308 karma · joined June 6, 2011
I put together a resource guide here: https://becomeautonomous.com
Hope that helps!
-------
If you want to get more serious about robotics then check out my website https://becomeautonomous.com
Here's a few thoughts:
* This isn't really open-source. The most important part, the neural network perception pipeline, is closed sourced. Without the training data and the network architectures, the supporting code is not that useful.
* This system is non-deterministic, which is a no-no in critical systems like autonomous vehicle. There's no guarantee that vehicle detection or throttle/brake commands will be executed on a acceptably safe timeline.
* I think the supporting tools like Panda and Giraffe are great for those wanting to get into car hacking.
------------------------------------------------
By the way, if you want to to become an autonomous vehicle engineer I put together a resource guide here: https://becomeautonomous.com
Here's a few thoughts:
* They say it has a 95 degrees vertical FOV and 360 degrees FOV. The vertical FOV is awesome actually. A Velodyne VLP-16 has roughly 30 degrees FOV. We need the vertical and horizontal angular resolution to make any first-take judgements though.
* One overlooked feature I want to point out is that the Minimum Range is 0. Many lidars perform well at far distances but struggle to see objects very close (<1 meter away). This poses a safety issue as someone or something is effectively invisible to the robot if they're too close. Of course you can use cameras to handle this case but this is a big win for the Waymo hardware team!
* They're going to be 15k+ (gut estimate)
------------------------------------------------
By the way, if you want to to become an autonomous vehicle engineer I put together a resource guide here: https://becomeautonomous.com
A Velodyne VLP-16 has roughly 30 degrees FOV. We need the vertical and horizontal angular resolution to make any real judgements though.
My sister was also battling cancer so this made it even more difficult from a time and emotional standpoint.
They also provide you with a lot of different perspectives from a practical system design level.
Good luck :-)
I’d recommend reading the papers from the DARPA 2007 challenge. https://www.springer.com/us/book/9783642039904
Pro-tip: You can google each chapter title and find PDFs for free.
Regarding the education comment...I’m going to disagree with you there. Employers don’t care where you went to school. They care about what you’ve done in the past (doesn’t matter if it’s at work or on the side) and what you can bring to the table. I had to compete with people with PhDs for my current AV job but I got the job because of the whole package of skills that I bring rather than theoretical knowledge in a single niche topic.
I took the 9-month Self-Driving Car Nanodegree and managed to find a job in the AV industry right before graduating.
I found the ND worth every penny because they broke down really dense topics that you’d usually need to be in a MS program at CMU or Stanford to undertake. Now I’m leading AV development at my current employer!
Like any course or degree, they can’t cover everything. That’s ok though, because you’ve learned enough to continue learning by reading current academic papers. You must be motivated though.
There’s a lot of nay-sayers on HN. I’d encourage you to ignore their excuses for not completing their ND. Usually these people don’t have enough grit or motivation to do something hard. If you really want it...you’ll work hard and persevere to get it.
1. The Ask A Manager Podcast and the audiobook on Audible.
2. Take a co-leadership course or read their book http://coleadership.com
Gotta say though...it's quite amazing to see 5-10 SDCs/day.
I was mostly speaking about math professors in college. I have degrees in Math and EE...advance math isn't an issue.
Having abstract intuition (aka "high level ideas") in math is incredibly value.
Sometimes things get hairy in the computation and you need to step back and consider the bigger picture again.
If you don't understand how the math works then you don't have the intuition and vice-vera.
I have video chat, a synced shared web browser, and support for Netflix & Hulu.
If anyone is interested: https://www.distanceflix.com
The setting that sounded best to my ear was:
Character variation 0.7
String damping 0.7
String damping variation 0.15
Pluck damping 0.5
Pluck damping variation 0.40
String tension 0.0
Stereo spread 0.1
String damping calculation Magic
Body resonation None
Simple Precache time 11.3 seconds