Learning vehicular dynamics, with application to modeling helicopters

By Pieter Abbeel, Varun Ganapathi, Andrew Y. Ng. 2005

92 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.

Citations in the corpus

Dual Policy Iteration (2018, methodology): “We tested our approach on several MDPs: (1) a set of random discrete MDPs (Garnet problems [27]) (2) Cartpole balancing [34], (3) Helicopter Aerobatics (Hover and Funnel) [1], (4) Swimmer, Hopper and Half-Cheetah from the MuJoCo physics simulator [35].”
Improving Multi-Step Prediction of Learned Time Series Models (2015, background): “They propose either a simpler but still non-convex approximation (Abbeel, Ganapathi, and Ng 2005) that can be iteratively optimized or Expectation Maximization (Abbeel and Ng 2005b) to ﬁnd a local optimum to the multi-step error.”
Approximate real-time optimal control based on sparse Gaussian process models (2014, background): “Their algorithm for model identification [25] is very efficient but requires considerable expert knowledge about the system to be controlled.”
Autonomous Helicopter Aerobatics through Apprenticeship Learning (2010, methodology): “We have proposed a method which performs as well as CIFER(1) when learning linear models for our helicopters, and which does allow application to the non-linear setting (Abbeel et al. 2006a).”
An Application of Reinforcement Learning to Aerobatic Helicopter Flight (2006, methodology): “Following [1] we learn a model from flight data that predicts accelerations as a function of the current state and inputs.”
Deep learning helicopter dynamics models (2015, methodology): “Equation 2 results in a linear or nonlinear least squares problem depending on the form of f (s,u,θ).”
Feedback controller parameterizations for Reinforcement Learning (2011, background): “Most uses of learning which have been successful on hardware involved learning an open-loop trajectory or control tape (see, for example [2], [13], [20], [23]).”
Improved Learning of Dynamics Models for Control (2016, background): “This comparison shows the difference between [20] (no exploration) and [18] (constant fraction exploration).”
Neuroevolutionary reinforcement learning for generalized helicopter control (2009, background): “Since such policies might still occasionally crash or behave erratically, we may need to consider more elaborate model representations, as the one employed here is known to be unreliable outside the region in which the helicopter strives to hover [1].”
Neuroevolutionary reinforcement learning for generalized control of simulated helicopters (2011, methodology): “Numerous methods based on apprenticeship learning [1, 2, 3, 84] and policy-search reinforcement-learning [6, 49, 57], including evolutionary approaches [18, 19], have been developed.”