Have you been to a lot of events? Not all events are set up consistently. Some regions are more consistent because they have dedicated volunteers - so I encourage all to volunteer at events …
That is not how engineering progresses through time what so ever. We still use principles learned thousands of years ago to help with modern solutions. It’s not as simple as “oh it’s outdated now so it’s completely obsolete and useless”. Secondly vex isn’t about giving kids all the things that a modern engineer will have. It’s about building the foundation of an engineers mindset, teaching kids to problem solve and think up creative solutions for challenging problems, to make a lot out of very little. This is what vex is meant to cultivate, but lately it seems we’ve been getting a little ahead of ourselves for the sake of accommodating to the rookie teams.
You misunderstand where fear is coming from. My fear is not for myself, my fear is for the community losing out on many great lessons that are being cut out due to shortcuts being added.
Not that I know how this sensor works or how REC intend to deploy it, but I do see potential for VAIC
LRT to happen.
Correct - I do not see this as a mandate for in-person EPs - but expect an opportunity for those Remote skills and Live Remote Tournament. VAIC will be undergoing whole hardware refresh this season - so can’t say it will be more or less disruptive to what they have seen.
Said every generation to the next generation
Theoretically, yes, this is what VEX was supposed to achieve. But, that’s simply not the reality. You can have the best engineers on the planet , but if you give them a vex classroom starter kit, no field, and no CAD, they probably wouldn’t be able to build a world class robot. VEX is quite expensive
As for the GPS sensor, I really don’t mind it. At the end of the day, abstraction will always occur. I heavily doubt that any of us prefer using old hardware over new hardware.
If the argument is that we will lose knowledge regarding odom, well, thats just a tradeoff that us, as engineers, make, in order to have a successful robotics program. I heavily doubt that anyone here would prefer manually multiplying, dividing, calculating trig in a program rather than have a method that does it. Back before these methods existed, programmers would have to manually do all of these functions. Now, we easily can use power and trig functions.
Also, I don’t see this as a loss of knowledge. Rather, I see it as an alternative way to gather data. I would much rather have a singular sensor that gives me data for position tracking than having encoders and gyro sensors. Less troubleshooting and less hassle. You would still have to do the task of moving the robot, but it makes the task easier.
I understand the frustration that some individuals felt after perfecting their odom and then seeing this sensor come out, allowing for newer, inexperienced teams to adopt a position tracking algorithm. But, they would still have to program their robot going to the certain coordinate. Without the actual movement of the robot, odometry is (no offense to the odom fans) kinda useless.
you’re missing the point, so i’ll give it another try.
when you do some long calculus for non-academic projects, do you use a calculator for the trig, integrals, and derivatives, or do you do it all by hand?
for me, the answer is, it depends. i’d encourage those who don’t have a full grasp on calculus yet, to not simply skip to plugging it into a calculator, but for those who do understand the concept already, to use the tools the engineers made for us to skip the tedious parts of the process so we can get to the actual engineering- the designing and application. this is analogous to the “GPS” vs traditional odom setup.
no one can stop newbies from skipping learning traditional odom to use the new sensor, but we can encourage them to learn how the “GPS” works, so that they can be best prepared to do the actual tough stuff- the motion profiling. this new sensor has the potential to raise the skill ceiling for everyone- including for those with no prior experience in position tracking.
this sums up my thoughts pretty well
What does that even mean? Vex is a starting point, not the end all be all of STEM education. It’s here to develop the fundamentals of engineering. Problem solving, the design process.
Besides, the best way to improve the future is to learn from the past. Age old lesson right there.
This is true, but this actually supports my argument. The first challenge of vex is actually starting, the first challenge is truly making something of nothing. It’s one of the biggest challenges, but the amount of experience and the number of lessons you learn from that trial is huge. When I first started the best tool I had was a Allen tool, and only had steel as an option. When I was finally able to get my hands on aluminum it was a game changer for me and I appreciated it more than I would’ve without my experience with only steel. And when I finally got better tools I appreciated them more than I ever would have. Things felt easier because of my experience with lesser tools, I appreciated them more because so, and was more skilled with them.
My concern is we are missing that experience. The trial of growing. We are allowing to many shortcuts.
If the trade off is between how easy your bot is to build and how much experience you gain then I will not classify that as a successful robotics program. At this point the program has lost sight of its purpose. The number of rewards you earn and how much experience you gain is not a causation, it is a correlation. The purpose of a vex robotics program is to ultimately learn.
Yes, but a far easier way to collect data that requires far less research, a point we both seen to agree on…
The concern is that it’s making things to easy on us to the point that we are losing out on experience. Experience that could help us greatly in future scenarios. That’s what vex needs to be about. The fundamentals.
