There were only 14 people present, though many regular members were there. I added Fred to the list, even though he was not present.  I could not easily find an online tool that would determine the center of a set of zipcodes, so I wrote my own using a combination of a geonames.org api and a google maps api and Python.  I basically find the latitude and longitude of each address or centroid of each zipcode; average them all, ignoring spherical distortion since this is a pretty small patch of the Earth; then convert that back into an address.

(Turns out the google maps api does not give the official census bureau zipcode centroid, so I use geonames’ api when all I have is a zipcode.  If I had a full address, I used the google maps api.)

Anyway, the center is surprisingly central east-west but further south than I expected. It’s within 2 miles of PCC Sylvania and also 2 miles north of the I-5 and Hwy217 interchange.

Also interesting is that even though google map’s idea of the center of each zipcode is different from the official center in most cases, the errors all seem to average out, so the address I get just using google’s api is within 0.1 miles of the more accurate one.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100

#
# center_of_zipcodes.py
#
# find address corresponding to center of a set of zipcodes using google maps api
# written by Pete Skeggs, 4/5/2011, for PARTS
# updated 4/13/2011 to use geonames.org API to get true census bureau centroids when just given a zipcode; also, use an address when known and look that up with the google maps api
#
import httplib
import json
import urllib
 
# define functions
 
def ztoll(z):
   conn = httplib.HTTPConnection("api.geonames.org")
   # note: you need to replace 'demo' below with a valid username; I decided not
   # to share mine
   url = "/postalCodeLookupJSON?postalcode=%s&country=US&username=demo" % z
   conn.request("GET", url)
   r = conn.getresponse()
   o = json.load(r)
   c = {'lat' : 0, 'lng' : 0}
   # format returned:
   #{"postalcodes":[{"adminName2":"Multnomah","adminCode2":"051","postalcode":"97210","adminCode1":"OR","countryCode":"US",
   # "lng":-122.703348,"placeName":"Portland","lat":45.530318,"adminName1":"Oregon"}]}
   c['lat'] = o[u'postalcodes'][0][u'lat']
   c['lng'] = o[u'postalcodes'][0][u'lng']
   c['addr'] = z
   return c
 
def addrtoll(addr):
   conn = httplib.HTTPConnection("maps.googleapis.com")
   url = "/maps/api/geocode/json?address=%s&sensor=false" % urllib.quote_plus(addr)
   conn.request("GET", url)
   r = conn.getresponse()
   o = json.load(r)
   c = {'lat' : 0, 'lng' : 0}
   # format returned:
   # { "status": "OK", "results": [{"types":["postal_code"],"formatted_address":"Portland, OR 97210, USA",
   # "address_components":[{"long_name":"97210","short_name":"97210","types":["postal_code"]},
   # {"long_name":"Portland","short_name":"Portland","types":["locality","political"]},
   # {"long_name":"Oregon","short_name":"OR","types":["administrative_area_level_1","political"]},
   # {"long_name":"UnitedStates","short_name":"US","types":["country","political"]}],
   #"geometry":{"location":{"lat":45.5552753,"lng":-122.7384041},"location_type":"APPROXIMATE",
   #"viewport":{"southwest":{"lat":45.5158719,"lng":-122.7897450},"northeast":{"lat":45.5858230,"lng":-122.6920030}},
   #"bounds":{"southwest":{"lat":45.5158719,"lng":-122.7897450},"northeast":{"lat":45.5858230,"lng":-122.6920030}}}}]}
   c['lat'] = o[u'results'][0][u'geometry'][u'location'][u'lat']
   c['lng'] = o[u'results'][0][u'geometry'][u'location'][u'lng']
   c['addr'] = o[u'results'][0][u'formatted_address']
   return c
 
def lltoz(lat, lng):
   conn = httplib.HTTPConnection("maps.googleapis.com")
   url = "/maps/api/geocode/json?latlng=%f,%f&sensor=false" % (lat, lng)
   conn.request("GET", url)
   r = conn.getresponse()
   o = json.load(r)
   addr = o[u'results'][0][u'formatted_address']
   return addr
 
# define variables
 
# list of zipcodes for PARTS members at April 2011 meeting
addresses = [
"97007",
"97210",
"97223",
"97034",
"97202",
"97232",
"97224",
"97071",
"97219",
"97086",
"97224",
"97211",
"97006",
"97224",
"97225",
"97229"]
lat = 0
lng = 0
count = 0
 
# main code
 
for a in addresses:
   if len(a) == 5:
      c = ztoll(a)
   else:
      c = addrtoll(a)
   print "%d. %s: %f, %f" % (count+1, c['addr'], c['lat'], c['lng'])
   lat = lat + c['lat']
   lng = lng + c['lng']
   count = count + 1
 
lat = lat / count
lng = lng / count
print "average of %d zipcodes: lat %f, lng %f" % (count, lat, lng)
print "address at center: %s" % lltoz(lat, lng)

This yields the following output (using a valid geonames.org username — you need to register for one of your own):

