"},{"issue":"3","key":"B55","doi-asserted-by":"crossref","first-page":"2455","DOI":"10.1109\/LRA.2018.2799426","article-title":"A geometric algorithm for robust multibody inertial parameter identification","volume":"3","year":"2018","journal-title":"IEEE Robot. Autom. Lett."},{"issue":"3","key":"B56","doi-asserted-by":"crossref","first-page":"735","DOI":"10.1137\/S0895479803436937","article-title":"A differential geometric approach to the geometric mean of symmetric positive-definite matrices","volume":"26","year":"2005","journal-title":"SIAM J. Matrix Anal. Appl."},{"issue":"2","key":"B57","first-page":"348","article-title":"Geometric robot dynamic identification: a convex programming approach","volume":"36","year":"2019","journal-title":"IEEE Trans. Robot."},{"key":"B58","volume-title":"Geometric methods for dynamic model-based identification and control of multibody systems","year":"2019"},{"key":"B59","doi-asserted-by":"crossref","first-page":"109773","DOI":"10.1016\/j.automatica.2021.109773","article-title":"Optimal excitation trajectories for mechanical systems identification","volume":"131","year":"2021","journal-title":"Automatica"},{"key":"B60","first-page":"364","article-title":"Bayesian estimation of inertial parameters of robots","volume-title":"1992 IEEE International Conference on Robotics and Automation","volume":"1","year":"1992"},{"issue":"3","key":"B61","doi-asserted-by":"crossref","first-page":"590","DOI":"10.1162\/neco_a_01360","article-title":"Implicit regularization and momentum algorithms in nonlinearly parameterized adaptive control and prediction","volume":"33","year":"2021","journal-title":"Neural Comput"},{"key":"B62","article-title":"A natural adaptive control law for robot manipulators","volume-title":"2018 IEEE\/RSJ International Conference on Intelligent Robots and Systems (IROS)","year":"2018"},{"issue":"4","key":"B63","doi-asserted-by":"crossref","first-page":"1307","DOI":"10.1109\/TRO.2020.2988642","article-title":"TossingBot: learning to throw arbitrary objects with residual physics","volume":"36","year":"2020","journal-title":"IEEE Trans. Robot."},{"key":"B64","first-page":"734","article-title":"Sim-to-real transfer with neural-augmented robot simulation","volume-title":"Proceedings of the 2nd Conference on Robot Learning","year":"2018"},{"key":"B65","first-page":"9474","article-title":"NeuralSim: augmenting differentiable simulators with neural networks","volume-title":"2021 IEEE International Conference on Robotics and Automation","year":"2021"},{"key":"B66","doi-asserted-by":"crossref","first-page":"3066","DOI":"10.1109\/IROS.2018.8593995","article-title":"Augmenting physical simulators with stochastic neural networks: case study of planar pushing and bouncing","volume-title":"2018 IEEE\/RSJ International Conference on Intelligent Robots and Systems (IROS)","year":"2018"},{"key":"B67","doi-asserted-by":"crossref","first-page":"319","DOI":"10.1007\/s10339-011-0404-1","article-title":"Model learning for robot control: a survey","volume":"12","year":"2011","journal-title":"Cogn. Process."},{"key":"B68","doi-asserted-by":"crossref","first-page":"357","DOI":"10.1017\/S096249290100006X","article-title":"Discrete mechanics and variational integrators","volume":"10","year":"2001","journal-title":"Acta Numer"},{"key":"B69","first-page":"352","article-title":"A linear-time variational integrator for multibody systems","volume-title":"Algorithmic Foundations of Robotics XII","year":"2020"},{"issue":"12","key":"B70","doi-asserted-by":"crossref","first-page":"1341","DOI":"10.1177\/0278364917731821","article-title":"Haptic rendering and interactive simulation using passive midpoint integration","volume":"36","year":"2017","journal-title":"Int. J. Robot. Res."},{"issue":"1","key":"B71","doi-asserted-by":"crossref","first-page":"2828","DOI":"10.1016\/S1474-6670(17)58105-7","article-title":"Energy based control of a class of underactuated mechanical systems","volume":"29","year":"1996","journal-title":"IFAC Proc"},{"key":"B72","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1146\/annurev-control-042920-094829","article-title":"An historical perspective on the control of robotic manipulators","volume":"5","year":"2022","journal-title":"Annu. Rev. Control Robot. Auton. Syst."