AWS DeepRacer League 2021 Update #5 End of Month Special

It’s all done! No more submissions for the March race and after some final pushes I can share who we will follow in the Pro Division in April and who will fight for the finals in March.

AWS DeepRacer

AWS DeepRacer is a 1/18th scale autonomous race car but also much more. It is a complete program that has helped thousands of employees in numerous organizations begin their educational journey into machine learning through fun and rivalry.

Visit AWS DeepRacer page to learn more about how it can help you and your organization begin and progress the journey towards machine learning.

Join the AWS Machine Learning Community to talk to people who have used DeepRacer in their learning experience.

DeepRacer News

AWS DeepRacer Women’s League India 2021

If you haven’t done so yet, there’s still time to register for the AWS DeepRacer Women’s League India 2021 which invites students in India to join the race for the ML knowledge, bragging rights and some pretty sweet prizes. More info on our blog.

This image has an empty alt attribute; its file name is 20210321_AWS_DR_W_L_India.png

AWS DeepRacer League March Finale

The final race will take place on the 8th of April at 5:30 pm PST (9th of April 1:30 am BST, 9th of April 10:30 pm AEDT). It will be presented live on the main AWS Twitch channel: https://www.twitch.tv/aws.

April races

Both Open and Pro Divisions have started their races. This month we’re facing the Lars Loop and Lars Circuit, inspired by the Thunderhill Raceway Park in California. Lars is racing as Duckworth in the league and is a very active community member, maintaining the local training stack and contributing to many other community projects.

Lars Circuit (Source: AWS DeepRacer Console)
Lars Loop (Source: AWS DeepRacer Console)

Head over to the AWS DeepRacer Console to start your autonomous racing adventure.

Pro Division

Po-Chun Super Speedway track (Source: AWS DeepRacer Console)

Race info:

  • Track named after the winner of 2020 AWS DeepRacer League
  • 89.24 meters long
  • 107 centimetres wide
  • Multiple difficult turns with the most tricky one exiting a chicane past the hairpin in the upper-right corner, branded Yellow Brick Road since it looks like the beginning of Yellow Brick Read from The Wizard of Oz
  • Head-to-bot format with four bot cars cruising at 2 m/s without lane changes
  • Racers are trying to complete three laps
  • Each collision and each driving off the track results in a five seconds penalty
  • Top 16 go on into a live race which will be streamed on Twitch on the 8th of April to fight for places in the League Finals and AWS DeepRacer EVOs

Let’s see the stats for the whole group:

All submissions
Update time: 2021-04-01 07:08:37.188048 UTC
Comparing to 2021-03-28 18:00:00 UTC

Racers: 70 (+8)
Total submission attempts: 71505 (+10446)
Best total time: 01:26.927 by JJ
Mean total time: 03:32.915828571 (+00:03.833360829)
Best lap: 00:28.676 by JJ
Mean best lap: 01:06.455142857 (+00:01.180062212)
Mean average lap: 01:10.971614286 (+00:01.277807834)

Eight racers joined in a last minute attempt which brings the total of participants to 70, none of them however managed to jump into top 20.

Top 20 submissions
Update time: 2021-04-01 07:08:37.188048 UTC
Comparing to 2021-03-28 18:00:00 UTC

Racers: 20
Total submission attempts: 46315 (+7210)
Best total time: 01:26.927 by JJ
Mean total time: 01:37.866450 (-00:01.377750)
Best lap: 00:28.676 by JJ
Mean best lap: 00:31.838500 (-00:00.373500)
Mean average lap: 00:32.621850 (-00:00.459150)

We’ve had some active submission attempts in the top twenty, most racers however haven’t improved their times. Let’s see what this means in terms of final standings:

Rank Racer Total Time Avg Lap Best Lap Resets Submissions
1 (↔0) JJ 01:26.927 (-00:00.000) 00:28.975 (-00:00.000) 00:28.676 (-00:00.000) 0 (-0) 4403 (+379)
2 (↔0) flatearth 01:29.800 (-00:00.000) 00:29.933 (-00:00.000) 00:29.325 (-00:00.000) 0 (-0) 1668 (+343)
3 (↔0) RogerRabbit 01:30.203 (-00:00.000) 00:30.067 (-00:00.000) 00:29.277 (-00:00.000) 0 (-0) 3606 (+397)
4 (↔0) DBro 01:31.268 (-00:00.000) 00:30.422 (-00:00.000) 00:29.935 (-00:00.000) 0 (-0) 3581 (+583)
5 (↔0) Fumiaki 01:35.555 (-00:00.000) 00:31.851 (-00:00.000) 00:31.202 (-00:00.000) 0 (-0) 2042 (+445)
6 (⬆8) PolishThunder 01:35.649 (-00:06.061) 00:31.883 (-00:02.020) 00:31.607 (+00:00.009) 0 (-1) 5210 (+557)
7 (⬇1) Duckworth 01:35.791 (-00:00.000) 00:31.930 (-00:00.000) 00:31.727 (-00:00.000) 0 (-0) 1260 (-0)
8 (⬇1) Karl-NAB 01:36.495 (-00:00.000) 00:32.165 (-00:00.000) 00:31.596 (-00:00.000) 0 (-0) 279 (-0)
9 (⬇1) Jochem 01:36.783 (-00:00.000) 00:32.261 (-00:00.000) 00:31.500 (-00:00.000) 0 (-0) 2528 (+555)
10 (⬆7) GT-DevelopersIO 01:37.110 (-00:09.098) 00:32.370 (-00:03.032) 00:31.756 (-00:01.746) 0 (-1) 513 (+449)
11 (⬇2) Ernesto 01:37.185 (-00:00.000) 00:32.395 (-00:00.000) 00:31.607 (-00:00.000) 0 (-0) 3574 (+281)
12 (⬇2) dartjason 01:38.257 (-00:00.211) 00:32.752 (-00:00.070) 00:31.831 (+00:00.228) 0 (-0) 3547 (+473)
13 (⬇2) RobinCastro-DBS 01:38.503 (-00:00.000) 00:32.834 (-00:00.000) 00:32.302 (-00:00.000) 0 (-0) 3218 (+633)
14 (⬇1) JPMC-DriftKing 01:38.829 (-00:02.018) 00:32.943 (-00:00.672) 00:32.246 (-00:00.417) 0 (-0) 401 (+148)
15 (⬇3) Maikel 01:39.672 (-00:00.000) 00:33.224 (-00:00.000) 00:32.740 (-00:00.000) 0 (-0) 1314 (+362)
16 (↔0) Joffe 01:41.768 (-00:03.589) 00:33.922 (-00:01.197) 00:32.559 (+00:00.083) 0 (-1) 470 (+348)
17 (⬇2) AJM-Model-1 01:44.016 (-00:00.000) 00:34.672 (-00:00.000) 00:33.823 (-00:00.000) 0 (-0) 1889 (+447)
18 (↔0) Yxcas-DBS 01:45.038 (-00:01.749) 00:35.012 (-00:00.583) 00:34.254 (-00:00.684) 0 (-0) 2808 (+556)
19 (⬆2) ep-dr-racer1 01:48.801 (-00:07.316) 00:36.267 (-00:02.438) 00:32.579 (+00:00.439) 2 (-1) 2338 (+398)
20 (⬇1) rosscomp1 01:49.679 (-00:00.000) 00:36.559 (-00:00.000) 00:36.228 (-00:00.000) 0 (-0) 1666 (+337)

No changes in top five and there are some shifts just below. PolishThunder has made some significant improvements that have moved him from 14th to 6th – those lower places are not very comfortable. Another significant jump cane from GT-DevelopersIO. Impressing nine-second speed-up puts the racer in 10th and marks the last successful dash into the top16. Unfortunately this means that we will not see AJM-Model-1 who tried really hard and shared a lot of his effort in the AWS DeepRacer Community. It’s good that there will be more races to fight in.

Let’s end with the time improvements graph:

Open Division

Open races will always be slightly easier than the Pro ones. Let’s see what’s on in March:

Po-Chun Speedway (Source: AWS DeepRacer Console)

Race info:

  • Po-Chun Speedway is a simplified version of the Pro track
  • 68.68 meters long
  • 107 centimetres wide
  • Simpler and shorter than the pro track but still pretty challenging with a variety of straights and turns, some of them pretty tight
  • Time Trial format
  • Competitors need to complete three laps and total time counts for their classification
  • Each off-track results in five seconds penalty.
  • Racers fight for qualification to the Pro group which is won by the top 10%
  • They will also receive a Pro Welcoming Kit which is a bit of a mystery

Let’s take a look at the overall stats:

All submissions
Update time: 2021-04-01 07:08:43.022931 UTC
Comparing to 2021-03-28 18:00:00 UTC