~$ python center_of_zipcodes.py
1. 97007: 45.450489, -122.865171
2. 97210: 45.530318, -122.703348
3. 97223: 45.447390, -122.795294
4. 97034: 45.409263, -122.684721
5. 97202: 45.484007, -122.636534
6. 97232: 45.528712, -122.636310
7. 97224: 45.409448, -122.801400
8. 97071: 45.144617, -122.858342
9. 97219: 45.457956, -122.707380
10. 97086: 45.444600, -122.537200
11. 97224: 45.409448, -122.801400
12. 97211: 45.565259, -122.644815
13. 97006: 45.520130, -122.860376
14. 97224: 45.409448, -122.801400
15. 97225: 45.498473, -122.778659
16. 97229: 45.548317, -122.827561
average of 16 zipcodes: lat 45.453617, lng -122.746244
address at center: 6655 SW Kingsview Ct, Tigard, OR 97223, USA
~$

Scott’s bringing the hot-plate reflow system and I’m bringing the parts.  You bring whatever parts you might have already gotten from me, some magnifying device if needed, a soldering iron (if you have one) and some cash if you haven’t paid for parts yet.

We’ll be at the bRainSilo:
bRainSilo.  7pm.

http://brainsilo.org/wiki/Location

See you there!

Monty

A small group of us have been regularly getting together every other Monday night at the bRainSilo to work on swarm bots.  We’d love to have more people come join us — especially if you built one of the BrushBotComm boards a while back!  We’re at the point where we have bidirectional IR comm code working and want to try it out w/ larger swarms.

This get together is not just about swarm bots though — rather than toiling away alone in the basement on that killer new robot, why not bring it down and work on it in the company of other robot enthusiasts?

7pm, bRainSilo: http://brainsilo.org/wiki/Location

Hope to see you there!

Monty

Have a great Labor Day!

Monty

See you there,

Monty

And if you DO have brushbots, bring them down and let’s get them talking!

7pm, bRainSilo:
http://brainsilo.org/wiki/Location

Hope to see you there!

Monty

7pm, bRainSilo:
http://brainsilo.org/wiki/Location

Monty

While Scott and I hearded bots around, Tim got his steering control of the brushbots (a homemade toothbrush-head based bot in his case) working. The really cool thing is that he’s only using a single vibrating motor, yet still getting steerability! Come on Tim… Hurry up and upload the details on how you’re doing this!

Hope to see you all tonight,

7pm, bRainSilo:
http://brainsilo.org/wiki/Location

Monty

Time: Monday 4/5, 7:00pm
Location: bRainSilo: http://brainsilo.org/wiki/Location

We will be continuing with the BrushBotComm boards — soldering on the final components, including the missing IR LED resistors and the IR
LED’s themselves; making programming cables; and learning how to load code onto them.

As I mentioned in this blog entry the ArdunioISP should work well for a programmer, and hopefully as a serial link for debugging.  So, if you have a spare Arduino that you can use for a programmer, bring it.  Or, I have a handful of cosmetically defective (but fully functional) Teensy2.0′s available for really cheap.

Bring:
*BrushBotComm boards and extra parts you’ve received.
*Soldering iron — if you have a decent one with a fine tip.  This is especially important this meeting as we have a fair amount of hand soldering to do.
*Desk lamp if you can — we didn’t really have enough light last time
*Vision enhancement devices — I’ll bring my stereo-microscope and what magnifiers I have, but I think we could have used more last time
*Money for Arduinos if you need one for programming.
*Laptop if you have one to use for programming.

Monty

I made up the following test setup:  For the receiver side, I took one of the IR receivers that we are using on the comma boards and mounted it in a solderless breadboard to make it easy to try resistor values.  I also hooked up a variable power supply and used an ATMega328 LillyPad Arduino to drive the IR LED.  The LillyPad was the only thing I had handy which would run at 8MHz and low voltage.  The power supply was set to 2.75VDC which is pretty much as low as you can go with a 328 running at 8MHz and is around the voltage that the batteries can put out.  I used a test program from the IRremote library (http://www.arcfn.com/2009/08/multi-protocol-infrared-remote-library.html) to repeatedly send out 0XD10 which is the Sony code for the ENTER key.

On the receiver side, I used a comm board hooked up to a variable power supply to test how the receiver coped with different voltages.   Monty had written a test program for the receiver side which recognizes Sony ENTER and blinks the LEDs in a particular pattern (although the key codes in irrx.h need to all be decremented by one to be correct).  With this program it is possible to move the comm board around and watch the LEDs to see where successful IR transmission occurs.

As a baseline test, I put in a 100 ohm resister into the IR transmitter circuit which gives about 20ma drive to the IR LED.  Using 5VDC for the receiver, the IR transmission works successfully at quite a long distance (probably across a room).

However, lowering the receiver voltage to 2.85VDC or below results in a failure of the IR receiver to work.  It still registers that an IR transmission is occurring but is unable to recognize the code.  Since the batteries put out more like 2.72VDC, this is a big problem.  After discussing it with Monty, we decided to short out R5, the resistor on the Vcc line into the IR receiver to try to get as much voltage into the receiver as possible.  Fortunately, this seems to work and allows the receiver to work down to 2.53VDC.  So I then hooked up the battery board to the comm board and (with R5 shorted) the receiver works.  The rest of the result below are all done with battery power for the receiver.

I then examined how changing the transmittter series resistor affected the range:

1K ohm            18 inches
4.7K ohm        5 inches
10K ohm         2.5 inches

These distances are for on-axis communications.  I’ll do some more experiments to see how the off-axis accuracy breaks down and update the table.