AI News, MIT Drone Flies Autonomously While Avoiding Obstacles

MIT Drone Flies Autonomously While Avoiding Obstacles

In just about every video featuring drones making aggressive maneuvers around obstacles there’s some amount of “cheating” going on.

With that goal in mind—andjust US $1700 in hardware—MIT PhD student Andrew Barryhas managed to fire a fixed-wing drone at some trees and not hit them, using only two cellphones worth of onboard computing hardware and real-time image processing.

Cameras that have the necessary frame rate and resolution to enable you to clearly see obstacles in the first place pour out ahumongousnumber of pixels, each one of which needs to be analyzed to determine whether the drone has to worry about it.

Using stereo filtering from a pair of 376-by-240-pixel resolution, 120-frames-per-second cameras spaced 34 centimetersapart, the drone focuses its attention (for robots, this equates to obstacle avoidance algorithms) on pixels that are about 10 meters away and nothing else.

It saves these pixels in memory, and the next image (taken 8.3 cm later if the drone is flying at 10 meters per second) adds more pixels beyond the previous set.

Because the detection horizon for obstacles is so short, the drone might not have enough time to take an effective evasive maneuver if (say) it approaches a building, but for trees and other relatively small and discrete obstacles, it seems like it should be able to continue avoiding things indefinitely.

As the researchers point out, the detection horizon is primarily constrained by computer processing power, so as that improves, they’ll be able to scan multiple depths to plan more complex paths around multiple obstacles at varying distances.

Autonomous vision-based drone navigation demo

Qualcomm's Snapdragon Flight platform demonstrates safe autonomous flight around obstacles using only onboard sensors plus 3D obstacle mapping and path planning algorithms. This footage was...

MINECON Earth 2017 Livestream

Pre-show starts immediately MINECON Earth show starts at 25:08 Post-show starts at 2:02:06 MINECON Earth is here! The ultimate Minecraft liveshow, beamed to every corner of the globe! Watch...

DARCE WEIR & LEE LUSTIG & THEIR SECRET SPACE DOCUMENTARY

I interviewed filmmakers Lee Lustig and Darce Weir about their new documentary on the secret space program and NASA called: BEYOND THE SPECTRUM Show their documentary to friends and family!...

Robust, Visual-Inertial State Estimation: from Frame-based to Event-based Cameras

I will present the main algorithms to achieve robust, 6-DOF, state estimation for mobile robots using passive sensing. Since cameras alone are not robust enough to high-speed motion and high-dynami...

MIT 6.S094: Deep Reinforcement Learning for Motion Planning

This is lecture 2 of course 6.S094: Deep Learning for Self-Driving Cars taught in Winter 2017. This lecture introduces types of machine learning, the neuron as a computational building block...

MIT 6.S094: Deep Learning

This is lecture 1 of course 6.S094: Deep Learning for Self-Driving Cars (2018 version). This class is free and open to everyone. It is an introduction to the practice of deep learning through...

The future of artificial intelligence and self-driving cars

Stanford professors discuss their innovative research and the new technologies that will transform lives in the 21st century. At a live taping of The Future of Everything, a SiriusXM radio...

Google Cloud Next '17 - Day 1 Keynote

Hear from Diane Greene, SVP of Google Cloud; Sundar Pichai, CEO of Google; Eric Schmidt, Chairman of Alphabet and Fei-Fei Li, Chief Scientist for Google Cloud Machine Learning and AI and Professor...

GIS in the Field: Using LiDAR Makes Sense

Presented by Steve Nechero, Supervisory Cartographer, Branch Leader, USDA NRCS National Geospatial Management Center, and USDA NRCS GIS Specialists Aaron Dushku, Massachusetts; Reed Sims, Vermont;...

PlayStation® Live From E3 2017 | Day 3

PlayStation® Live From E3 2017 | Day 3.