Racers: 972 (+57)
Total submission attempts: 10473 (+814)
Best total time: 01:10.649 (-00:03.378) by ZhengYi-NYCU-CGI
Mean total time: 05:33.590667695 (-00:02.029640501)
Best lap: 00:22.356 by Yoway-NYCU-CGI
Mean best lap: 01:45.122076132 (-00:00.587923868)
Mean average lap: 01:51.196564815 (-00:00.676547753)

57 new racers in the last three days and a total of 97 racers (most likely) qualifying into the Pro Division. We’ve got a significant time improvement in the top 20:

Top 20 submissions
Update time: 2021-04-01 07:08:43.022931 UTC
Comparing to 2021-03-28 18:00:00 UTC

Racers: 20
Total submission attempts: 1770 (+56)
Best total time: 01:10.649 (-00:03.378) by ZhengYi-NYCU-CGI
Mean total time: 01:29.140100 (-00:04.354200)
Best lap: 00:22.356 by Yoway-NYCU-CGI
Mean best lap: 00:28.615250 (-00:01.275500)
Mean average lap: 00:29.713100 (-00:01.451400)

A 4 seconds reduction in mean total time when only four racers had better times than on Sunday – pretty impressive. Overall top 20 racers were pretty inactive towards the end – not much incentive in racing when qualification is already secured.

Now have a look who joined the top of the table at the very last moment:

Rank Racer Total Time Avg Lap Best Lap Resets Submissions
1 (–) ZhengYi-NYCU-CGI 01:10.649 (–:–.—) 00:23.549 (–:–.—) 00:23.314 (–:–.—) 0 (–) 4 (–)
2 (⬇1) Mungi 01:14.027 (-00:00.000) 00:24.675 (-00:00.000) 00:24.249 (-00:00.000) 0 (-0) 13 (-0)
3 (–) Yi-LI-NYCU-CGI 01:15.256 (–:–.—) 00:25.085 (–:–.—) 00:23.405 (–:–.—) 1 (–) 1 (–)
4 (⬇1) AIDeepRacer 01:15.793 (-00:00.282) 00:25.264 (-00:00.094) 00:24.459 (+00:00.321) 0 (-0) 206 (+69)
5 (⬇3) Yoway-NYCU-CGI 01:15.805 (-00:00.000) 00:25.268 (-00:00.000) 00:22.356 (-00:00.000) 1 (-0) 1 (-0)
6 (⬇2) BeaconHillTheFastestVer2 01:17.297 (-00:00.000) 00:25.765 (-00:00.000) 00:25.263 (-00:00.000) 0 (-0) 14 (-0)
7 (⬇2) TimJ-KMITLXRAI 01:18.232 (-00:00.000) 00:26.077 (-00:00.000) 00:25.606 (-00:00.000) 0 (-0) 24 (-0)
8 (⬇2) VaughanDiesel 01:19.407 (-00:00.000) 00:26.469 (-00:00.000) 00:25.720 (-00:00.000) 0 (-0) 451 (-0)
9 (⬇2) s1thind 01:19.551 (-00:00.000) 00:26.517 (-00:00.000) 00:26.198 (-00:00.000) 0 (-0) 110 (+2)
10 (⬇2) Jerec 01:20.433 (-00:00.000) 00:26.811 (-00:00.000) 00:26.461 (-00:00.000) 0 (-0) 231 (-0)
11 (⬇2) GregRacing 01:20.956 (-00:00.000) 00:26.985 (-00:00.000) 00:26.126 (-00:00.000) 0 (-0) 26 (+1)
12 (⬇2) xuqian 01:34.356 (-00:00.000) 00:31.452 (-00:00.000) 00:30.056 (-00:00.000) 0 (-0) 127 (+4)
13 (⬇2) Jackandwho 01:40.376 (-00:00.000) 00:33.458 (-00:00.000) 00:33.205 (-00:00.000) 0 (-0) 6 (-0)
14 (⬇2) satussy777 01:41.325 (-00:00.000) 00:33.775 (-00:00.000) 00:28.867 (-00:00.000) 5 (-0) 121 (+1)
15 (⬇2) muzi 01:42.262 (-00:00.000) 00:34.087 (-00:00.000) 00:32.139 (-00:00.000) 1 (-0) 205 (+12)
16 (⬇2) Redfalcon5 01:46.018 (-00:00.000) 00:35.339 (-00:00.000) 00:34.997 (-00:00.000) 0 (-0) 49 (-0)
17 (⬇2) jmp543 01:47.158 (-00:00.000) 00:35.719 (-00:00.000) 00:35.609 (-00:00.000) 0 (-0) 61 (-0)
18 (⬇2) hypersky 01:47.265 (-00:00.000) 00:35.755 (-00:00.000) 00:35.128 (-00:00.000) 0 (-0) 84 (+5)
19 (⬆243) SorinB 01:47.985 (-02:15.549) 00:35.995 (-00:45.183) 00:33.998 (-00:46.993) 1 (+1) 22 (+11)
20 (⬇3) DanMeon 01:48.651 (-00:00.000) 00:36.217 (-00:00.000) 00:35.149 (-00:00.000) 0 (-0) 14 (-0)

