PBP projects for R/C models

[rmteo]

....Ever tried to explain to a thirteen year old how a microwave oven works?

Ken

How Microwave Cooking Works

[mackrackit]

Do not under estimate their ability, the kids that is.
One example.
http://www.picbasic.co.uk/forum/showthread.php?t=10200

When YOU take the attitude that something is going to be difficult for someone to learn, then as long as YOU are doing the teaching it will be hard for them to learn.

Children are much easier to teach new concepts to than adults are. Adults "think" they know how things work. Remember the trouble you had with BASIC? That was because your understanding was based on pre-conceived notions from what you learned in the past. Children will not have that problem.

[Kenjones1935]

However, the only way to reach the students in a public school is through the teachers and the administrators. My question is what is the best way to communicate to them. It is a marketing issue. A teach-the-teacher issue.

Ken

[mackrackit]

How do I communicate that to adults schooled in Education, not Engineering or Computer Science.

I do not think that should be a problem, many of the users on this forum do not have a formal background in Engineering or Computer Science. I will bet there are plenty of teachers that have RC or some other gadget related hobby. Do they have a "shop" class.

Market to the school district... How many school board members are in local industry? A google search makes me think that Parker has a lot going on in your area, New England Dairy Association, and of course the famous secondary schools in your area should make it easy to convince the schools of the importance.

[Kenjones1935]

I am getting closer. This video shows a 'real' car that can parallel park itself. Using bits from old videos plus some voice over I try to show that making radio control vehicles autonomous is relevant, practical, and fun. I believe there is great potential here for our public schools which are struggling with STEM.

[Kenjones1935]

I've actually made three robots. new

* Reply
* More
o edit
o You can't chill your own stuff
o Report Spam

This morning I realized that I have actually made three different robots. Three different RC cars with three different physical charcteristics dictated this.

First I built the 1/10 scale model level car based on an HPI SPRINT. It has detailed miniature suspension and handling features, great speed, and proportional control systems based on pulse width modulation. My PIC kit for this car is simple because the controls are digital signals. All that is needed is the PIC and a DPDT switch. The switch - under PIC control - selects which PWM signals (those from the PIC vs those from the radio receiver) go to the steering and driving electronics. The PIC code is quite simple because of the car's capabilities. At $200 - $300 for the car and my kit this seems too expensive for our public schools.

Second I built a 1/12 scale toy level car. This car has bang bang controls. It requires the PIC, the SONARs and four DPDT relays to steer the DC current into the correct connections. It runs slowly enough for the PIC and the SONARs to keep it within reasonable behavior limits. It can be stopped in mid 'flight' by turning ON the radio transmitter. The price is right for our schools, but I feel the car is not fast enough nor exciting enough. It is also too small to carry my KIT under the car skin.

Third I built the 1/10 scale toy car. This car is much faster than the 1/12 scale toy. It too has no proportional controls. The added electronics kit is the same as the 1/12 car, but the code needed adjusting. ie the 150millisec pulses. I think the $50 car price is within budget. My KIT is complex because of the four relays, but hopefully doable by a dedicated teacher and class. The engineering is interesting because, as one of my videos shows, the behavior is a bit unpredictable. Maybe if I had some local help and a more sophisticated program ------

Ken

[Kenjones1935]

Happy New Year all.....

I have added another video It shows the performance of the 1/10 scale TOY car and two versions of the 1/10 scale MODEL car.

It shows the TOY and a MODEL racing. The lap counter (me) is lacking, but I think the excitement and the attendant appeal to middle school kids is evident.



Ken

[ScaleRobotics]

If you used a hall effect sensor like this Melexis 90217 Hall-Effect Sensor:

and put a magnet or two on one of the wheels, then you could use the CCP module with DT_INTS using CCP1_INT to measure the time between pulses. You would measure the circumference of the wheel and do a calculation to get speed. You would do something similar with your front sonar. You would take a reading, and then at a specified time later, take another reading and compare. If you were headed straight for a wall, the distances would be pretty similar, etc, etc.

Here is a hardware example, using hall effect sensor, a lathe, and a basic stamp: http://www.parallax.com/Default.aspx?tabid=420

[ScaleRobotics]

At high speed, I worry that without some sort of reference point, your car will get "lost". If you add a cheap line sensor, you could add something like a 4" black tape reference point to one of the walls or floor beside one of the 4 walls. How accurate are your sonar readings? Can it measure the room from each corner successfully? If not, then you may need to add a cheap wheel sensor to measure distance. Really both of these would be dirt cheap, but yet add a new dimension to your car finding itself, as well as sensing it's speed. (although you could use your sonar to get speed. It would not be accurate as you turn a corner though. It will think you are going backwards as it senses the wall getting further away)

Has the "Robotics for Dummies" paper back been published?

