Here are some amazing software projects which warrant being assimilated into MRL !  RESISTANCE IS FUTILE !  

Please add your favorite open source Java project to the list as a comment !

Who's up next ?

JHLabs - visual BufferedImage filters - http://www.jhlabs.com/ip/filters/index.html

SLAM

Inverse Kinematics

Speech Deep Learning

Physics & Kinetics

AI

Arduino/AdaFruit Sensors

Competitions

Q-Learging

raver1975: so the next time the robot is in a similar state, it will know what the best action
is
GroG: like subsumption architecture with memory
raver1975: the robot is in a state, and there is a list of actions the robot can take while in that state. The robot leanrs through trial and error, which actions are the best to take when it is in a certain state
raver1975: but if it is holding a small child, and it takes the "eat" action, the robot is rewarded
raver1975: the grids don't have to be robot positions, they are robot "states", and the 4 directions are the 4 "actions" the robot can take in each state. so if the robot is holding poison, and it takes the action "eat", it is bad.
Vision & Video H.264 

 

3D Representations

Looking for a 3D simulator + physics engine for WebGL - this one I think combines the bullet engine with its own framework more research needed - http://cjcliffe.github.io/CubicVR.js/cubicvr/samples/vehicle_physics_demo/stunt_track1.html

more about it - http://www.cubicvr.org/45-news/webgl/82-cubicvr-js-vehicle-physics-demo

Review of different WebGL js frameworks - http://weblog.benjaminsommer.com/blog/2012/05/13/comparison-of-webgl-framework-apis/

 

GUI 2.0 for MRL needs to be based on HTML 5, WebGL & WebSockets !
Above is an amazing demonstration of what can be done with WebGL (http://madebyevan.com/webgl-water/)

Here is a kinect point cloud display demo - http://www.mrdoob.com/lab/javascript/webgl/kinect/

These and many more demos are @ http://www.chromeexperiments.com/

Point Cloud Library PCL - http://pointclouds.org

WebGUI over Smart Phones

The new gui will be web based allowing it to work on any phone, tablet, or computer.  MRL's GUI system is already client / server based - so it will fairly straight forward to make the existing architecture work with a new web base gui view.

To mitigate the differences between different mobile operating systems - phonegap or similar framework will probably be used.

Distributed Computing

Interface Related

Video or OpenGL related

 

SIMULATORS

KINECT

OCR

Speech Recognition (Speech to Text)

Speech Synthesis (Text to Speech)

RXTX Replacement

Genetic Programming

Wavefront Algorithm

TLD+

Security

Others

Point Cloud Reconstruction 

RasPi

I2C
Cortex Algorithms

Ring Buffer

OpenCV

USB over IP

Machine Learning 
Cloud Robotics
MineCraft Display

morrows_end

11 years 2 months ago

To add to the wavefront algorithm, here are a few others that may be interesting:

Bug1 and Bug2: http://www.cs.cmu.edu/~motionplanning/lecture/Chap2-Bug-Alg_howie.pdf

A* (in python): http://www.laurentluce.com/posts/solving-mazes-using-python-simple-recursivity-and-a-search/

Simple maze-solving optomization (think micromouse competitions): http://campus.murraystate.edu/academic/faculty/bob.pilgrim/245/lab_06.html

 

We've already got solr support...  I'll try and get elasticsearch support baked in.  https://www.elastic.co

We're already using LWJGL in the oculus display class, that is based on JOGL.  Our support for LWJGL comes from the "Thin Matrix" tutorials on working with java open gl stuffs.   https://www.youtube.com/playlist?list=PLRIWtICgwaX0u7Rf9zkZhLoLuZVfUksDP

They are really excellent 

What would be good with the SLAM is to be able to take data from OpenCV and YOLO so that it remembers where objects are and can find its way back to them