},{"key":"B73","article-title":"Deep Lagrangian networks: using physics as model prior for deep learning","year":"2019","journal-title":"In The Seventh International Conference on Learning Representations"},{"key":"B74","doi-asserted-by":"crossref","first-page":"7718","DOI":"10.1109\/IROS40897.2019.8968268","article-title":"Deep Lagrangian networks for end-to-end learning of energy-based control for under-actuated systems","volume-title":"2019 IEEE\/RSJ International Conference on Intelligent Robots and Systems (IROS)","year":"2019"},{"key":"B75","first-page":"21910","article-title":"Extending Lagrangian and Hamiltonian neural networks with differentiable contact models","volume-title":"Advances in Neural Information Processing Systems 34","year":"2021"},{"issue":"6","key":"B76","doi-asserted-by":"crossref","first-page":"101","DOI":"10.1109\/MCS.2008.929425","article-title":"Revisiting the LuGre friction model","volume":"28","year":"2008","journal-title":"IEEE Control Syst. Mag."},{"key":"B77","article-title":"Fast and accurate data-driven simulation framework for contact-intensive tight-tolerance robotic assembly tasks","year":"2022"},{"key":"B78","first-page":"378","article-title":"Data-augmented contact model for rigid body simulation","volume-title":"Proceedings of the 4th Annual Learning for Dynamics and Control Conference","year":"2022"},{"key":"B79","first-page":"2279","article-title":"ContactNets: learning discontinuous contact dynamics with smooth, implicit representations","volume-title":"Proceedings of the 2020 Conference on Robot Learning","year":"2021"},{"issue":"5","key":"B80","doi-asserted-by":"crossref","first-page":"2780","DOI":"10.1109\/LRA.2023.3257707","article-title":"Differentiable physics simulation of dynamics-augmented neural objects","volume":"8","year":"2023","journal-title":"IEEE Robot. Autom. Lett."},{"key":"B81","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1146\/annurev-control-060117-104848","article-title":"Toward robotic manipulation","volume":"1","year":"2018","journal-title":"Annu. Rev. Control Robot. Auton. Syst."},{"key":"B82","first-page":"4470","article-title":"Graph networks as learnable physics engines for inference and control","volume-title":"Proceedings of the 35th International Conference on Machine Learning","year":"2018"},{"key":"B83","first-page":"4509","article-title":"Interaction networks for learning about objects, relations and physics","volume-title":"Advances in Neural Information Processing Systems 29","year":"2016"},{"key":"B84","first-page":"1157","article-title":"Graph network simulators can learn discontinuous, rigid contact dynamics","volume-title":"Proceedings of the 6th Conference on Robot Learning","year":"2023"},{"key":"B85","article-title":"Sample efficient grasp learning using equivariant models","volume-title":"Robotics: Science and Systems XVIII","year":"2022"},{"key":"B86","first-page":"427","article-title":"SE(2)-equivariant pushing dynamics models for tabletop object manipulations","volume-title":"Proceedings of the 6th Conference on Robot Learning","year":"2023"},{"key":"B87","first-page":"21910","article-title":"Learning physical dynamics with subequivariant graph neural networks","volume-title":"Advances in Neural Information Processing Systems 34","year":"2022"},{"key":"B88","first-page":"2995","article-title":"Sample efficient dynamics learning for symmetrical legged robots: leveraging physics invariance and geometric symmetries","volume-title":"2023 IEEE International Conference on Robotics and Automation","year":"2023"},{"key":"B89","article-title":"Adaptive-control-oriented meta-learning for nonlinear systems","volume-title":"Robotics: Science and Systems XVII","year":"2021"},{"issue":"10\u201311","key":"B90","first-page":"1123","article-title":"Learning stabilizable nonlinear dynamics with contraction-based regularization","volume":"40","year":"2021","journal-title":"Int. J. Robot. Res."}],"container-title":["Annual Review of Control, Robotics, and Autonomous Systems"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.annualreviews.org\/content\/journals\/10.1146\/annurev-control-061523-102310?crawler=true&mimetype=application\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,10,3]],"date-time":"2025-10-03T04:06:16Z","timestamp":1759464376000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.annualreviews.org\/content\/journals\/10.1146\/annurev-control-061523-102310"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,7,10]]},"references-count":90,"journal-issue":{"issue":"1","published-print":{"date-parts":[[2024,7,10]]}},"URL":"https:\/\/doi.org\/10.1146\/annurev-control-061523-102310","relation":{},"ISSN":["2573-5144"],"issn-type":[{"value":"2573-5144","type":"electronic"}],"subject":[],"published":{"date-parts":[[2024,7,10]]}}}