Robot Wall Racers

Wicked!What sensors? Let me

Wicked!

What sensors? Let me warn you, even at the low speeds mine are driving by, I was getting close to the limit of the SRF05’s - You will be suprised at how much distance is needed to react.

And since the SRF05’s in reality do not do more than 1.5 meters, 50 mph will need some good grip from the tyres :slight_smile:

I know
First of all Thats what the remote overide is for! Im thinking of using a program that uses 2 webcams to see distance It says it can see up to 4 meters. it has wiked steering and can stop on a dime. Its experimental but its gonna be cool!

you have to coolest
you have to coolest ideas!! --TNT

awsome

haha, seems fun both to make an race :slight_smile:

 

but

"If something closer than 30 cm to the left, turn left." wouldnt he drive into every wall then?

You where absolutely right

You where absolutely right :slight_smile:

I altered the text, so I think it should fit now. Please mind that it was written before I actually made the robots :wink:

Nicely Done!

Very international. You have the European-standard advertising on the crash walls, and the good ol’ boy NASCAR ideal of “Go fast, turn left. Mostly.”

Looking forward to seeing the details.

Hey Frits, check out these:

Hey Frits, check out these: http://battlemachinesrc.com/

Laser tag wall racers, anyone?

Learning Telemetry

What is needed for these to eventually learn the track? If they learn the track they could fine tune their driving to eventually run the perfect time around the course.

I would be interested to know what you would need to build into the bots (or sent via wifi to a computer) for this to be possible.

Thanks.

Uh yes, I’d love to

Uh yes, I’d love to something like that :slight_smile:

Perhaps it coul dbe fun to just control some sort of canon on a robot, that drowe around without you controlling the driving? :slight_smile:

Sorry, I am an artie, I

Sorry, I am an artie, I cannot answer your techie stuff :slight_smile:

Anyone else?

ladar averages

well, in theory it couldn’t be that difficult. with components small enough, if they exist, i would assume a fantastic way to do it would be averaging ladar signals, to “paint” the track. an average of the corrections might allow the bots to virtually paint a line to follow it. interesting concept.

laser range finding and stuff is cool and all but to be sentinent, it’s not good enough for learned machines.

sounds like a fun project! anyone in london want to get into the research with me to make it happen??

processing, openframeworks and opencv would be good places to start with this…

I think you should try to

I think you should try to make a wall racer first.

I think you will be surprised, when theory no longer works in reality.

Speed and sliding… a fast sliding robot is something else to program. Try that first, and see if you can make that work at all :slight_smile:

Then think about mapping.

A RC car that can also be an autonomous car.

Last fall I started to plan, design, and purchase the parts to make a hobby level 1/10 size RC electric car that can run both under radio control and under PIC control. This car has a Tomahawk Reverse electronic speed control driving the wheels. I wish to keep that system.

All is going well. I have a Microchip PICkit 2 attached to the car. I have two batteries. One 7.2V for the driving wheels and another 6V pack for the PIC and the radio receiver.

The RC still works fine. I am having problems with the PIC control. I have no access to an oscilloscope. Under PIC control I can drive the wheels backwards. I can brake them. But for some reason I can not get them to go forward.

Has anyone in this thread accomplished what I am trying to do and has some code that they will share?

Ken

My experience is that it’s

My experience is that it’s hard to control the speed controllers with PWM.

A good tip is to use your computers soundcard to see what’s going on (see post “you may no tneed an oscilloscope”, somewhere)

Check what your old RC was sending to the speedcontroller, and see what you send to it.

Another tip is to take steps one at a time; very slowly work your way through PWM-speeds.

Reason it is hard is that they react on the speed that the PWM is changing…

**Success controlling ESC with from a PICkit2 16F887 **

By setting a comparator/counter to interrupt every 20,000 usec (20msec) ie 50 per second I have found a mechanism that is accurate enough to drive my Tomahawk Reverse ESC. I set up pulses ranging from 1.2ms to 1.8ms. The wheels respond with full back, medium backs, neutral, medium forwards and full forward. I am trying to say that the response is not bang bang, but continuous. This same system also drives the steering servo.

fritsl, did you ever post the code for your wall racers? If so, I have not been able to find it.

Many years ago I was coding in assembly, pascal and C. I am now retired and sitting in front of my homebased PC. Gone are all my old corporate resources and buddies. I have never used BASIC. Does anyone have a reasonably complex BASIC program that they would be willing to share. I’d like to get the feel.

Ken

To put it short here are

To put it short here are some pointers for you:

At the top of this page, look for "How to make Robot Wall Racers yourself, see here"

At the top of that page, look for

Attachment Size
wall_racers.bas 6.51 KB

 

Do not use big RC’s

This just strikes me; From the knowledge I have from you, I guess that you are working on perhaps a scale 1:10, some sort of “big” RC.

I must warn you regarding the physics; The cheap distance sensors (not laser or camera, but ultrasound or infrared) have a limited effective distance at max 2 meters in reality, and at it’s best.

A heavy car traveling at perhaps just 20 kilometers pr haour is VERY fast to “eat” 2 meters, especially when you are using them to navigate and not just a pure halt.

This is why it is smart with small light cars - and it is fun too if you tune them, as I did, feeding the motors an extra volt. This gives “radical” driving… where as making the larger ones “radical” and skidding, while only having 1-2 meters to react is impossible IMHO.

You should also consider that driving by nature gives a lot of false readings - this makes it even harder to scale up in inertia with these relative short distance sensors.

PS: Please take any further talks about building to the “how to make…”-page, thanks :slight_smile:

RE:
(baseoverapex’s picture)This is a Tamiya Blackfoot chassis, which was also used on a few other spin-off RC vehicles. Watchout for the “dog bones” also called half shafts that power the rear wheels. These are famous for wearing out on these chassis’s. They sell high quality replacements online using CV joint technology. I know off topic, but I’ve owned 10 of these and had these problems over and over…

RE wall racers
I built a simple wall follower robot for maze solving, and I use it in the same fashion as these wall racers. During presentations I like to have a kid (I do a lot of robotics mentoring at a HS school) sit in the floor and let the robot drive around them (circle them) The kids love it, as they feel they are a part of the robots actions, etc. Its like HUMAN - NASCAR!!! -lol

Video of this?
Video of this?