Two more NYCU students jumped straight to the top at the very last moment. Pretty impressive. I’m wondering if it’s three students racing this year and that’s it or if we’ll see more join. I’m pretty sure we will be dedicating some time to their results in the months to come. Have a look at the top 20 racers open division times throughout the month:

Every leader we’ve had since the second day of March has qualified from the looks of it. We could also see some impressive improvements along the way.

And something more interesting: top 97 racers’ times:

Let’s end this post with the list of top 97 racers. We’ll be observing them in the Pro Division starting April:

Rank Racer Total Time Avg Lap Best Lap Resets Submissions
1 ZhengYi-NYCU-CGI 01:10.649 00:23.549 00:23.314 0 4
2 Mungi 01:14.027 00:24.675 00:24.249 0 13
3 Yi-LI-NYCU-CGI 01:15.256 00:25.085 00:23.405 1 1
4 AIDeepRacer 01:15.793 00:25.264 00:24.459 0 206
5 Yoway-NYCU-CGI 01:15.805 00:25.268 00:22.356 1 1
6 BeaconHillTheFastestVer2 01:17.297 00:25.765 00:25.263 0 14
7 TimJ-KMITLXRAI 01:18.232 00:26.077 00:25.606 0 24
8 VaughanDiesel 01:19.407 00:26.469 00:25.720 0 451
9 s1thind 01:19.551 00:26.517 00:26.198 0 110
10 Jerec 01:20.433 00:26.811 00:26.461 0 231
11 GregRacing 01:20.956 00:26.985 00:26.126 0 26
12 xuqian 01:34.356 00:31.452 00:30.056 0 127
13 Jackandwho 01:40.376 00:33.458 00:33.205 0 6
14 satussy777 01:41.325 00:33.775 00:28.867 5 121
15 muzi 01:42.262 00:34.087 00:32.139 1 205
16 Redfalcon5 01:46.018 00:35.339 00:34.997 0 49
17 jmp543 01:47.158 00:35.719 00:35.609 0 61
18 hypersky 01:47.265 00:35.755 00:35.128 0 84
19 SorinB 01:47.985 00:35.995 00:33.998 1 22
20 DanMeon 01:48.651 00:36.217 00:35.149 0 14
21 MindWaris 01:50.346 00:36.782 00:33.997 1 11
22 greyracer 01:53.440 00:37.813 00:34.543 1 33
23 TheMontrealTurtle 01:53.528 00:37.842 00:35.141 1 13
24 LawrenceWong-hk 01:53.724 00:37.908 00:36.864 0 19
25 AM730 01:54.192 00:38.064 00:34.193 2 356
26 Heisenberg 01:59.980 00:39.993 00:36.650 2 84
27 Sl1deways 02:01.086 00:40.362 00:39.326 0 3
28 TestVinodRacer 02:02.146 00:40.715 00:38.675 1 43
29 amanelso4 02:03.675 00:41.225 00:39.458 1 7
30 m-hosokawa-nes 02:04.586 00:41.528 00:37.468 2 7
31 Groovy-DNPds 02:04.983 00:41.661 00:41.331 0 29
32 jgamblin 02:05.170 00:41.723 00:41 0 44
33 drunken-monkey 02:06.378 00:42.126 00:41.226 0 18
34 WaveLabCup-test 02:06.605 00:42.201 00:37.131 5 4
35 Kalky 02:07.901 00:42.633 00:40.704 1 14
36 GoodmanKR 02:07.976 00:42.658 00:39.862 4 255
37 HarinWu 02:08.178 00:42.726 00:39.942 1 51
38 Arcjude 02:08.256 00:42.752 00:37.752 2 4
39 hiroisojp 02:08.276 00:42.758 00:42.465 0 1
40 onepau 02:08.906 00:42.968 00:42.733 0 41
41 js4064 02:09.176 00:43.058 00:40.991 1 84
42 zsracer 02:10.204 00:43.401 00:41.732 4 115
43 Ultrafast 02:10.207 00:43.402 00:35.269 5 80
44 Linnaeus 02:10.460 00:43.486 00:43.067 3 3
