Real Time Robotics
PoseCNN, Dense Articulated Real time tracking and Riemannian Motion Policies
The challenge of the kitchen environment is that the robotic arms need to work in unstructure environment with real time detection of the object and generate the fast human-liked motion in order to acheive the goal.
Perception side (แนวความคิด)
- PoseCNN
— taking the image in front of robot
— detecting where object are in the space - Dense Articulated Real time tracking
— continuous perceptual (เกี่ยวกับการรับรู้ได้) feedback
— where object are in any moment - Riemannian Motion Policies
— generate real time fast reactive, adaptive
— human-like motion
Automated Kitchen Price
Our futuristic kitchen is feasible! It is available right now with the price of $330000!
As a programmer and software engineer, 10M per kitchen will make a fortune of your life. Just keep thinking about 10M baht, what if you have 1000 order around the globe. 10,000M baht will suddenly be yours. Joining the force of your ten friends which can be shared equally of 1,000M baht, a sudden millionaire, tycoon will be you!
How the automated kitchen companies make $330000 from selling the automated kitchen?
Information courtesy of google.com and NVIDIA’s video.(lol) This should be edited.
PoseCNN is 6d pose object detection
It is basically an 8 mins video which tell you the experience, and the steps to do the 3D detection.
This video explained about RGB D data.