This week's main task is to understand the current paper as well as the NetMatch algorithm. Since the basic code with random labels and nearest neighbors is working, it is time to start implementing QuickMatch distributedly on different nodes.
I spent a lot of time talking with Zack and reading about the current paper in publication. Here is a summary of Distributed QuickMatch (NetMatch) algorithm:
The next step would be to start implementing k-means partitioning and make sure everything still works the same.
0 Comments
Part 1: iCreate DocumentationThe documentation of how to build a customized iCreate roomba can be found here. The documents include:
Part 2: Code Modification The code has been modified according to the requirements from the last meeting.
The current node:
The goal of this week is to modify the previous node to use the extract features from pre-existing image set, publish and subscribe the features through a rostopic.
The current node:
Feedback:
The goal this week is to modify the existing code for distributed random numbers to distributed random vectors. There will be a pre-defined 2 x 3 matrix as a parameter, within which each column represent a 2d vector/point in the coordinate plane (there will be a total of 3 column vectors in the matrix). Each node should have a column vector, which is its coordinate (x, y). At least 30 points/vectors will be generated for each node (90 points/vectors total)
Part 1: Distributed Random Numbers in ROSThe main assignment this week is to create a centralized version as well as a distributed version of random number generators. For both versions, the same amount of numbers from 1 to 10 will be generated randomly, and their frequencies will be recorded. The differences between the two versions are listed below.
Part 2: Experimenting with iRobot Create roombaWith the help of Zack and the Robotics Lab wiki pages, I was able to connect to the iRobot Create roombas and try out the following two things.
Q & A:
6 hours worked
Questions:
|
Archives
May 2020
Categories |