A simple learning strategy for high-speed quadrocopter multi-flips

By Sergei Lupashin, Angela P. Schoellig, Michael Sherback, Raffaello D'Andrea. IEEE International Conference on Robotics and Automation, 2010

392 citations recorded by the source corpus.

We describe a simple and intuitive policy gradient method for improving parametrized quadrocopter multi-flips by combining iterative experiments with information from a first-principles model. We start by formulating an N-flip maneuver as a five-step primitive with five adjustable parameters. Optimization using a low-order first-principles 2D vertical plane model of the quadrocopter yields an initial set of parameters and a corrective matrix. The maneuver is then repeatedly performed with the vehicle. At each iteration the state error at the end of the primitive is used to update the maneuver parameters via a gradient adjustment. The method is demonstrated at the ETH Zurich Flying Machine Arena testbed on quadrotor helicopters performing and improving on flips, double flips and triple flips.

Subfields: Computational Theory and Mathematics, Control and Systems Engineering

View Paper: https://www.semanticscholar.org/paper/d9c54dd53a5cdf8aa41cc9f4e2d8b0aa002d5523

How researchers see this paper

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

In Computational Theory and Mathematics

In Control and Systems Engineering

Citations in the corpus

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.”
Vision and IMU Data Fusion: Closed-Form Solutions for Attitude, Speed, Absolute Scale, and Bias Determination (2012, neutral): “The data are provided together with a reliable ground-truth, which has been obtained by performing the experiments at the ETH Zurich Flying Machine Arena [17], which is equipped with a Vicon motion capture system.”
Data-Driven MPC for Quadrotors (2021, neutral): “Other approaches use iterative learning control to perform highly aggressive trajectories [9], but they are constrained to a single maneuver and do not generalize.”
Generation of collision-free trajectories for a quadrocopter fleet: A sequential convex programming approach (2012, background): “More details on the testbed can be found in [20].”
Onboard IMU and monocular vision based control for MAVs in unknown in- and outdoor environments (2011, methodology): “Arena” [23] to obtain ground truth for the indoor experiments.”
Vision-Controlled Micro Flying Robots: From System Design to Autonomous Navigation and Mapping in GPS-Denied Environments (2014, background): “Using this infrastructure, several groups have demonstrated aggressive maneuvers and impressive acrobatics [4], [5].”
Camera-based navigation of a low-cost quadrocopter (2012, neutral): “Significant progress has been made, recent examples include aggressive flight maneuvers [1, 2], ping-pong [3] and collaborative construction tasks [4].”
Minimum snap trajectory generation and control for quadrotors (2011, background): “Some work in this area has addressed aerobatic maneuvers [ 3 , 6, 9, 10].”
Estimation, Control, and Planning for Aggressive Flight With a Small Quadrotor With a Single Camera and IMU (2017, background): “In [9], the authors focus on the hovering stabilization problem after a flip maneuver, whereas [8], [10] address the learning and adaptive iteration of the control law, which is refined based on different flight trials.”
Mixed-integer quadratic program trajectory generation for heterogeneous quadrotor teams (2012, background): “In particular, there have been many recent advances in the design [6], control [9] and planning [7] for quadrotors, rotorcrafts with four rotors.”
Tracking a depth camera: Parameter exploration for fast ICP (2011, methodology): “We run the experiment in a room equiped with a Vicon tracking system7 [12].”
Aerodynamics and Control of Quadrotors (2017, methodology): “search since 2004 Mahony et al. [2008]; Mellinger et al.”
A General Safety Framework for Learning-Based Control in Uncertain Robotic Systems (2017, background): “quadrotor in [6] was enclosed in a safety net.”
Opportunities and challenges with autonomous micro aerial vehicles (2012, methodology): “Iterative learning has been used effectively in [26, 29] for acrobatic maneuvers.”
Automatic re-initialization and failure recovery for aggressive flight with a monocular vision-based quadrotor (2015, background): “Autonomous Micro Aerial Vehicles (MAVs) will soon play a major role in remote inspection and search-andrescue missions.”