Register    Login    Forum    Search    FAQ     Formula Hybrid Home Page

Board index » General » Everything else




Post new topic Reply to topic  [ 11 posts ] 
Author Message
 Post subject: Driver Requierments
 Post Posted: Wed Jun 06, 2012 6:32 pm 
Offline

Joined: Thu Sep 09, 2010 6:32 pm
Posts: 68
Location: University of Idaho
Is there a rule the requires that the drivers be human? We were thinking about getting a sponsorship from Honda for an ASIMO (http://world.honda.com/ASIMO/control/) or hitting up Google for one of their self-driving vehicles.

Also if they do have to be human, how human are we talking? Can they have robot hands, legs, etc.?

And furthermore, if no robotic limbs are allowed could we just have a human driver in a vehicle that drove itself and the driver was only there for backup?


Top 
 Profile  
 
 Post subject: Re: Driver Requierments
 Post Posted: Wed Jun 06, 2012 7:41 pm 
Offline

Joined: Mon Mar 12, 2012 7:07 am
Posts: 75
There are references to human body parts in the rules such as skin, feet, head, and shoulders. These could potentially be replicated by a robot, but that would probably be outside of the scope of the competition.

Using robotic limbs for precision operation of the vehicle would probably have to be allowed. If they were not, FHSAE might come under fire for discriminating against handicapped or disabled individuals.

A car that drove itself, with a person along for the ride, is probably the most reasonable. Autonomous vehicles are a very interesting research subject, and if successfully built, would be an impressive feat of engineering for this competition.


Top 
 Profile  
 
 Post subject: Re: Driver Requierments
 Post Posted: Thu Jun 07, 2012 8:19 am 
Offline

Joined: Wed Jun 03, 2009 1:29 am
Posts: 117
Location: Texas A&M
A self-driving hybrid would be pretty impressive, and if you could go driver-less, you could get rid of a lot of weight out of the car (I think the driver is the single heaviest thing in the car by a large margin), which would improve performance and energy efficiency.

However, has anyone (Google, etc) made a autonomous driving system that can actually drive faster around a course than a talented human? I haven't kept up too much on the latest technology, but a really good human driver does so many things to make the car fast, I'm not sure we currently have the technology to beat one.

_________________
- Texas A&M 2009 FH Crew Chief
- 1st place FH '09: 981 pts


Top 
 Profile  
 
 Post subject: Re: Driver Requierments
 Post Posted: Mon Jun 11, 2012 10:40 am 
Offline

Joined: Mon Mar 12, 2012 7:07 am
Posts: 75
In this last competition there were only 6 teams that competed in the acceleration events, 9 in autocross, and 14 in endurance. If a self-driving vehicle were to compete, it really wouldn't matter if it was an extremely fast driver; as long as it finished every dynamic event it would be a top ten car minimum, possibly a top five.

Also, suppose energy consumption would be monitored in real time. In the endurance event, if it calculated there would be excess fuel by the time it finished, or an insufficient amount of fuel preventing it from finishing, the vehicle could quite possibly be programmed to increase or decrease its average speed. Realistically it doesn't have to be the best or fastest driver to win the competition, it has to be the smartest and most accurate. I believe that a self-driving car would be very hard pressed to preform better than a human in the autocross event, but in the acceleration and endurance events, maintaining a certain torque or fuel consumption could be done more accurately by a computer.


Top 
 Profile  
 
 Post subject: Re: Driver Requierments
 Post Posted: Tue Jun 12, 2012 10:56 pm 
Offline

Joined: Wed Jun 03, 2009 1:29 am
Posts: 117
Location: Texas A&M
Riviera wrote:
In this last competition there were only 6 teams that competed in the acceleration events, 9 in autocross, and 14 in endurance. If a self-driving vehicle were to compete, it really wouldn't matter if it was an extremely fast driver; as long as it finished every dynamic event it would be a top ten car minimum, possibly a top five.

Also, suppose energy consumption would be monitored in real time. In the endurance event, if it calculated there would be excess fuel by the time it finished, or an insufficient amount of fuel preventing it from finishing, the vehicle could quite possibly be programmed to increase or decrease its average speed. Realistically it doesn't have to be the best or fastest driver to win the competition, it has to be the smartest and most accurate. I believe that a self-driving car would be very hard pressed to preform better than a human in the autocross event, but in the acceleration and endurance events, maintaining a certain torque or fuel consumption could be done more accurately by a computer.


If the competition at hybrid ever gets stiffer, you'll need all the driver performance you can get to win. Driver and tires are the two most important aspects to a car's performance, for hybrid or regular FSAE.

As far as fuel consumption, a computer could aim for a target easier, no question, however a good human driver will have a significantly higher corner speed than a computer, which won't be able to use all the grip the tires provide. Higher corner speeds allows you to run decent laps without wasting energy accelerating/decelerating. Then again, the car losing 140 pounds of driver weight will definitely help in efficiency.

