Control-limited differential dynamic programming

By Yuval Tassa, Nicolas Mansard, E. Todorov. 2014

568 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Biomedical Engineering, Computer Vision and Pattern Recognition

How researchers see this paper

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

In Artificial Intelligence

In Biomedical Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

Model-based reinforcement learning with parametrized physical models and optimism-driven exploration (2015, methodology): “The velocity cost is weighted by a diagonal weight matrix Qv, and the controls are penalized both after squashing under R and before the squashing, under P, as recommended in prior work [30].”
Graph networks as learnable physics engines for inference and control (2018, methodology): “As an MPC baseline, with a pre-speciﬁed physical model, we used a Differential Dynamic Programming algorithm (Tassa et al., 2008; 2014) that had access to the ground-truth Mujoco model.”
A Survey of Motion Planning and Control Techniques for Self-Driving Urban Vehicles (2016, methodology): “This method is known as the shooting method, and a version of this approach has been applied to planning parking maneuvers in [83].”
Neural probabilistic motor primitives for humanoid control (2018, methodology): “These can be obtained by sample-based optimization (e.g. Ding et al. (2015)) or by differential dynamic programming (Morimoto & Atkeson, 2003; Tassa et al., 2012; 2014).”
Plan Online, Learn Offline: Efficient Learning and Exploration via Model-Based Control (2018, background): “Trajectory optimization and model predictive control (MPC) have a long history in robotics and control systems [11, 12]1.”
An Online Learning Approach to Model Predictive Control (2019, methodology): “Various MPC algorithms have been proposed, using tools ranging from constrained optimization techniques [7, 20, 27] to sampling-based techniques [30].”
Parting with Misconceptions about Learning-based Vehicle Motion Planning (2023, methodology): “Trajectories in nuPlan are simulated by iteratively retrieving actions from a LQR controller [63] and propagating the ego vehicle with a kinematic bicycle model [64, 65].”
Reinforcement learning for non-prehensile manipulation: Transfer from simulation to physical system (2018, unknown): “Control methods that depend on physical models, whether or not with reduced or simplified models, are able to produce dynamic actions [9], [10], [11],”
Manipulators and Manipulation in high dimensional spaces (2016, methodology): “In light of this observation, control limits were enforced on the controller using [99] and iterative model learning methods were used to refit the pneumatic models to the restricted range of operation.”
A Review of Robot Learning for Manipulation: Challenges, Representations, and Algorithms (2019, background): “, 2016) and Differential Dynamic Programming (Tassa et al., 2014; Yamaguchi and Atkeson, 2015), in which nonlinear dynamics and nonquadratic costs can be approximated as locally linear and locally quadratic, respectively.”
Meta Learning via Learned Loss (2019, background): “6-a, when compared to a classical iLQRbased trajectory optimization [29].”
Synthesizing Neural Network Controllers with Probabilistic Model-Based Reinforcement Learning (2018, background): “Trajectory optimization approaches [7], [8], [9] can produce highly effective controllers for complex robotic systems when precise analytical models are available.”
Differentiable MPC for End-to-end Planning and Control (2018, methodology): “To solve the MPC problem in (10) and reach the ﬁxed point in (11), we use the box-DDP heuristic [Tassa et al., 2014].”
Optimizing Simulations with Noise-Tolerant Structured Exploration (2018, methodology): “In section III, we consider three classic continuous control problems: Cartpole and Acrobot balancing, and a car maneuvering problem studied in [22] (see Figure 2, right).”