Video: maker uses 7 Arduinos to turn his Ford Focus into a self-driving car

The real challenge inherent in getting cars to drive themselves is not the act of driving, but in chipping away at that final one percent margin of error that could result in death. Automating the driving aspect in and of itself doesn’t require millions of dollars of R&D, but about $1000, as was the case with Keran McKenzie, a tinker from Melbourne Australia who claims to haved automated his Ford Focus using a bunch of Arduinos.

Musing over the pointlessness of the “home button” on the Ford Focus’ dashboard, a button whose sole purpose is to plot the trip back to your home, McKenzie wondered: “Why can’t I push that button and have it take me home. I don’t want to drive; I just want to go there.” And so a project was born.

McKenzie, who once led a team to victory in a 2013 self-driving car hackathon, set to work by replacing the five short-range recessed ultrasound sensors around the perimeter of his car with five 150-degree camera lenses connected to five Arduino Unos.

The next step was to neatly link them all together with a master processor—also an Arduino—situated in a compartment behind the brake fluid control systems, in the back of the motor. Then he piped some feedback into a miniature OLED chip-on-chip display placed within the cabin to nab feed off the control system and make sure everything’s properly working.

McKenzie doesn’t go into any detail over the programming challenges involved in training the vehicle to identify and discern pedestrians from street signs and cars, but as we see in the video below, his budgeted approach at a self-driving car was not half-bad. The vehicle certainly drove itself, although its safety on public roads is questioned within only 10 seconds of driving.

So, is this clever rouse poking fun at makers and the race to automate, or an authentic self-driving vehicle? Let us know what you think in the comments below.

Looking into the future, it’s no secret that the autonomous driving era will wipe out many of the traditional driving jobs once it gains further momentum; on the flipside, analyst firms like Boston Global Consulting are predicting an outpouring of 100,000 new jobs by 2025 as the market value settles at $42 billion. Projects like McKenzie’s are an excellent way to gain some hands-on experience that could help prepare you for a new career in this growing field.

Those interested in joining the cutting-edge of machine learning may find that a traditional college education may not always be able to keep pace with the ever-changing nature of the technology. With massive tech companies like Google and Baidu racing to upstart the autonomous vehicle market, highly-focused nano degrees could provide a more focused learning experience in less time. One such example includes a 27-week course from the online-learning platform Udacity promises to turn anyone with $2400 and a basic coding experience into a deep learning engineer. The curriculum requires prospective students to possess a basic understand of statistics and the Python programming language.

Source: IEEE Spectrum and Wired