I remember that awesome day of March 21st when I got the amazing email notification that contained the honorary invitation to the PIX Moving Challenge in Guiyang, Guizhou, China. I knew immediately that this was a once in a life-time opportunity, so I decided to accept the challenge and fly to China.
Getting to the PIX factory was overwhelming. PIX is an Autonomous mobility company that is reinventing the way we move; the way things move and the way we live. All the technology in the factory was just impressive. Talking to the different people and hearing from them gave me a wonderful perspective on things, and you could also feel their passion of building a cool product that will change people’s lives.
Check this out!
Personally, I think that the PIX team have a great challenge in front of them but all the development, the research and the enthusiasm of the team will make the project a great success in the near future.
On May 22nd we got to the PIX moving factory to start the challenge. Engineers from China, Turkey, Brazil, India, Spain and Mexico were attending the event. We were split into 3 teams. PIX provided a car with a drive-by-wire, a lidar, a camera, a GPU Server and the power station installed for each team and the challenge was to install, configure and develop all the needed components to run the car autonomously and accomplish all the 7 missions within 5 days.

Here’s a glimpse of the wonderful experience that panned out over the next 5 scintillating days: “Experiential learning” was the common denominator of the first 2 days. We started facing a lot of problems while installing and configuring the software. Simultaneously we were trying different approaches to do the object localization from the images collected by the camera. By the end of the second day, the car was collecting data through the lidar and the camera, and we were ready to map and integrate with the object localization.
The third day arrived, and to sum-up the third day in one word I would choose “Frustration”. We arrived at the factory on the third day with our expectations high. The plan was to integrate and have something working by the end of that day. But, soon we figured out that nothing was working as expected, the environment in our computers was different to the car computer. We did so many changes to the code, but our accuracy dropped from 94% to 0.001% in the car computer. The way points were not accurate, and the car was not behaving as expected. We decided to clear/reset our minds and go back to the hotel, hoping to get a fresh start for the next day.
For the final 2 days of the challenge, the word I constantly listened from the team was "Persistence". We changed our approach for Object Localization and the configuration in the car started to work as expected. We were able to run our first tests. As if the already existing problems weren’t enough, rain decided to show up and added an extra complexity factor to the challenge. We worked extremely hard on the car, running so many tests, pair coding to help each other. At nights the wind and the rain did their usual job and we were up through midnight on the street trying to calibrate the lidar to detect objects (we did not know we can do this on the computer).
Sometimes I felt we were pushing the car limits so hard, and that could have been the reason for the many mistakes we did. However, the strategy was clear for PHD. Ma (Sam), we will win this challenge because we will be FASTER, and Persistence was the key, so we continued trying, testing and improving for rest of the event. Those were long days, I think I slept 6 hours during the final two nights, which I feel was too much compared to my other teammates. :)
Somebody asked me, what did you win at the challenge? I can say I won a unique experience, having the opportunity to run some code on those incredible cars. Working with the Shanghai Jiao Tong University team (Sam, Leon, Ting), spending all that time with them, and after that short but intensive period hearing "Bye my friend, see you next time" was just priceless.
Connecting with smart people like Meenakshi, Italo, Mustafa, Phd. Ali, Phd. Kerem and learning from the experts Phd. Alexander, Sam and Eduardo are simply fantastic, and this does not compare to any other materialistic prizes.
I cannot stop expressing my gratitude to the PIX factory team, specifically to Angelo Yu [CEO of PIX] for inviting me to the event, and to Nancy Lee [community manager] of Pix Challenge. Thank You!
Thanks to my friend Vicky Hora for his support.