That goalie for the black team is right on the edge of being spectacular. He just needs to improve timing a bit. I saw a couple real nice dives that just came up short. #5 yellow showed some real potential out there, emerging superstar
Hi, we're a team who partecipate at Nao Challenge 2015-2016 in Italy. A part of this challenge consist in programming Nao to run faster than the other one(it is a race with three or more Nao). We thought that you have a lot of experience and we would like to ask you if you have some advise to give us. Thank you very much.
wow thats impressive! is this a new tech with the bots having a "swarm" mind? ahh i hate that i cant get involved in anything like this around central illinois >:(
***** interesting. thanks for the info. i need to look more into machine vision. ive had a raspberry pi 2 sitting around for months now, not know what to do with it (in a practical sense) but i think after seeing the list of used technologies for these leagues its time to bust it open. i take it the robocup challenges are purely software skill based and the teams cant manipulate the hardware in any way? besides the obvious software interaction with the hardware.
-iDol-Hands The Standard Platform League (which you saw UNSW winning in this video) uses robots built by Aldebaran Robotics, teams provide the software, but in most Robocup leagues teams develop their own software and hardware. There is a pure simulation league, if you're not comfortable with hardware
they are slow but today at Deerfield elamantry school I saw a nao robot and it did dancing it did a lot of stuff my favorite thing that it did was it walked down the hallway
+UNSW CSE Newb refs this time /smh Some random rambling: Firstly much better, much more entertaining. Huge improvement. The bots need a safer routine for moving the ball with more control and less distance. They need to be able to trade distance for control, instead of always stopping, pausing, and trying to kick it. That sniper kick routine needs to have some proximity criteria for deployment. Really should move away from the hive mind approach imo. I assume we are watching one organism with several parts. I remember in the other video a router crashed and they all completely clumped because they had no real awareness of each other. I think it needs to be changed. I can see where it would be tempting to have your pilot AI think like a single organism so that it can properly place the individual robots without them having to figure out where to go and what to do, but I think it would be better in the long run to make them as isolated from each other as possible for future realism and emergent swarm logic type behavior, also faster exploitation of brief rare opportunities and recovery from failed plans. After all, that's how humans play. Also that would requiring giving the robots a way to talk to each other, preferably using sound like people, but maybe ultrasonic so they aren't constantly making noise to our ears :) You could still have something like a coach AI that gives general orders to the individuals, but the bulks of the processing needs to be locally by unit, or else we're not really watching robots, we're watching two AIs play with remote controlled drones. I think that constraint, having to fit your program on hardware that walks, as opposed to a nearby tower, would lead to much tighter code and more elegant solutions, the kind of which you find in nature. Obviously you've got something interesting here :) And gz on the win, especially that last second goal :)
I don't think they use a hive-mind approach. As part of the rules, you are required to do all the processing on the individual robots themselves. In addition, there are strict packet limits (Wi-fi) on how much you can communicate between robots and no communication is allowed between the team and the robots during play. Typically, the robots are programmed with certain behaviors which activate based on some rules. So yeah, they are very much individual.
Why do they keep removing some of the robots?
That goalie for the black team is right on the edge of being spectacular. He just needs to improve timing a bit. I saw a couple real nice dives that just came up short. #5 yellow showed some real potential out there, emerging superstar
Why do we keep hearing Homer Simpson yelling "Doh!" every few seconds???
+Steven Dunn lol I noticed that too.
+Steven Dunn maybe someone's shutter sound
I guess the robots are programmed to do that noise when they get kicked.
Hi, we're a team who partecipate at Nao Challenge 2015-2016 in Italy. A part of this challenge consist in programming Nao to run faster than the other one(it is a race with three or more Nao). We thought that you have a lot of experience and we would like to ask you if you have some advise to give us. Thank you very much.
13:17 the real moment, you are welcome.
i dont know whats going on right now but this is awesome. are they player controlled or autonomous?
wow thats impressive! is this a new tech with the bots having a "swarm" mind? ahh i hate that i cant get involved in anything like this around central illinois >:(
***** interesting. thanks for the info. i need to look more into machine vision. ive had a raspberry pi 2 sitting around for months now, not know what to do with it (in a practical sense) but i think after seeing the list of used technologies for these leagues its time to bust it open. i take it the robocup challenges are purely software skill based and the teams cant manipulate the hardware in any way? besides the obvious software interaction with the hardware.
-iDol-Hands The Standard Platform League (which you saw UNSW winning in this video) uses robots built by Aldebaran Robotics, teams provide the software, but in most Robocup leagues teams develop their own software and hardware. There is a pure simulation league, if you're not comfortable with hardware
I LOVE ROBO THERE IS SO COUTE
Thrilling sports entertainment.
they are slow but today at Deerfield elamantry school I saw a nao robot and it did dancing it did a lot of stuff my favorite thing that it did was it walked down the hallway
Did anyone else see that ludicrous display last night?
13:17 the reason im coming here (from 9gag)
zecurox Tks captain!
+zecurox holy shit balls
You pick the ball up and they all look around 🤣
that s amazingly more than true soccer
So cool
The thing about yellow team is that they always try to walk it in
FUCKING REF FUCKED IT! PAID OFFF!
Mola
dr idooooooooooo
UNSW team know how to play.
This is so nerd... I love it.
The referee have something against the UNSW team...
B-human is a great team but because referee always fuck up in their favor, I always want them to lose
Like real soccer isn't
boring enough ?
+UNSW CSE
Newb refs this time /smh
Some random rambling:
Firstly much better, much more entertaining. Huge improvement.
The bots need a safer routine for moving the ball with more control and less distance. They need to be able to trade distance for control, instead of always stopping, pausing, and trying to kick it. That sniper kick routine needs to have some proximity criteria for deployment. Really should move away from the hive mind approach imo.
I assume we are watching one organism with several parts. I remember in the other video a router crashed and they all completely clumped because they had no real awareness of each other. I think it needs to be changed. I can see where it would be tempting to have your pilot AI think like a single organism so that it can properly place the individual robots without them having to figure out where to go and what to do, but I think it would be better in the long run to make them as isolated from each other as possible for future realism and emergent swarm logic type behavior, also faster exploitation of brief rare opportunities and recovery from failed plans. After all, that's how humans play.
Also that would requiring giving the robots a way to talk to each other, preferably using sound like people, but maybe ultrasonic so they aren't constantly making noise to our ears :)
You could still have something like a coach AI that gives general orders to the individuals, but the bulks of the processing needs to be locally by unit, or else we're not really watching robots, we're watching two AIs play with remote controlled drones.
I think that constraint, having to fit your program on hardware that walks, as opposed to a nearby tower, would lead to much tighter code and more elegant solutions, the kind of which you find in nature.
Obviously you've got something interesting here :) And gz on the win, especially that last second goal :)
I don't think they use a hive-mind approach. As part of the rules, you are required to do all the processing on the individual robots themselves. In addition, there are strict packet limits (Wi-fi) on how much you can communicate between robots and no communication is allowed between the team and the robots during play.
Typically, the robots are programmed with certain behaviors which activate based on some rules. So yeah, they are very much individual.
@@renderererer3572 Ahh ok thanks! That definitely puts the whole thing in a different light.
GO UNSW
i agree
Waste of human time.