Robust walking based on MPC with viability guarantees
Max Planck Institute for Intelligent Systems（马克斯·普朗克智能系统研究所）；University of Trento（特伦托大学）；K. N. Toosi University of Technology（K. N. 图西理工大学）；New York University（纽约大学）
Mohammad Hasan Yeganegi, Majid Khadiv, Andrea Del Prete, S. Ali A. Moosavian, Ludovic Righetti
Model predictive control (MPC) has shown great success for controlling complex systems such as legged robots. However, when closing the loop, the performance and feasibility of the finite horizon optimal control problem (OCP) solved at each control cycle is not guaranteed anymore. This is due to model discrepancies, the effect of low-level controllers, uncertainties and sensor noise. To address these issues, we propose a modified version of a standard MPC approach used in legged locomotion with viability (weak forward invariance) guarantees. In this approach, instead of adding a (conservative) terminal constraint to the problem, we propose to use the measured state projected to the viability kernel in the OCP solved at each control cycle. Moreover, we use past experimental data to find the best cost weights, which measure a combination of performance, constraint satisfaction robustness, or stability (invariance). These interpretable costs measure the trade off between robustness and performance. For this purpose, we use Bayesian optimization (BO) to systematically design experiments that help efficiently collect data to learn a cost function leading to robust performance. Our simulation results with different realistic disturbances (i.e. external pushes, unmodeled actuator dynamics and computational delay) show the effectiveness of our approach to create robust controllers for humanoid robots.
Variable Horizon MPC with Swing Foot Dynamics for Bipedal Walking Control
Max Planck Institute for Intelligent Systems（马克斯·普朗克智能系统研究所）；Amirkabir University of Technology（阿米尔卡比尔理工大学）；Tandon School of Engineering, New York University（纽约大学坦登工程学院）
Discrete time model predictive control for humanoid walking with step adjustment
Robert Bosch Centre for Cyber Physical Systems, Indian Institute of Science（印度科学学院罗伯特·博世网络物理系统中心）；Computer Science and Automation, Indian Institute of Science（印度科学学院计算机科学与自动化系）
Fast Online Optimization for Terrain-Blind Bipedal Robot Walking with a Decoupled Actuated SLIP Model
Imperial College London（帝国理工学院）
我们提出了一个高度反应控制器，使双足机器人盲目地走过各种不平坦的地形，同时抵抗推动。高级运动规划器使用解耦致动弹簧加载倒立摆（aSLIP）模型对脚步位置和质心（CoM）高度进行快速在线优化。解耦的aSLIP模型简化了原始aSLIP，在水平状态下具有线性倒立摆（LIP）动力学，在垂直状态下具有弹簧动力学。运动规划可以用离散时间模型预测控制（MPC）来表示，并在1 k ~HZ的频率下求解。使用降阶模型的运动规划器的输出被馈送到基于逆动力学的全身控制器中以用于在机器人上执行。该控制器的一个关键结果是机器人的脚是顺应性的，这进一步扩展了机器人对未观察到的地形变化的鲁棒性的能力。我们评估我们的方法在模拟与双足机器人SLIDER。实验结果表明，该机器人能够在斜坡、波浪场和楼梯等各种不平坦地形上实现盲走。它也可以抵抗推，而走在不平坦的地形。
Bipedal Robust Walking on Uneven Footholds: Piecewise Slope LIPM with Discrete Model Predictive Control
Faculty of Computing, Harbin Institute of Technology（哈尔滨工业大学计算学部）；State Key Laboratory of Robotics and System, Harbin Institute of Technology（哈尔滨工业大学机器人技术与系统国家重点实验室）
Extended Hybrid Zero Dynamics for Bipedal Walking of the Knee-less Robot SLIDER
Robot Intelligence Lab, Dyson School of Design Engineering, Imperial College London, UK（机器人智能实验室，设计工程学院，伦敦帝国理工学院，英国）；Department of Aeronautics, Imperial College London, UK（航空系，伦敦帝国理工学院，英国）
Dynamic Modeling and MPC for Locomotion of Tendon-Driven Soft Quadruped
Department of Electrical Engineering, IIT Gandhinagar（电子工程系，印度加尔各直辖区理工学院）；Department of Electronics and Computer Engineering, Sreenidhi Institute of Science and Technology（电子与计算机工程系，斯里尼迪学院科学与技术研究所）；Department of Mechanical Engineering, IIT Gandhinagar（机械工程系，印度加尔各直辖区理工学院）
Real-time Whole-body Model Predictive Control for Bipedal Locomotion with a Novel Kino-dynamic Model and Warm-start Method
Department of Intelligence and Information, Graduate School of Convergence Science and Technology, Seoul National University（智能信息系，融合科学与技术研究生院，首尔国立大学）；Advanced Institutes of Convergence Technology(AICT)（融合技术先进研究院）
Dynamic Mirror Descent based Model Predictive Control for Accelerating Robot Learning
Indian Institute of Science Bangalore（印度科学理工学院）；Manipal Institute of Technology（马尼帕尔理工学院）；National Institute of Technology Karnataka, Surathkal（卡纳塔克邦国立技术学院苏拉特卡尔）
Whole-Body Model-Predictive Control of Legged Robots with MuJoCo
Department Aeronautics and Astronautics, Massachusetts Institute of Technology（航空宇航与工程力学系，麻省理工学院）；Google DeepMind（谷歌DeepMind）；Department of Electrical and Computer Engineering, Carnegie Mellon University（电气与计算机工程系，卡内基梅隆大学）；Robotics Institute, Carnegie Mellon University（机器人研究所，卡内基梅隆大学）
TD-GRPC: Temporal Difference Learning with Group Relative Policy Constraint for Humanoid Locomotion
University of Texas at Arlington（德克萨斯大学阿灵顿分校）；VinRobotics；Intelligent Autonomous Systems Lab, TU Darmstadt（图宾根大学智能自主系统实验室）；German Research Center for AI (DFKI)（德国人工智能研究中心）；Automation & Control Institute (ACIN), TU Wien（维也纳技术大学自动化与控制研究所）
ECO: Energy-Constrained Optimization with Reinforcement Learning for Humanoid Walking
State Key Laboratory of General Artificial Intelligence, Beijing Institute for General Artificial Intelligence (BIGAI)（通用人工智能国家重点实验室，北京通用人工智能研究院（BIGAI））；Department of Automation, Tsinghua University（自动化系，清华大学）；Department of Automation, University of Science and Technology of China（自动化系，中国科学技术大学）；Department of Computer Science, Harbin Institute of Technology（计算机科学系，哈尔滨工业大学）；Institute for Artificial Intelligence and School of Artificial Intelligence, Peking University（人工智能研究院和人工智能学院，北京大学）
A Three-Stage Offline SDRE-Based Control Framework for Human Motion Reproduction on a Suspended Bipedal Robot
Department of Applied Mathematics, National Yang Ming Chiao Tung University（应用数学系，阳明交通大学）；Department of Electrical Engineering, National Central University（电气工程系，国立中央大学）
