28 Jan

This improvement is based on the previous posts, corresponding to the first step: Step 1: Deal with information problem: using only closest neighborhood information. Code is here The main improvent is to set the observation field. Current mean-field assumes all agents' information are av...

28 Jan

Hi everyone! I'm a new postdoc with LEMUR, and I'll be working on modern tools for circuit board level electronics design. So that's a lot of words, let's unpack that. Modern board design tools (like KiCad) consist of schematic capture (drawing the schematic) followed by layout (placing components...

24 Jan

This week, we started to get all the new members know how to build a blimp from sketch. Next week, we will be working on the altitude control using ultrasonic sensor. Meanwhile, some ideas that we are thinking about: (1) Modify the propulsion system, use only 1 motors at the side of blimp...

20 Jan

Paper Target: IROS 2022 Current open source mean-field reinforcement learning code, which is based on MAgent Blue: Group 1 (Proposed Algorithm)Red: Group 2 (Opponent) Algorihtm Information Collision Task Current using all agents as neighborhood One state (position) only...

11 Jan

Problems that I've been thinking about: How can collaborative code development be more efficient? How do avoid conflict when multiple people are developing? How make the code more readable, easy to maintain and debug? A successful Git branching model: (Simplified version) A example...

07 Jan

A quantum Kalman filter is a recursive Bayesian filter used to predict and estimate the state of a quantum system, and was first introduced by Viacheslav Belavkin. It is described in his work (Belavkin, 1999). Another interesting work is the quantum extended Kalman filter. However, much of the r...

07 Dec

Tasks to finish before 16/12/2021 These tasks mostly to get to speed on the roco pipeline and begin work on adding new modules Build and fix existing cars to produce .wbo output in the Device branch. Making the API compatible to work with it. Build some new WMR models Figure out why one pytest...

30 Nov

Last week, I explained why typical distributed gradient descent leads to inexact solution, even though the convergence rate is linear (in log scale). There is a very important paper that shows that we can in fact have exact solution by incorporating the previous gradient. This algorithm is called...

09 Nov

proposal writing Focus Implementing a generalized control and planning output for arbitrary folded robots designs generated by a computational design framework Approach I am looking at papers that are focusing on generating the kinematic and dynamic model of robots with modular components (...