self-learning remote control car

Thread Starter

seesy123

Joined Mar 14, 2008
13
hey,all of you.do you know how to make self-learning remote control car.For example if i control the car using remote control and let it move to 1 track or 1 around.when i press a switch ,it will follow back the previous track that i control without controlling by user.The problem is i don't familar with the PIC16F84..i don't know how to store the previous instrcution that i control the track and save it in memory..any idea.or do you have any the schematics about remote control car..thanks..
 

beenthere

Joined Apr 20, 2004
15,819
While we won't design this project for you, we will lend advice.

How were you planning to utilize the PIC? How will you interface it to the controller/car?
 

Thread Starter

seesy123

Joined Mar 14, 2008
13
this is what i draw for this project.ya..i want someone teach me cause i don't know deeply about them.i stuck with the PIC programing.i just know how to write the programing let the car move,but the problem is i don't know how to let the car follow the previous track without controlling.:confused:
 

Attachments

beenthere

Joined Apr 20, 2004
15,819
You may need to learn a bit more about hardware, too. Your H bridge controller will not work as drawn, nor do I understand what it is supposed to accomplish.

Before you get much farther - how do you propose to track the RC car? Will the microcontroller use the same RC controller to run the car? Is it a commercial RC car, or do you propose to construct it from scratch?

As my tag suggests, you need to work out all the details about the hardware before you start selecting the microcontroller.
 

Thread Starter

seesy123

Joined Mar 14, 2008
13
Actually it's sch project..Taking the simplest case, say RC car only moves forwards and backwards. For every command i send, i store the values in memory, and RC car moves for a certain time for each.

When you send the repeat command, it read the values back from memory one by one and move accordingly.
 

SgtWookie

Joined Jul 17, 2007
22,230
Actually it's sch project..Taking the simplest case, say RC car only moves forwards and backwards. For every command i send, i store the values in memory, and RC car moves for a certain time for each.

When you send the repeat command, it read the values back from memory one by one and move accordingly.
That would work, if it were for a railroad, and the engine was geared to the track, or could count the railroad ties - so that it would always know precisely how far it had to move in which direction, and which switches had to be thrown in order for it to arrive back in the original position.

However, you are talking about a model car that is not positively coupled to the ground. There will be errors induced by wheel spin, wheel slip, small irregularities in the surfaces it drives over - and the limitations of the uC itself due to the number of bits of resolution. I'm afraid that you will not be very happy with the results.

You would really need some kind of GPS device - but even those are not accurate enough; the best you can hope for is to get within about 10 feet of actual location. That won't work well for reverse navigation down a narrow hallway.
 

blocco a spirale

Joined Jun 18, 2008
1,546
It would be easier to start with a 2-wheel robot buggy chassis rather than a remote control car and perhaps use opto-sensors to count wheel revolutions.

You should think about how you represent instructions and data in memory e.g. 2 bits of a byte could represent up to 4 instructions such as drive left wheel, drive right wheel, drive both wheels, drive no wheels etc and the remaining 6 bits could contain the wheel revolution count value.

Each time you send a command to the buggy it saves a byte of date in RAM. During playback step through each byte and execute the instruction for x wheel revolutions.

Maybe, forget the remote control bit and enter commands via a keypad e.g forward 10, left 2, right 1 etc etc etc.
 
Last edited:

Thread Starter

seesy123

Joined Mar 14, 2008
13
i confuse which one PIC should i use in my project.i want build a remote control car where it will repeat the command that i send to it without any controlling(let it move to 1 round.when i press a switch ,it will follow back the previous track that i control without control by user)..Each time i using the remote control to control it then move in any direction where controlled by user and at the same time it will automatic save the data in RAM.during playback step(without any controlling),it will read out the command or instruction that have save previous in RAM..so what type of PIC i need..and do you have any website about writing a program to read out the command from RAM..thanks
 
Last edited:

jpanhalt

Joined Jan 18, 2008
11,087
As SgtWookie so aptly described, repeating the signals will not make the car repeat its path.

I suggest you try your idea with an RC car and see what happens when you try to enter the same controls manually.

If you want it to follow the same path, you need a tracking mechanism.

John
 

beenthere

Joined Apr 20, 2004
15,819
Perhaps if the motors were stepping motors and the steering were very precise, it might be possible to bring the car back to the initial starting position without sensing. That raises the cost pretty significantly.

Do you want this car to be able to go anywhere and be controlled back, or to run on a defined course?
 

jpanhalt

Joined Jan 18, 2008
11,087
Yes, Seesy123,

You have re-stated your intent several times. Data logging the received commands is quite doable and commercial units are available for that. (Look up data logging for RC airplanes.)

The problem I and others have pointed out is that your car will not follow the exactly the same path using the same sequence of commands on different trials. There are too many other variables.

John
 
Top