Standard Drivetrain Comparison

[spiffyspleen]

As I was watching the best football team in the NFL win the super bowl last night , I was thinking about all the debate that has gone on in these forums between different types and iterations of drive trains. It came to me that there should be a quantitative way to compare these drive trains. There would be different categories which each drive train would be ranked in. These categories would be weighted differently for different games(i.e. in last years game incline climbing ability was very important but this year it is useless). There would also be an overall score which would be all of the categories added up and weighted in some way.
Here are some categories I have come up with so far:

Weight(different weight ranges could receive different amounts of points)

Speed

Agility(time to complete some sort of a simple obstacle course with cones)

Pushing Power( This would be a hard one to test, I was thinking it could be the amount of weight it could push on carpet, but then everyone has to have something with the same coefficient of friction)

Motors Needed

Pneumatic Cylinders needed

Obviously this is just an idea right now, so I would really appreciate the help of more experienced FIRST Participants in refining it. I am sure that there are a lot of things that I haven't thought of. If we could actually create a standard comparison between drivetrains though, I think it would help reduce a lot of confusion and be a lot of help to teams trying to choose a drivetrain. Thanks for your help, and GO PACK

[EricH]

My thought on a standard drivetrain comparison is that I could take two identical drivetrains, hand one set of controls to, say, ShaneP, take the other myself, and he'd beat me every time with identical drivetrains. Then I take one with "better" capabilities and he beats me with the first one. We switch, and he beats me again.

It's not the drivetrain you use that necessarily makes it better. It's how you use it. It's how you implement it. While some drivetrains have advantages in certain types of game or styles of play, an expert driver can neutralize those advantage with a solid drivetrain that they are extremely familiar with.

[Hawiian Cadder]

this is fairly simple

Co-axial is the best
swerve is second
octomanium, or some other wheel changing - third
slider or other 5-9 wheel omni setup - 4th
mechanum, omni, and tank all tie for 5th

[Chris is me]

Originally Posted by Hawiian Cadder

this is fairly simple

It is?

Co-axial is the best

How is it clearly the best? Are you saying there are no drawbacks to a coaxial swerve drive?

swerve is second
octomanium, or some other wheel changing - third
slider or other 5-9 wheel omni setup - 4th
mechanum, omni, and tank all tie for 5th

I don't mean to be harsh - but listing arbitrary "ranks" for types of drivetrains is not a comparison or evaluation by any means. Drivetrains have non absolute tradeoffs and benefits - there is not simply a "best" drivetrain, and certainly not one so obvious you can just list it without an ounce of data to back it up.

Qualitative statements about "best" and "worst" have no place in an engineering discussion.

[Andrew Schreiber]

Originally Posted by Chris is me

I don't mean to be harsh - but listing arbitrary "ranks" for types of drivetrains is not a comparison or evaluation by any means. Drivetrains have non absolute tradeoffs and benefits - there is not simply a "best" drivetrain, and certainly not one so obvious you can just list it without an ounce of data to back it up.

Qualitative statements about "best" and "worst" have no place in an engineering discussion.

Correct, squishy things like feelings have very little place* in the decision making process. You need quantitative data rather than your feelings.

I'm not going to pretend that I know all the answers to this but as part of one of my classes I am evaluating drive systems (specifically "swerve" systems) and their programming. This is for autonomous mobile navigation using on board sensors. As part of my proposal I determined some basic criteria with which to evaluate the various options. These are:

The drive systems will be evaluated on the following criteria:
4.1 Mobility
This will be evaluated by analyzing the degrees of freedom of the system and determining if the system is holonomic or not. It will also take into account whether or not the system can move at all.
4.2 Accuracy
This will be measured by determining where the system should end up for a given set of inputs based on the encoder readings of wheels. I will make several runs of a course. This course may be as simple as an "L" shaped course 2' long by 1' wide.
4.3 Ease of Use
I would like to measure how simple it is to code and/or drive each system but do not currently have an objective way of measuring this.

The goal of my project is not to find the "best" solution but to provide subjective data on the available solutions to be evaluated later. Interestingly, Chris and I were discussing this project (I defer to his experience with the VEX system when I have questions) and he suggested evaluating 6wd, nonadrive, and mecanum too. Sadly I just don't have time/budget to do those. Also, Vex mecanums are not available for purchase and designing them myself is beyond the scope of my project.

I digress, I think I provided a decent start on a list of evaluation criteria. Please add on criteria you would like to see (with the caveat that I might steal them for my evaluation).

PS: We also discussed having an autonomous challenge track where a 6wd would drive in auton against the various systems. This was only removed because I didn't feel like building an additional drive train + coding it + building a course. (My budget is coming out of my own pocket)

[spiffyspleen]

Originally Posted by EricH

My thought on a standard drivetrain comparison is that I could take two identical drivetrains, hand one set of controls to, say, ShaneP, take the other myself, and he'd beat me every time with identical drivetrains. Then I take one with "better" capabilities and he beats me with the first one. We switch, and he beats me again.

It's not the drivetrain you use that necessarily makes it better. It's how you use it. It's how you implement it. While some drivetrains have advantages in certain types of game or styles of play, an expert driver can neutralize those advantage with a solid drivetrain that they are extremely familiar with.

I definitely get where your coming from, which is why I tried to make the categories unaffected by the driver. Except for agility, all the other ones should be unaffected by the driver. I guess we need a better test for agility/maneuverability though.

[Joe Ross]

One way to do this comparison mathematically is through the use of a pugh matrix. JVN wrote a whitepaper about it. http://www.chiefdelphi.com/media/papers/2175?