Not knowledge per say, but experience that can lead to knowledge.
Exactly our point, it skips a step that can teach a lot.
One again, exactly our point. Odometry is a fundamental system used in so many industry systems all over the planet, something that any programmer will need to know, but vex gave us a shortcut. In the long run it will punish teams who try to learn it, and reward teams for taking the easy way out.
The issue here is that VEX creates an “artificial solution,” that the students should have the capability to solve and receive a competitive advantage for utilizing their intellect, as compared to “let’s solve all of the STEM problems for the students instead!”
There are many problems with creating an artificial solution to problems, because they will create long-term damage if the artificial solution is swiftly dropped and students are now having to rely on their head. This is why students are not allowed calculators in math class, to encourage intellectual thinking (even though calculators are everywhere). The further VEX pushes students from the natural world of society, the less students are prepared after hitting a brick-wall after graduation realizing that what they learned in VEX barely put a dent in preparation towards STEM, ESPECIALLY SINCE COMPUTING WILL EVENTUALLY DOMINATE THE STEM-FIELD SOONER IF NOT LATER.
By continuing to eliminate practices and needs in computer coding in the VEX ecosystem and nowhere else, opportunity is lost. The aspect of “I need this made, so I should construct code from my math understanding” can help students figure out and practice how to apply mathematical aspects into code. The GPS sensor creates an artificial component left to the individual to figure out on their own after being pushed into the real-world, because VEX decided to artificially solve one of many essential real-world problems for students.
Yet again, if there is artificial solutions where money is introduced, then the aspect introduced should translate into the real-world. Calculators are everywhere and you can utilize them on nearly every device now, where as the VEX GPS sensor can only be utilized in the VEX ecosystem and nowhere else. YET calculators are still not allowed in math class to encourage students to use their head (and if you argue that “VEX is not meant to be academic” then you really do not understand that VEX is meant to provide real-world examples of academics, not a gamer-boy real-world PS4 competition). Things like a GPS sensor should only be made, if and only if it is an “artificial solution nobody has to solve because it’s accessible everywhere they go” except for the GPS sensor it isn’t, and only accessible and usable in VEX.
Compare the grades of a class where you tell the students in class 1 “You don’t need to learn this,” then tell the students in class 2 “This is essential for your life and it would prepare you for the future.” The students in class 1 would be less motivated to learn because they would feel like learning it would be a waste, while class 2 would learn it as they feel importance in the material learned. In similarity, VEX is pretty much saying “You don’t need to learn odometry anymore as there is no drastic and rewarding impact to learn it for your 4 years in VEX robotics,” so how the hell are students going to learn the material of odometry for futher learning? I call BS about what you saying about the GPS sensor “Raising the ceiling” In fact, it will DEFINITELY lower the ceiling as compared to raise.
I agree with @holbrook and others that it would be premature to render judgement on V5 GPS sensor before you see its full specs, programming API, and test statistics. But if you must arrive at some conclusions now, based on the partial information from the patent document and a handful of forum posts from VEX engineers…
First of all, I think, there is no reason to declare it all doom and gloom for the odometry experts and aficionados.
V5 GPS sensor will return position errors that, at the very least, are as large as the field assembly tolerances, but likely even larger. To account for such errors on each field you would need to calibrate the filter by sampling multiple sensor solutions (x,y,theta) from the fixed field position(s). But, so far, there was no indication that its internal filter is going to have any calibration capabilities - that’s why we need to wait for the detailed specs and API release.
While you would expect (theta) to be somewhat accurate, I don’t think (x,y) components are going to be that good and able to beat well tuned odometry, based on the quad encoders, IMU, line followers, perimeter resets, and Vision Sensor looking at the game objects. That’s my personal guess - we need independent test statistics from the multiple field setups to know for sure.
So my best recommendations, based on what we know now would be:
1. If you are running a school engineering class (like PLTW) or you have a large club with multiple up and coming teams and you have extra funding, then I would buy a couple of V5 GPS units per approx 10 teams.
Competitive robotics is not the biggest VEX market and you cannot underestimate usefulness and inspirational value of giving students a taste of full robot position knowledge, available just after a few weeks of instructions. Think of all those autonomous passenger, freight, farm, and mining vehicles out there.
However, I wouldn’t spend too much money on GPS. And, if budget is constrained, I would rather buy extra motors and mechanical components, because they could be used all season long and it would be a waste to have expensive V5 GPS units to sit on the shelf most of the time - one GPS per school or club is still plenty for inspiring students.
2. If you are more experienced VRC team, that is planning to use odometry and motion profiling, then you may get V5 GPS, if it is ruled VRC legal for programming skills, but only if you have plenty of funding. It could buy you some easier life at the beginning of the season.
