Here is another idea i have seen in robot club i zm in,does take some scripts and a map

Using a map with a grid with furniture on it and using ,sonar and IR and set a path planing to the charger.

I will look for the link and post it,i think it was the big robots club in calif or washington area

It have much more details on it and how it works.

I know EZB has a map,but i never tried it yet.

That's a good thought but my house is so small and cluttered with furniture it might not work for me. Still it would be fun to try. Have you found a source for a good distance sensor?

Hi Fred,

That's interesting. I''l take a look at the link when you find it.

There are many distance sensors out there. I have three different types on my current robot although two of them are only short range (40cm max). Also, ultra sonic doesn't pick up everything and ir doesn't pick up everything but combined they should do, so it may require both sensor types mounted together and some scripting to work out the distance.

I think small and cluttered may actually be better for this method, it would mean having a much larger array of area dimensions for the script to check against but it would be more accurate (I think).

The basic script would be something like; $arealength = $distance1+$distance2+$robotlength $areawidth = $distance3+$distance4+$robotlength If ($arealength >490 and $arealength < 500 and $areawidth > 690 and $areawidth < 700) $area = "nook" ElseIf ($arealength >780 and $arealength < 790 and $areawidth > 390 and $areawidth < 400) $area = "kitchen" ...

and so on for each area. The two values of > and < give a bit of a tolerance for the distances too, although if the sensors are good enough it wouldn't be needed (I would always add in a slight tolerance personally though)

Then something like If($area="kitchen")

commands to navigate from the pinpointed location to the base

ElseIf($area="nook")

commands to navigate from the pinpointed location to the base

...

REX Might be hard with almost any type of design if a house has a lot of clutter in it.

May be the best solution for you is like the roomba design,its a little best then pololu beacon type ,one reason is 360 deg ,with pololu beacon type is kinda close only it has 4 points ,N,S,E,W at 360 deg

Another is high cost,looking at $60 with shipping ,on ebay on ROOMBA looking at near $20 to $40 I bought many and a good test to buy also, plus it has a charger base that only needs a few changers ,and EZB has the scripts too

With a map design you can set it for any size room ,even one room And if furniture is changed can easy fix the MAP,and use any sensorss

WHAT i know best at is sensors ,i said so many times,and for a good navigation or object detection you need both IR and SONARS since each one is good for certain types of material.

ALso the placement is very important as orientation of the sensor

I bought awhile back a optical bench for running all types of tests on all sensors,will be putting the info up soon and try to make it clear for others to understand it.

Yes I think it probably will end up with a good mixture of several types of sensors. Your script looks interesting. I am just getting to the point where I can start experimenting with script. I just got the ping radar working. Next I need to get the IR sensors mounted so I can get the robot to go through doorways. So I still have a ways to go.

Rich have you watched the videos of a robot finding its home beacon I posted earlier? It looks as though the home beacon only needs power to transmit. But it is not entirely clear.

I am very excited about this docking project and we are seeing a lot of cool ideas and suggestions.

Going back to earlier suggestions from McJeff0125 and Alan. McJeff0125 had suggested getting the EZ-B to trigger L.E.D. lights to come on. Alan suggested blink patterns.

I know some would not like to put something on their ceiling or wall but what if each room had a small box mounted on the ceiling for instance. The box would have a simple circuit that flashes a high intensity IR L.E.D. for a fixed number of pulses.

The box in each room has a unique number of flashes that would identify which room the robot is in. Say the kitchen box blinks it's IR LED 3 times pauses for 5 seconds then repeats, the living room box blinks it's IR LED 4 times then repeats and so on for other rooms.

On the robot would be a IR sensor hooked to an EZ-B input. A script is created that monitors this sensor and counts the number of blinks and pauses which tells the EZ-B which room it is in..... Problem to solve here. What can the robot use to trigger the blink box to activate? Assuming you don't want the boxes to be on 24/7. Could the EZ-B use an output to trigger something like a car key fob mounted on board the robot and the fob transmitter activates the boxes?....

OK, say the robot is in the kitchen and it's battery monitor script tells it to go to the charger (which is in the living room). It checks this input to tell it what room it's in. Then it uses Rich's idea of a compass module and determines that it needs to go west (for example) to get to the charger. So now it knows what room and it knows which direction it needs to go. Its radar ping helps it avoid obstacles on the way..... Now here is yet another problem. If you need IR sensors on the front and sides of the robot to help with getting through doorways, will the high intensity IR L.E.D.s on the ceiling give these false triggering?....

Anyway the robot is heading to the living room. When it enters the living room it turns on the IR Beacons from POLOLU and they guide it to the charger like in the video I posted earlier. The robot enters the charging nest and a hall effect (magnetically triggered)switch (suggested by Ray) mounted on the robot detects a magnet mounted on the nest wall. A script tells the robot it is docked and also that the charger is connected because of the higher voltage level. All drive motion is stopped and robot sits there until the script tells it that the battery charge is finished..... Another problem to solve here. How does it know the difference between the charger voltage level or when the battery is full?.... Sorry for the long post

I haven't watched the video yet, I had a quick look at 2am this morning when I got in and that's pretty much it, I will watch it when I have more time though as I'm very interested in that idea too.

As for the problem of detecting if it's charging or if it's not, if it's fully charged etc. that's one that does need looking at. I have a solution for mine but it wouldn't work on all. Mine uses a 12v LiPo balance port charger to charge, so I plan to tap off the green and red LEDs which indicate that it's 1) charging and 2) fully charged. This wouldn't work for other chargers though. But it's something to look at.

I think the best way is to tackle each of the steps one at a time. Step 1 - Knowing it needs charging. Addressed with the monitor circuits and scripts. Step 2 - Finding and navigating to the charger. Lot's of ideas but all need further thought. Step 2.1 - Identifying robot location in relation to charging dock Step 2.2 - Navigating to the charging dock whilst avoiding objects and obstacles. Step 3 - Lining up and docking, setting the charger to on and the robot to charge mode. To be discussed. Step 4 - Knowing it's fully charged and releasing itself from the charger. To be discussed.