Okay, the Competition is now over, it’s time for a couple of final updates:
-
we’ve sent out tiny (digital) gifts to the Teams on the podium (best 3 Teams) - please reach out if you have not received anything
-
for the Finalists (best 5 Teams) we can also offer a 1 year, 100 device Husarnet plan if you start it this year. Please reach out if you’re interested
-
we’ll also be preparing participation certificates for you this month - if the member lists have changed - please send the updated complete (not just a difference) lists to erc@husarion.com and teams@roverchallenge.eu - ideally in a week or so. Please remember to provide same GDPR forms that were used during the Registration for each of the new Team members
-
just a notice - virtual machines for the competition have been destroyed already
Next, I’d like to give a short summary of the whole Competition from our point of view. It won’t be as detailed as the ones we’ve got from you, but I hope it’ll still provide some insight into how things looked from our side.
We’ve started strong - with hardware setups for both the Challenge 1 and 2, so we can prepare Technical Handbook while already having some things ready and known to work. But the software… software turned out to be the difficult part. It had made us not only ditch the Zed cameras but also forced us to prepare “something functionally similar” at the last minute. This took so many man-hours that other aspects of the Competition started to take a toll and we decided that, if there are so many unexpected hiccups on our side, there will be some on your side too, thus we need to limit the scope of the whole Competition. Challenge 2 have been the one we internally longed for, and we’ve already made many tests of the setup - but the sheer number of “moving parts” in those “mini-games” made us project the potential amount of issues on your side and we decided not to do that part this year. Challenge 1 however, seemed like a “known problem” in the robotics area therefore you’d have more resources available than just our notes. Partially it turned out to be true - you’ve built your pipelines upon existing solutions, and the main issue we’ve noticed at the end was… simple lack of time with the robot to test things - especially the remote connection part. While we won’t have any “easy” solution for that, we will be keeping it high in our notes.
Scoring was also one of the difficult parts, because of two main reasons. First - as this was the first competition of this kind organized by us - we were intentionally vague about the requirements and wanted to see what you will come up with. While we were positively surprised by the quality of your work - it also made some of the deliverables almost incomparable. This is also why, for the most part, we’ve tried to see all of the submissions before we started scoring them. For the reports it was relatively easy - read all of the documents, note the common parts, note the outliers, balance the points for each of those categories and then do the actual scoring. For the Test Drives it wasn’t so easy as there was no “back to the beginning” button in real life. This is why you may have felt that some of the scoring felt “artificial” - we had to come up with something before we’ve seen your work so we had at least some starting points for scoring each of you. Second difficult thing in that area was balancing the points for different categories. This is something that have come up to us at the very last stages of the Competition - and also something that made us glad that we’ve left us some leeway regarding the previous point. After we’ve seen what you’ve done outside the main events in all possible aspects - especially technically, while preparing for the Challenge 2 and socially, for the community outreach - we now know what you’re capable of and have some ideas how the Rulebook should have been changed to make rewarding outstanding results easier. We haven’t made those changes this year because we didn’t want to introduce more changes to the Competition, and also we’ve seen that plenty of you had adjusted the strategy based on those points so it wouldn’t be fair.
While talking about the organization and scoring we have to notice one important thing. From the start we knew (and communicated) that this Competition will not be perfect. This is way we aimed for fairness instead. This is why we’ve made many things in certain way, but also why we have not have done many things. I.e. if one of the Teams figured out what was the problem with the setup but communicated the solution only to us and not the Community Forum - we did the exact same steps debugging problems with all the Teams that stumbled upon them later in the day. Has it made the process more frustrating for some of you (and for us)? Most likely, but it seemed to be the only fair solution in that case.
When it comes to the things that we’re happy about I’d like to talk about two main things. First is the Community Forum - I was really glad to see your professional and helpful posts/inquiries. Even while you’ve described bugs and issues - you’ve tried to be helpful and provide data about the exact situation, which made it much easier to analyse. More importantly though I was able to see your work in between Test Drives and see that you’ve managed to collaborate and build upon materials from other Teams. To the outsider this may seem like an odd move - showing your cards before the game - but in reality you’re opening yourself for feedback that can make you stronger. On the other side being able to see what other Teams have made allows you to improve your own solutions - not by copying their work - but by observing the problems they’ve found and overcome - but this mechanism only works if you’re Community is mature enough to first open their work to others - and I’m glad you were. Second thing I’m happy about is that the “technology bingo” we’ve had in mind designing the Challenges has worked out decently. Our goal was to take the challenges typical for rover competitions from all around the world but then not only to make them solvable in the Remote Formula but also to add some industrial/corporate trends - for example this is why the final report was meant to be more of a “retro” from agile methodology, why we’ve encouraged you to use large language models and so on. While it may seem that some of these things are trends that may not last very long - they’re definitely something that current employers are looking for - and already having documented experience with them gives you a huge advantage - especially while landing your first jobs. We’ve seen your reports, we’ve seen your code on Github and we’re certain that you’ll all succeed in your professional careers. We’ve also seen that some of you were able to utilize this Competition to advance your university progress i.e. by writing thesis about parts of the Competition - and we’re really happy about it.
Finally I’d like to thank each and every of you that spent the summer with us - despite the enormous amount of work on both sides - it was really good seeing you progress and achieve more and more. Have a fantastic year and I hope to hear from you again,
Paweł on behalf of Husarion Team