We’re gearing up for an exciting year in the AlphaPilot Challenge! As we close out 2018 and prepare for the year ahead, we would like to share more details on what to expect in the coming months.

Please see below for a list of new/updated frequently asked questions as well as check out the FAQ section for all answers. Additional details will be made available in mid-January. If you have a question that is not answered yet, please post in the forum and we’ll do our best to address it in the next release.

 

Do I need my own drone/equipment?

Answer: Following the qualifier challenge, the final teams selected will be provided with kits including software and hardware development assets. They do not need to supply any of their own equipment to write or test competition code. Onsite, all drones, technology, and associated components will be provided by the AIRR League
 

What are the specifications of the drone?

Answer: Though we’re not releasing specifications on the competition drone model yet, we can confirm it will utilize an NVIDIA Xavier chipset for onboard computing and multiple sets of stereoscopic cameras. Each set of cameras will operate at 720p @ 60fps. Look out for more details in the coming months.

 

How quickly will the drones go?

Answer: DRL’s stock racing drones reach speeds in excess of 80mph, but the AlphaPilot Challenge and AIRR League will require development of a significantly different drone model and technology-base. Look out for more details on its specifications and capabilities in the coming months.

 

What can I do to prepare for the challenge?

Answer: In preparation for the virtual qualifier challenge and AIRR League, we urge participants to practice solving challenges related to computer-vision, creation of neural networks, and image classification. Developing an in-depth understanding of the Xavier Jetson is also encouraged, as is utilization of free autonomous vehicle related simulators - all of which will be valuable experience in preparation for the AlphaPilot Challenge and AIRR League.
 

How do I find teammates or form a team:

Answer: Teams can be formed through the HeroX forum, through in-person contacts with other challenge participants, or any way you see fit to approach the challenge. Regardless of which team you’re on, we encourage collaboration between participants in the forum.

 

Where can I find more information about citizenship eligibility?

Answer: For more information on participation regulations in the challenge, reference the AlphaPilot Terms and Conditions. All team members and any entities they represent must comply with contest regulations as they relate to these restrictions. Applicants holding dual citizenship in a restricted country will not be allowed to participate.

 

Will I have access to a simulator?

Answer: The final teams selected for the AlphaPilot challenge will have access to a development kit - one component of which will be a simulator for code deployment and testing. Unfortunately, contest administrators will only be able to provide this AIRR League development kit for the final qualifying teams.

 

What will course routing and course elements consist of?

Answer:  Racing will take place on AIRR League courses - designed and produced by the Drone Racing League. Competition elements such as gates, length, and flight path complexity are being developed by the same course designers that shape DRL’s cutting edge courses. However, they will incorporate a visual fiducial system and be designed specifically with the intention of hosting autonomous drone competition.
 

What does the virtual qualifier consist of?

Answer: The virtual qualifier will consist of three parts: Test 1 will be a 3-5 minute video submission accompanied by several written responses. It will be used to assess each team’s knowledge level, relevant experience, and cohesiveness. Tests 2 and 3 will be made available simultaneously shortly after, and will be designed as objective tests of applicant skill. Test 2 will consist of a computer vision exam, where teams must develop and submit image classifier algorithms for a specific race-related object (using a provided image library). Test 3 will involve an open-source Unity-based simulator, where teams must develop and submit algorithms for guidance and control of drones. The simulator provided in Test 3 will be separate from DRL’s simulation environment, which will only be available to finalist teams as part of a development kit. However, the Test 3 simulator will remain available for public use as tool for training and testing autonomous control algorithms following the 2019 AlphaPilot qualifier. Details on the qualifier challenges, timelines, and scoring criteria will be made available in mid-January.