Receding Horizon Differential Dynamic Programming

By Yuval Tassa, Tom Erez, William D. Smart. 2007

151 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Reinforcement Learning

Citations in the corpus

Learning Continuous Control Policies by Stochastic Value Gradients (2015, background): “Swimmer is a chain body with multiple links immersed in a fluid environment with drag forces that allow the body to propel itself [5, 27].”
Embed to Control: A Locally Linear Latent Dynamics Model for Control from Raw Images (2015, methodology): “When combined with receding horizon control [3], and machine learning methods for learning approximate system models, such algorithms are a powerful tool for solving complicated control problems [3, 4, 5, 6]; however, they either rely on a known system model or require the design of relatively low-d”
Ensemble Bayesian Decision Making with Redundant Deep Perceptual Control Policies (2018, methodology): “trol Differential Dynamic Programming (iLQG/MPC-DDP) [22] algorithm was used as an expert.”
Agile Autonomous Driving using End-to-End Deep Imitation Learning (2017, methodology): “Upon convergence, DDP returns a locally optimal control sequence { ˆ a ∗ t , ..., ˆ a ∗ t + T h − 1 } , and the MPC expert executes the ﬁrst action in the sequence as the expert’s action at time t (i.e. a ∗ t = ˆ a ∗ t ).”
Graph networks as learnable physics engines for inference and control (2018, methodology): “While this iLQG planner does not work optimally when the dynamics involve complex contacts, for relatively smooth dynamics as found in the swimmer, differential dynamic programming (DDP) style approaches works well (Tassa et al., 2008).”
Agile Off-Road Autonomous Driving Using End-to-End Deep Imitation Learning (2017, methodology): “The computation is realized by the trajectory optimization algorithm, Differential Dynamic Programming (DDP) [23]: ).”
Early Failure Detection of Deep End-to-End Control Policy by Reinforcement Learning (2019, background): “Iterative LQG/Differential Dynamic Programming (iLQG/DDP) [9] is a dynamic programming method which provides locally optimal solutions to the optimal control problem.”
Reinforcement learning and adaptive dynamic programming for feedback control (2009, background): “Note that performing (74) on (71) yields exactly (54).”
Robust Locally-Linear Controllable Embedding (2017, background): “Algorithms from this class such as differential dynamic programming (DDP) [3], iterative linear quadratic regulator (iLQR) [8], and iterative linear quadratic Gaussian (iLQG) [14] have been successfully applied to a variety of complex control problems [1, 13, 7, 9].”
Learning variable impedance control (2011, background): “The work on Receding Horizon DDP (Tassa et al. 2008) provided an alternative and rather efficient way of computing local optimal feedback controls.”
Linear Bellman combination for control of character animation (2009, methodology): “Model predictive control may be preferable to combination when feasible as in autonomous control of aircraft [Milam 2003] or character models over short time horizons [Tassa et al. 2008; da Silva et al. 2008a].”
Probabilistic Differential Dynamic Programming (2014, background): “In the last decade, variations of DDP have been proposed in both control and machine learning communities [2][3][4][5][6].”
A probabilistic approach to mixed open-loop and closed-loop control, with application to extreme autonomous driving (2010, methodology): “While there are many possible methods for minimizing the cost function (1), one method that is particularly wellsuited to this task is the Linear Quadratic Regulator (LQR) algorithm (sometimes also referred to as the Linear Quadrat ic Tracker when the task is specifically to follow some target traje”
A robot path planning framework that learns from experience (2012, background): “The most common approaches to solving such tasks is to construct a motion plan using no prior knowledge of the task and the environment; a planning algorithm is only given the robot and environment models along with the start and goal configurations and asked to generate a path.”
Approximate real-time optimal control based on sparse Gaussian process models (2014, methodology): “For performing online trajectory optimization we employ a receding horizon variant of the iterative LQR (iLQR) algorithm [5], [14].”
Space-indexed dynamic programming: learning to follow trajectories (2008, methodology): “An alternative approach is to use a local control method, such as DDP, in order to estimate the value function along several trajectories, and use these local estimates to build an approximate global model of the value function (Atkeson, 1994; Tassa et al., 2007).”
Safe end-to-end imitation learning for model predictive control (2018, unknown): “Iterative LQG/Differential Dynamic Programing (iLQG/DDP) [11] is a dynamic programming method which provides locally optimal solutions to the optimal control problem.”