Am I crazy?

[TPGRacing75]

Call ,me crazy. but no matter the scale, I want at least 12-16 ghost cars to be ran at once. As I'm a solo racer, I would ideally like a reasonably accurate field of cars, and 6 just does not cut the mustard. I have been wondering how to get around this problem, as I am in a unique situation.

Most people don't care about having only 6 cars at once because rarely will you have more racers, especially needing marshals. But as a solo racer who can go at their own pace, having more cars is a boon. I have come up with many crazy ideas, one of which is the stupid thought of hooking up 2 ARC Pro powerbases, running 2 separate races, and having a external central lap counter keep track of the cars. So while technically I am running 2 races, one with me, one with just ghosts, The central lap counter would keep track of both races and merge them into one tracker.

Of course the more sensible idea is BLST and Oxigen, which supports 20 cars natively, and avoids any nasty crashes caused from a having so many ghost cars.

Am I just crazy, or are their others who share my ambition? :think:

[TuscoTodd]

The idea of having to compete with 12-16 ghost cars sounds pretty entertaining when running a limited (in your case one) number of human drivers! Definitely would make it a lot more challenging and fun in my book! :D

[patwa]

Like TuscoTodd says the idea sounds great.
I find that just running four or five ghost cars is a nightmare, no pun intended, due to the fact that they end up in a train, as they all run at slightly different speeds.
You can handicap the speeds a little to keep the cars apart, but eventually they all end up nose to tail.
Also the ghost cars are not as fast as you so they stay on the track in the turns. This can be altered slightly with different strength magnets.
I find the best way is to programme 1 car per lane and then practice clean overtakes, weaving through the traffic without having to slow down and wait for a passing spot.
Sure helps when the other drivers do show up as well!
Have fun.

[GRUNZ]

it depends what you mean with 12-16 pace cars...do you want them to be able to overtake each other (to reduce the train formation) and to avoid crashing into one another? in order to do that, the software controlling the cars has to have a clear idea where each car is on the track by sensing the cars.
All the digital systems "see" a car when it passes over the start/finish line. The only system that allows the race management software to see cars also when they go over lane changers is Scorpius. I know that with Scorpius one can record a pace car lap and then the system will replay it and because of the extra sensing it is quite accurate. But I am not sure if it can handle 12 or 16 different ghost cars.

BLST is not a system per se, it is more a technology on how the lane changers are operated: it still needs to be embedded on a digital system like ssd, carrera, or oxygen (here as a digital system I mean a digital protocol).

[TPGRacing75]

I had a feeling it would just be a pipe dream. I still wait for the day when I can race solo against 12 or so other computer driven cars outside of racing video games. I feel like a BLST track with an oxygen digital system might be the closest I can get.

[buspor63]

It could probably be done with Carrera and some DRS Lane Gates on each LC. Better get the 10 amp power supply though.

[slothead]

As I see it GRUNZ has it right - the biggest problem for starters is position tracking. Knowing where a car is on the track is needed for effective intelligent control. One of the reasons I got back into slot car racing in 2005 was to use it as a test bed for artificial intelligence projects. At that time I had a Scalextric track set up in my lab and was working with a computer science student on his senior thesis. We used images from a digital camera hung over the track to determine the position and direction of a car. It's doable but doing it quickly and being accurate are not easy. Speed and speed changes can be calculated by using info from several previous frames. The student did very good work, but it was complex and we never finished the project. He did earn an A on his thesis based on what was accomplished!

So being able to race against a bunch of intelligently controlled ghost cars is possible, but would require computational processing speed and good software that could either learn to drive a car on it's own (machine learning) or capture a human driver's actions and extract patterns. A circuit would be needed to convert the computer output signal into a voltage value to be sent to each car. The problem is actually simple considering there's only one output variable per car - voltage. Quad copter guidance systems have to compute and output multiple variables in real time.

An alternative would be having intelligent cars rather than a master control system. Sun Microsystems sponsored a contest about 15 years ago in which teams were given small circuit boards that could measure speed and direction changes to mount in slot cars. The teams then had to write programs to load on the circuits to process the data and control voltage to the motor. Contestants had no advance knowledge what the track used on test day would be like so there were mixed results. I prefer the master control system approach because each car doesn't need to be altered and the 'system' can also control the overall race. Such as if it noticed there was no current draw for a given car it could turn on caution lights and lower voltages for all cars.