Adam Bry

P

End-to-end learning of geometry and context for deep stereo regression

Proceedings of the IEEE international conference on computer vision

Mar 2017

We propose a novel deep learning architecture for regressing disparity from a rectified pair of stereo images. We leverage knowledge of the problem’s geometry to form a cost volume using deep feature representations. We learn to incorporate contextual information using 3-D convolutions over this volume. Disparity values are regressed from the cost volume using a proposed differentiable soft argmin operation, which allows us to train our method end-to-end to sub-pixel accuracy without any additional post-processing or regularization. We evaluate our method on the Scene Flow and KITTI datasets and on KITTI we set a new state-of-the-art benchmark, while being significantly faster than competing approaches.

T

Aggressive flight of fixed-wing and quadrotor aircraft in dense indoor environments

The International Journal of Robotics Research

Mar 2015

In this paper, we describe trajectory planning and state estimation algorithms for aggressive flight of micro aerial vehicles in known, obstacle-dense environments. Finding aggressive but dynamically feasible and collision-free trajectories in cluttered environments requires trajectory optimization and state estimation in the full state space of the vehicle, which is usually computationally infeasible on realistic time scales for real vehicles and sensors. We first build on previous work of van Nieuwstadt and Murray [51] and Mellinger and Kumar [38], to show how a search process can be coupled with optimization in the output space of a differentially flat vehicle model to find aggressive trajectories that utilize the full maneuvering capabi

I

State estimation for aggressive flight in GPS-denied environments using onboard sensing

IEEE International Conference on Robotics and Automation

May 2012

In this paper we present a state estimation method based on an inertial measurement unit (IMU) and a planar laser range finder suitable for use in real-time on a fixed-wing micro air vehicle (MAV). The algorithm is capable of maintaing accurate state estimates during aggressive flight in unstructured 3D environments without the use of an external positioning system. Our localization algorithm is based on an extension of the Gaussian Particle Filter. We partition the state according to measurement independence relationships and then calculate a pseudo-linear update which allows us to use 20x fewer particles than a naive implementation to achieve similar accuracy in the state estimate. We also propose a multi-step forward fitting method to id

R

Polynomial trajectory planning for aggressive quadrotor flight in dense indoor environments

Robotics Research The th International Symposium ISRR

We explore the challenges of planning trajectories for quadrotors through cluttered indoor environments. We extend the existing work on polynomial trajectory generation by presenting a method of jointly optimizing polynomial path segments in an unconstrained quadratic program that is numerically stable for high-order polynomials and large numbers of segments, and is easily formulated for efficient sparse computation. We also present a technique for automatically selecting the amount of time allocated to each segment, and hence the quadrotor speeds along the path, as a function of a single parameter determining aggressiveness, subject to actuator constraints. The use of polynomial trajectories, coupled with the differentially flat representa

About

Detail

Timeline

Résumé