Qualcomm Chips Make Drones Smart
Written by Harry Fairhead   
Friday, 01 January 2016

The past year, 2015, has seen amazing advances in AI and it looks as if the experimental systems are ready to make it into consumer devices in 2016. Qualcomm is planning to release a new device that will make drones smarter and cheaper.

The GPU is usually thought of as something that speeds up 3D graphics operations, but many of the same computations are useful in computer vision tasks. As a result Qualcomm have designed a special "drone" version of its Snapdragon system on a chip.

 

 

QualcomDrone

 

 

The Snapdragon includes an ARM multicore processor and a GPU. It isn't difficult to guess that the new chipset aimed at drones makes heavy use of the GPU to perform computer vision tasks but there isn't any hard information. However we do have this the "Sneak Peek" video: 

 

 

It is a preview of what Qualcomm hopes to be demoing at the 2016 Consumer Electronics show in Las Vegas. The key features are an optical flow camera, visual inertial odometery, motion planning, obstacle mapping and more. Basically what this means is that the drone can see where it is and auto-guide itself around obstacles. 

If you want to know what Visual Inertial Odometry is all about see the next video:

 

  

It should make high quality drones cheaper and more capable. Exactly what applications this might open up is a matter of guesswork. 

It is worth pointing out that the same chip set could come in handy for any autonomous device that needs to navigate. A few software tweaks and it could be useful in anything from a house security guard robot, to a self driving car.

If the price of the chipset comes down you aren't even going to be restricted to a single unit per robot. Now imagine the robot vac that doesn't bag its head against the furniture and manages to avoid getting tangled in the cables while working out which room in your house needs the next attention.

 

 qualcomdroneicon

AI and robotics will become a lot easier as the algorithms are transferred to the hardware and all that is left for the software to do is integrate the system. This is how robots will become a commodity item. 

 

More Information

Snapdragon Flight robotics dev platform

Related Articles

Drones Build A Rope Bridge

Flying Ball Drone Wins $1 Million Prize

Not A Drone Army But a Drone Orchestra

Nano Quadcopter - Your Personal Flying Robot

Juggling Quadrotors

Flying Neural Net Avoids Obstacles

Quadrotors Play Ball

You're Never Alone With A Joggobot Quadrotor

A Swarm of Quadrotors 

 

To be informed about new articles on I Programmer, sign up for our weekly newsletter, subscribe to the RSS feed and follow us on, Twitter, FacebookGoogle+ or Linkedin

 

Banner


ZLUDA Ports CUDA Applications To AMD GPUs
18/04/2024

ZLUDA is a translation layer that lets you run unmodified CUDA applications with near-native performance on AMD GPUs. But it is walking a fine line with regards to legality.



Android 15 Developer Preview Updated
25/03/2024

Google has released Android 15 Developer Preview 2 with changes including better handling of automatic language switching and updates for OpenJDK 17.


More News

raspberry pi books

 

Comments




or email your comment to: comments@i-programmer.info

 

Last Updated ( Thursday, 31 December 2015 )