I think it's scheduled to be released right after Pic Basic Pro for dummies.

[Kenjones1935]

The WEB says that a standard hockey rink is 200' by 85' with curved corners on a 10' radius. Here in Fitchburg we have both indoor ice hockey rinks and outdoor street hockey rinks where the players use roller blade skates. The rinks are bounded by an approximately 4' high solid wooden wall.

The advantage here is speed is possible; 25mph +- for electric cars well over that for nitro powered gas engined cars.

I just found a WEB site extolling the 2D IR proximity sensor. Two potential problems - outdoors in bright sunlight and multiple cars. What do you think?

I did not read the specs for IR sensing. They are talking inches to a few feet. I need at least ten feet, probably more.

http://www.societyofrobots.com/senso...pirrange.shtml

Ken

[Kenjones1935]

I think my little car needs to know where it is relative to the wall it is trying to follow and what maneuver it is presently trying. Right now the code has three states:
1. Trying to go straight along the wall.
2. Trying to turn left sharply because something is in front, but not close enough for state #3.
3. Trying to brake and backup and turn to the left enough so it can return to #1 above.
There is a forth unofficial state in which it is completely lost. It usually gets there from state #2 by turning too successfully. It thinks it is in state #1 but is too far from the wall. If the room is too big it just keeps turning right making a big circle.

It has problems in each case.
In #1 it swerves rather than going straight. If it knew better its rate and angle of approach it could better decide how hard to make its corrective turn.
In #2 if it turns too hard and too successfully it becomes lost. If it does not turn hard enough it finds itself in state #3.
#3 works pretty well except when the front of the car is touching the wall. Something strange sometimes happens then. It just stops. I am not sure why.

My model level car has a POT with which I tell the PIC how fast to go. The top POT position is no where near top speed for this particular model. Associated with each of the seven POT positions are three distances: The ideal distance from the wall. The distance at which the car must enter state #2. The distance at which the car must go to state #3.

I have SONARs, PICBASIC PRO, and 16F887 to work with. Can my cars race inside a hockey rink?

Ken

[ScaleRobotics]

I just found a WEB site extolling the 2D IR proximity sensor. Two potential problems - outdoors in bright sunlight and multiple cars. What do you think?

I did not read the specs for IR sensing. They are talking inches to a few feet. I need at least ten feet, probably more.

http://www.societyofrobots.com/senso...pirrange.shtml

Ken

Yeah, that was discussed back on post # 617 of this thread. http://www.picbasic.co.uk/forum/show...6363#post96363

I thought the object was to be within two feet of the wall, but maybe that changed? I was thinking if you are more than two feet from the wall, turn right. So I am not sure that you really need more than 30 inches of sensing for the side view. Front view, yes, you need more like 15 - 20 feet. Using one sonar and one IR, at least you would be able to take both measurements at the same time, giving you twice the data in the same amount of time.

I would think multiple cars with IR sensors would be a lot happier than multiple cars with sonar. The IR is a spot, where as the sonar is pretty much all over the map with echos etc. As for sunlight, this website says "This ranging method is almost immune to interference from ambient light, and offers amazing indifference to the color of the object being detected. In other words, the sensor is capable of detecting a black wall in full sunlight with almost zero noise." But then they go to say "(UPDATE: despite popular belief, it is quite possible for both direct and indirect sunlight to significantly affect results. I learned this the hard way!) " Sounds like some testing is required!


I had another thought about the importance of a wheel speed sensor. If you are measuring vehicle speed with a wheel sensor, you can differentiate the speed that your sonar is picking up to see if the wall is moving toward you at a faster or slower rate than your vehicle is moving. This will tell you if the wall is turning toward you (curved surface), or if you are turning toward a straight wall. This would be helpful for a lot of things. I would say required for better control

I have SONARs, PICBASIC PRO, and 16F887 to work with. Can my cars race inside a hockey rink?

As long as they get enough traction, and don't get checked by any of the hockey players.

[Kenjones1935]

Yes, two feet from the wall is fine when the car is moving at walking speed. At 20mph or more as would be appropriate in a hockey rink I think the car needs more maneuvering room.

The model level cars have rubber tires that can be studded. Since 'indoor' hockey rinks have good enough ventilation to run the Zamboni ice scrapping machine it might be possible to race gas (nitro) driven models in them.

The issue seems to be speed of sensing and accuracy of sensing and using PBP to do the math. I have no idea how to attach a speed measuring device to the wheels of a model car.

Ken