However, be prepared to be disappointed, if one of your fellow teams beats you at programming skills, relying purely on encoders, IMU, and clever field resets. V5 GPS is definitely not for teams with constrained funding - you can get more bang for the buck by investing into mechanical components, learning more in depth algorithms, and using sensors that you already have.
V5 GPS alone is not going to let you magically win out of the box in any competitive region. Because, I think, it will have large field specific errors and you have to learn more advanced stuff to close the control loop anyway.
3. If you are one of the advanced teams, that have experience with odometry and considering learning and using Kalman Filter to up your game, then your decision to buy V5 GPS should be purely based on VEX making time stamped raw measurements from the optical and IMU sensors available with their API.
If API only provides the final solution from their internal filter and no raw measurements - then I wouldn’t buy it.
Even if you have ample funding to burn - it will be waste of your money and time, accompanied with a lot of frustration trying to reverse engineer hidden internal filter states. You will have much better success running your own EKF with raw measurements available from other sensors and you will have all the freedom of properly calibrating it on each field if necessary.
Yeah, (note: Not sarcasm) and in a couple generations in the future it is inevitable that AI will outperform Engineers. Eventually, all you likely only gotta do is type in what you want invented, and all you have to do now is click a button that says “Run” and the AI does the rest.
How far do we want to go with this? What is enticing about competing in VRC if it is just gonna become a VEX IQ but with metal parts?
just a brief correction that motion profiling actually doesn’t require odom feedback at all to function
An artificial solution to an artificial problem. VEX’s constraints for position tracking are completely made up. Encoder odometry is not a skill used in all but the most niche professions. Anything a student could gain from learning how to do it could easily be gained from one of the million other things you can do with VEX.
i have an idea
If you don’t want the new sensor, don’t buy it
Pretty amazing divergence in the topic since Friday when a picture, a price, a CAD drawing and a 6 week ship date were posted. I appreciate the dissection of the patent (yes @technik3k the legalspeak was a killer) and James answering some questions.
But want to jump in on the competition part. I’m here (and have been here) for one simple reason. That is to inspire people to look at a STEM based career. Sadly we all can’t work at fast food places, some you will need to take on STEM jobs, be they electrician, programming, bio-sciences, etc. jobs.
To be successful at it, you’ll need to do some math or at the very least have a general overview of how the math works. But you can still compete and be inspired without being able to write your own PID loop or Kalman filter. But I think it’s great that there is a API (I hope) that says “You are here”) but also access to the low level data to be able for the more advanced teams to roll their own.
But I think it was @Connor that posted if VEX / RECF was going to allow / use a $200 sensor, it was time to open up 3D printing. I would put higher value on 3D printing since that it takes some success in design skills (a key component in the competitions) the ability to use CAD (a direction lots of teams are going in now) and an alternative fabrication process (which is a big thing that is happening NOW)
So I’d like to see a real campaign start with the GDC @Grant_Cox and RECF @DanMantz to get 3D printing added to the rules. The VEXU rules around it would be fine, I’d like to see one addition and that is “After the robot season ends, all printed designs need to be posted to a public site”. That way we can all learn. I’ve seen some of the design challenge parts and would love to use them on some of my robots.
Looking forward to @DRow update to the “Position Location System” with descriptions, API, sample code, etc. But also passing on Kudo’s to you buddy, you got a huge buzz with under a 100 words of text and a picture, some major marketing moves going on there.
Something something “The Only Motor You’ll Ever Need”
That sounds fine enough… but how in the world would you even begin to enforce that? And personally, I’d be against posting the actual 3d files, merely pictures, so that way people can’t just download the files and use them compared to making their own parts.
Again, this seems to be gatekeeping behavior. To me, sharing the 3D files (presumably created by students) is the epitome of student-led learning.
Again, I’m not an educator and maybe my views on students “learning the fundamentals” is skewed. But the history of progress is generally one of making small improvements to what came before you. Artificially creating “resets” with each season seems to artificially reward teams or organizations with institutional memory.
I can see that. Obviously, in a perfect world, they’d take designs, improve them, whatever, go crazy. But realistically, most teams are just going to download the file and print that straight without even modifying it, which kinda defeats the purpose of the whole thing, since they’re not really learning how to design parts nor how to CAD, which would be my main concern.
I want the CAD available after the season for EXACTLY THAT REASON SO OTHERS CAN PRINT ONE!
Not much different than using premade code OKAPI anyone?
We want to lift teams, I get your CAD, make improvements to it and the IMPROVED one gets posted. Lather, rinse, repeat. If they don’t improve it, it’s also fine, they are still using 3D printing, that may inspire someone to make a different part.
Like mine is skewed too Learning happens in lots of different ways.