TE
TechEcho
Home24h TopNewestBestAskShowJobs
GitHubTwitter
Home

TechEcho

A tech news platform built with Next.js, providing global tech news and discussions.

GitHubTwitter

Home

HomeNewestBestAskShowJobs

Resources

HackerNews APIOriginal HackerNewsNext.js

© 2025 TechEcho. All rights reserved.

Show HN: Autopilot written in Keras for Self Driving Cars

144 pointsby autonomydriverover 5 years ago

10 comments

pathsjsover 5 years ago
&gt; At the current state of my model the model basically just clones the human driver as good as possible. That means the the amount of brake is higher in curves<p>I read in another comment that you are still in high school, so maybe the above is because you do not have actual driving experience. But this is not how human drivers drive.<p>Human drivers brake <i>before</i> the curves, while usually accelerate during curves. This improves stability.<p>This is something you may want to consider for your next iterations. In any case, congratulations for your impressive work!
评论 #21564877 未加载
评论 #21564898 未加载
评论 #21565341 未加载
评论 #21564498 未加载
czrover 5 years ago
good start op. a good next step is add in temporal context (previous frame information) so model may resolve ambiguous cases. see e.g. breakdown of comma.ai openpilot model here (from their twitter, [0]), or the karpathy talk as well.<p>also, when presenting results, prefer to include a longer demo or sub-demos that show strengths and failure cases, and move to the top of the post rather than bottom imo. for a given reader, implementation details are either confusing and uninteresting (if not subject-matter expert) or predictable and uninteresting (if an expert, you&#x27;ve seen many similar before); audience for implementation details is very small number of people who want to sit down and replicate or check your work. but demos &#x2F; analysis is always novel and interesting to anyone, so lead with that :)<p>[0] <a href="https:&#x2F;&#x2F;medium.com&#x2F;@chengyao.shen&#x2F;decoding-comma-ai-openpilot-the-driving-model-a1ad3b4a3612" rel="nofollow">https:&#x2F;&#x2F;medium.com&#x2F;@chengyao.shen&#x2F;decoding-comma-ai-openpilo...</a>
评论 #21561949 未加载
评论 #21563051 未加载
devitover 5 years ago
Seems really scary to exclusively use a neural network for safety-critical tasks like this, without having an explicit method guaranteeing safety.<p>You can&#x27;t prove that a trained neural network is always correct, and thus this is likely going to kill someone at some point.<p>I think you definitely need a LIDAR (or in general something that can give an accurate 3D map of all surroundings) and some explicitly written code that can be proven to result in never hitting a car going the opposite direction in the opposite lane (provided they stay in their lane), as well as never hitting stationary obstacles and never going off the side of a mountainside road.
评论 #21563010 未加载
评论 #21563040 未加载
nikodunkover 5 years ago
Great job! This is amazing. Love how I’m you used off-the-shelf components. Congrats and keep going!!
评论 #21562229 未加载
syntaxingover 5 years ago
You do something similar in Udacity&#x27;s SDC nanodegree but you use their simulator rather than a new car. Interestingly through trial and error on my project, ELU activation functions was the only thing that prevent vanishing gradient problems. You use the same activation function and I&#x27;m curious why you selected that one. I always wonder why it was the only function that worked for me.
评论 #21564542 未加载
评论 #21562228 未加载
rpmismsover 5 years ago
Holy crap.<p>First: you&#x27;re awesome. Keep it up. Second: anyone else think this is a great example of how autonomous driving is a last-mile problem?
评论 #21565616 未加载
billsmithaustinover 5 years ago
How big of a risk is overfitting for these kinds of models? I think common practice is to hold back some of the data for testing.
评论 #21563900 未加载
verumnover 5 years ago
Just out of curiosity how old are you? Keep up the good work!
评论 #21565279 未加载
daenzover 5 years ago
&gt;A few days ago @karpathy presented their workflow with PyTorch and also said some numbers, to train the Autopilot system with all it neural networks you would have to spend 70,000 hours with a decent gpu - that is around 8 years (depending on which GPU you are using). In total the Autopilot is a system of 48 Neural Networks. When we compare this to what I will show you, you are gonna see that this is insane.<p>I&#x27;m very confident that it is not insane, for reasons that you have yet to discover, and the arrogance of calling insane inspires a fear in me considering what you appear to be doing on roads with other people.<p>Your project is very cool, but also very irresponsible. Are you using this on the road with other drivers? For the love of all things good, what are you thinking? Please clarify this. It&#x27;s one thing to trust your own life to your creations, it&#x27;s another thing to endanger everyone else&#x27;s.
评论 #21561670 未加载
评论 #21561713 未加载
评论 #21562178 未加载
评论 #21561608 未加载
评论 #21561515 未加载
jacquesmover 5 years ago
Neat project, at a minimum this should get you hired somewhere fancy. I&#x27;d love for all the self-driving car vendors to take a leaf out of your book and keep their software in-house until it agrees with what the real drivers do &gt; 99.99% of the time and the other 0.01% led to an accident of sorts.
评论 #21563043 未加载