Home

Welcome to the autonomous wiki!

• collected data is smaller than models data set that could cause some training issues
• The windows is cropped in the middle and doesn't see obstacles at the right moment
• Position camera less pointing to ground

Explanation of the car model: Dan Does Data: Self Driving Car Model

Papers for review:

• The DAVE network, http://images.nvidia.com/content/tegra/automotive/images/2016/solutions/pdf/end-to-end-dl-using-px.pdf

Data Format:

Python Development:

Python Best Practices (Digested)

PEP8

PEP20

Data Examples:

Otto Tractor Bay Area Makerfaire 2017

Suggested Development Environment(s)

Developer Setup Suggestions (Atom)

Date Read Example Code:

import numpy as np
import matplotlib.pyplot as plt
a = np.load("imgs_2017-05-15_22:13:47.npz")
a.keys()
b = a['arr_0']
b.shape
plt.ion()
plt.imshow(b[0].transpose(1,2,0))
b[0]

Installation Documentation

[TX2 Installation] (https://github.com/fubarlabs/autonomous/blob/master/installation.md) [Raspberry PI Installation TBD]

Enable Raspberry PI USB Camera

Add Video for Linux 2 (v4l2) to the PI:

sudo apt-get install v4l-utils

Add USB video camera kernel module

sudo modprobe bcm2835-v4l2