It would be interesting to see how it would shake out. How will the computer read the sea of cones for the auto-x/endurance course?

_________________
- Texas A&M 2009 FH Crew Chief
- 1st place FH '09: 981 pts


Top 
 Profile  
 
 Post subject: Re: Driver Requierments
 Post Posted: Wed Jun 13, 2012 8:17 pm 
Offline

Joined: Thu Sep 09, 2010 6:32 pm
Posts: 68
Location: University of Idaho
With traction control and ABS systems being what they are today I would argue that a computer could outperform an amateur driver.

I agree that this is only if the vehicle/computer had a way to interpret the cones and this is no small feat to say the least.

There are probably several different ways to tackle such a problem but one that we have thought of is GPS way-points. This could be done when drivers are allowed to walk the course before the events. A driver could have a gps unit in their pocket and walk the racing line or the center line of the track. The vehicle could then be given that data to determine max speeds, breaking zones, etc. and the system could adjust on the fly to stay within the traction of the tires. Passing lanes and other issues such as other vehicles could also be worked into the driving algorithm or inputs could be sent to the vehicle via a team member on the side of the course.

That said I think its well within the ability of the current technology out there to build a vehicle that could out drive a human.


Top 
 Profile  
 
 Post subject: Re: Driver Requierments
 Post Posted: Thu Jun 14, 2012 8:48 am 
Offline

Joined: Wed Jun 03, 2009 1:29 am
Posts: 117
Location: Texas A&M
A GPS track map could work, but I wonder if it would be accurate enough. Commercial grade GPS is only accurate to ~3 feet, which isn't quite accurate enough to lay out a perfect run, which would ideally be something like less than 3 inches from each (important) cone.

_________________
- Texas A&M 2009 FH Crew Chief
- 1st place FH '09: 981 pts


Top 
 Profile  
 
 Post subject: Re: Driver Requierments
 Post Posted: Thu Jun 14, 2012 6:34 pm 
Offline

Joined: Thu Sep 09, 2010 6:32 pm
Posts: 68
Location: University of Idaho
What if the system relied on a local point of contact rather than a satellite i.e. a tower off of the team trailer? This would probably increase the accuracy of the system. I'm not sure to what degree, 3" might be a little bit difficult but with the loss of the weight of the driver and a buffer built in, it might just work.


Top 
 Profile  
 
 Post subject: Re: Driver Requierments
 Post Posted: Fri Jun 15, 2012 1:52 am 
Offline

Joined: Wed Jun 03, 2009 1:29 am
Posts: 117
Location: Texas A&M
wos9722 wrote:
What if the system relied on a local point of contact rather than a satellite i.e. a tower off of the team trailer? This would probably increase the accuracy of the system. I'm not sure to what degree, 3" might be a little bit difficult but with the loss of the weight of the driver and a buffer built in, it might just work.


That's a good question. Obviously the closer it can get, the better. If it is accurate to say 6", to guarantee it will never hit a cone it has to be programmed 6" off the cone. If it errors 6" in the other direction, it will be a full 12" off of the cone, which in a slalom could start to get pretty ugly. Obviously this is a worst case setup, but something to consider.

But losing a driver is a lot of weight...

_________________
- Texas A&M 2009 FH Crew Chief
- 1st place FH '09: 981 pts


Top 
 Profile  
 
 Post subject: Re: Driver Requierments
 Post Posted: Thu Jun 21, 2012 1:53 am 
Offline

Joined: Thu Sep 09, 2010 6:32 pm
Posts: 68
Location: University of Idaho
Just watched a TED talk on the Google self driving car and they showed a few clips, one of which was this one, http://www.youtube.com/watch?v=-cj375UZyaI&feature=fvwrel obviously it still has some bugs to work out but perhaps with more prepossessing power and more precise sensors it could be as competitive if not more in comparison to a human driver.


Top 
 Profile  
 
 Post subject: Re: Driver Requierments
 Post Posted: Fri Jun 22, 2012 1:32 pm 
Offline

Joined: Mon Mar 12, 2012 7:07 am
Posts: 75
wos9722 wrote:
Just watched a TED talk on the Google self driving car and they showed a few clips, one of which was this one, http://www.youtube.com/watch?v=-cj375UZyaI&feature=fvwrel obviously it still has some bugs to work out but perhaps with more prepossessing power and more precise sensors it could be as competitive if not more in comparison to a human driver.


That is impressive. At the end notice the Max lat, which I assume to be the maximum lateral acceleration was 1.08 g's. Imagine putting that same system on a formula car! Plus they were "nicking" cones, not knocking them over. To me that says sensors are accurate enough to steer as well as a human, but can it brake, accelerate, and manage its shifting as well as a human? I think that would be the bigger challenge.


Top 
 Profile  
 
Display posts from previous:  Sort by  
 
Post new topic Reply to topic  [ 11 posts ] 

Board index » General » Everything else


Who is online

Users browsing this forum: No registered users and 1 guest

 
 

 
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
cron