Rendered at 05:56:58 GMT+0000 (Coordinated Universal Time) with Cloudflare Workers.
daww 2 days ago [-]
Author here.
I've spent the last six months replicating the paper "Champion-level drone racing using deep reinforcement learning" and now I'm writing down the blog posts I wish I had along the way.
Any feedback is welcome, especially as I'm a bit unsure if I struck the right balance between being concise and not requiring too many prerequisites.
Also if you're working on RL and robotics (especially aerial), let's connect!
avidiax 7 hours ago [-]
I assume you are going to start introducing all the 2nd and 3rd order effects? One big one is ground effect, and another is vortex ring state/settling with power and the related translational lift, and the props themselves have p-factor and the dirty air effect for the rear props.
quibono 5 hours ago [-]
Are there any resources you know of on modelling ground effect? I’m curious how this would change the dynamics from the post.
dvh 8 hours ago [-]
Isn't it just bi-copter?
palata 7 hours ago [-]
No, it's a quadcopter setup, but simulated in a 2D world (I guess for simplicity). A bi-copter would require tiltrotors, which is different.
echoangle 6 hours ago [-]
In 2D, a bi-rotor is equivalent to the quadcopter in the post. There are 2 thrusts you control to guide the thing.
kabir_daki 8 hours ago [-]
Physics simulations from scratch are great learning projects.
Did you implement your own PID controller for stabilization?
That's usually where things get interesting — tuning the
gains without it oscillating to death.
I've spent the last six months replicating the paper "Champion-level drone racing using deep reinforcement learning" and now I'm writing down the blog posts I wish I had along the way.
Any feedback is welcome, especially as I'm a bit unsure if I struck the right balance between being concise and not requiring too many prerequisites.
Also if you're working on RL and robotics (especially aerial), let's connect!