45 DH 02:10.513 00:43.504 00:41.463 1 46
46 WAATRAFFICNISHAKRISHNAM 02:10.619 00:43.539 00:41.476 1 3
47 WCStarks 02:12.001 00:44 00:43.333 0 6
48 funny 02:12.451 00:44.150 00:40.136 2 46
49 Euodial3 02:12.963 00:44.321 00:43.734 6 6
50 scottsfirstmodel 02:13.204 00:44.401 00:41.072 2 11
51 ANDREW 02:13.245 00:44.415 00:39.387 5 103
52 KaiRacer 02:13.910 00:44.636 00:38.810 6 14
53 GG12138 02:14.198 00:44.732 00:43.341 1 6
54 SeungJo 02:15.124 00:45.041 00:42.657 1 9
55 CaliSchwiizer 02:17.297 00:45.765 00:44.809 0 4
56 gill-bates 02:17.840 00:45.946 00:43.661 2 17
57 MeganRajagopal 02:18.172 00:46.057 00:43.871 1 6
58 2HIGuys 02:18.576 00:46.192 00:45.879 6 4
59 Dharani1 02:18.977 00:46.325 00:44.335 7 23
60 Wasabi-DevelopersIO 02:19.320 00:46.440 00:42.658 2 694
61 secondattemptwaypoint 02:19.520 00:46.506 00:44.529 4 22
62 seren-kim 02:20.090 00:46.696 00:43.010 2 9
63 Brady11 02:20.159 00:46.719 00:44.513 7 7
64 LeagueRacerName 02:20.326 00:46.775 00:43.453 2 6
65 FightingJ2 02:20.602 00:46.867 00:45.148 0 18
66 awsracer3 02:20.883 00:46.961 00:43.479 2 1
67 CVA 02:22.045 00:47.348 00:43.331 5 29
68 NCMadMax57 02:22.073 00:47.357 00:47.064 0 3
69 PaulGomes01 02:22.097 00:47.365 00:47.139 6 6
70 GunjitBedi 02:22.383 00:47.461 00:44.019 8 15
71 Anser 02:23.735 00:47.911 00:45.393 1 65
72 RunBabyRun 02:24.072 00:48.024 00:47.476 0 7
73 Doug-Racer 02:24.076 00:48.025 00:47.802 6 19
74 GB2000 02:24.264 00:48.088 00:47.538 9 11
75 NathanLiang 02:25.225 00:48.408 00:46.191 4 5
76 kbdyd 02:25.756 00:48.585 00:42.746 3 23
77 AllenRacer 02:25.765 00:48.588 00:45.466 1 24
78 gg01 02:25.896 00:48.632 00:48.212 0 11
79 dhuTest 02:25.955 00:48.651 00:46.596 7 3
80 rnarukulla-1 02:26.253 00:48.751 00:45.385 4 29
81 Andrew-Gould 02:27.455 00:49.151 00:46.470 1 7
82 NowSomeDonkeyCar 02:27.590 00:49.196 00:45.255 7 145
83 nikhilkala 02:28.097 00:49.365 00:48.723 0 152
84 Markova 02:28.677 00:49.559 00:47.546 4 148
85 bradHorton 02:28.736 00:49.578 00:42.653 3 4
86 emerald-hill 02:30.023 00:50.007 00:41.554 5 15
87 aayush 02:30.615 00:50.205 00:47.990 1 3
88 xh 02:31.130 00:50.376 00:49.663 6 33
89 AaronAylor 02:31.316 00:50.438 00:48.515 2 8
90 racergang 02:32.958 00:50.986 00:49.800 3 6
91 JoesRacerX 02:33.033 00:51.011 00:45.403 2 21
92 SVDeepRacer 02:33.301 00:51.100 00:44.209 4 6
93 My-First-Deep-Racer 02:33.320 00:51.106 00:46.936 2 1
94 DRacerNeerajU 02:33.427 00:51.142 00:48.915 1 57
95 Jacob-Johnston 02:33.639 00:51.213 00:49.135 7 4
96 RobRacer 02:34.145 00:51.381 00:47.401 5 62
97 sbhimalingam 02:34.494 00:51.498 00:50.395 3 37

If you’d like to learn some Machine Learning skills while racing for glory, be sure to check out the AWS DeepRacer Page for hints on how to get started. Be sure to join the AWS Machine Learning Community on Slack for more tips and tricks on how to improve your model.

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.