1
00:00:06,320 --> 00:00:11,499
[Music]

2
00:00:16,320 --> 00:00:19,760
welcome back to the open hardware

3
00:00:18,080 --> 00:00:22,080
minicon so next up we have a

4
00:00:19,760 --> 00:00:26,080
presentation about open hardware weather

5
00:00:22,080 --> 00:00:27,920
radar and our speaker is tisham who you

6
00:00:26,080 --> 00:00:29,359
may have seen previously at an open

7
00:00:27,920 --> 00:00:31,599
hardware mini conf he did a very

8
00:00:29,359 --> 00:00:32,399
interesting talk about energy monitoring

9
00:00:31,599 --> 00:00:33,920
and

10
00:00:32,399 --> 00:00:36,239
if my memory is correct tish you were

11
00:00:33,920 --> 00:00:37,280
using fpgas for that project i think is

12
00:00:36,239 --> 00:00:39,520
that correct

13
00:00:37,280 --> 00:00:41,760
i got to a stage where i was like

14
00:00:39,520 --> 00:00:43,840
experimenting with them yeah this has a

15
00:00:41,760 --> 00:00:45,680
fair bit of fpgas in it as well

16
00:00:43,840 --> 00:00:47,280
okay great so

17
00:00:45,680 --> 00:00:49,360
following on from questions after that

18
00:00:47,280 --> 00:00:51,520
last session about what are applications

19
00:00:49,360 --> 00:00:53,840
for fpgas perhaps

20
00:00:51,520 --> 00:00:55,920
this is an example i'm very interested

21
00:00:53,840 --> 00:00:56,960
to see this presentation so tish take it

22
00:00:55,920 --> 00:00:59,840
away

23
00:00:56,960 --> 00:01:03,039
thank you thank you um yeah so my slides

24
00:00:59,840 --> 00:01:05,040
are up uh so i uh

25
00:01:03,039 --> 00:01:07,840
i'm tisham i

26
00:01:05,040 --> 00:01:10,000
build things for fun and this is one of

27
00:01:07,840 --> 00:01:11,280
the things i built and

28
00:01:10,000 --> 00:01:13,360
i'll start

29
00:01:11,280 --> 00:01:15,360
before i introduce myself i will start

30
00:01:13,360 --> 00:01:17,119
with an acknowledgement to the country

31
00:01:15,360 --> 00:01:19,600
i'm presenting today from the land of

32
00:01:17,119 --> 00:01:21,600
the ghana people of the adelaide plains

33
00:01:19,600 --> 00:01:25,119
and i acknowledge their

34
00:01:21,600 --> 00:01:27,920
people the past present and current

35
00:01:25,119 --> 00:01:29,920
and the connection to the land and water

36
00:01:27,920 --> 00:01:32,159
is very crucial to life

37
00:01:29,920 --> 00:01:36,079
and this presentation even though it's

38
00:01:32,159 --> 00:01:38,320
about hardware it's really about water

39
00:01:36,079 --> 00:01:41,439
and we'll see a lot of that going

40
00:01:38,320 --> 00:01:43,439
through and yeah enjoy

41
00:01:41,439 --> 00:01:45,520
so i started

42
00:01:43,439 --> 00:01:47,200
building this project

43
00:01:45,520 --> 00:01:49,759
many years ago there's a little bit of a

44
00:01:47,200 --> 00:01:51,600
timeline here but uh slowly been

45
00:01:49,759 --> 00:01:54,320
chipping away at it and

46
00:01:51,600 --> 00:01:56,479
i'm at a stage where i need help and

47
00:01:54,320 --> 00:01:59,280
this is in addition to what you have

48
00:01:56,479 --> 00:02:01,759
done and learned it's also about getting

49
00:01:59,280 --> 00:02:03,600
this out to the community and getting

50
00:02:01,759 --> 00:02:06,560
interest in hardware

51
00:02:03,600 --> 00:02:07,439
in rf frequencies

52
00:02:06,560 --> 00:02:11,360
bigger

53
00:02:07,439 --> 00:02:13,520
and looking at water that's right

54
00:02:11,360 --> 00:02:17,360
so there's a bit of a flow of the ideas

55
00:02:13,520 --> 00:02:18,800
i my mind works in a very random

56
00:02:17,360 --> 00:02:20,640
branching fashion and i have to

57
00:02:18,800 --> 00:02:22,879
backtrack and linearize it our

58
00:02:20,640 --> 00:02:25,760
presentation by its very nature is very

59
00:02:22,879 --> 00:02:28,560
linear so i've tried to put down all of

60
00:02:25,760 --> 00:02:30,400
the different ideas that are

61
00:02:28,560 --> 00:02:31,360
going to be talked about in a linear

62
00:02:30,400 --> 00:02:33,680
fashion

63
00:02:31,360 --> 00:02:36,239
so i will talk a bit about the history

64
00:02:33,680 --> 00:02:38,480
of weather radars where they come from

65
00:02:36,239 --> 00:02:41,120
it's a bit of a superpower seeing things

66
00:02:38,480 --> 00:02:44,160
in radar uh we don't really have extra

67
00:02:41,120 --> 00:02:46,480
vision or super people these days but uh

68
00:02:44,160 --> 00:02:48,560
radar is the closest you can come to

69
00:02:46,480 --> 00:02:52,480
extravision seeing through walls and

70
00:02:48,560 --> 00:02:55,440
clouds into into the what's behind

71
00:02:52,480 --> 00:02:58,480
uh talk about a bit of ocean in the sky

72
00:02:55,440 --> 00:03:00,080
and how the radar interacts with the

73
00:02:58,480 --> 00:03:01,920
moisture

74
00:03:00,080 --> 00:03:04,319
some stuff about solid state radars and

75
00:03:01,920 --> 00:03:08,319
where things are at in the commercial

76
00:03:04,319 --> 00:03:11,120
industry in this arena a bit of project

77
00:03:08,319 --> 00:03:13,280
timeline of my and my partner gashini's

78
00:03:11,120 --> 00:03:15,280
work on this so i didn't do all this by

79
00:03:13,280 --> 00:03:18,400
myself i had a lot of help

80
00:03:15,280 --> 00:03:20,959
and all of the rest of it is around

81
00:03:18,400 --> 00:03:23,120
how iteratively you try to solve

82
00:03:20,959 --> 00:03:25,040
problems i was watching

83
00:03:23,120 --> 00:03:26,480
john's presentation on dealing with

84
00:03:25,040 --> 00:03:28,080
issues that you come about when you're

85
00:03:26,480 --> 00:03:30,000
trying to make hardware

86
00:03:28,080 --> 00:03:31,920
some of the issues when you're trying to

87
00:03:30,000 --> 00:03:33,680
get a new idea of the ground or even

88
00:03:31,920 --> 00:03:35,440
bigger their business issues and making

89
00:03:33,680 --> 00:03:36,959
people understand that this is actually

90
00:03:35,440 --> 00:03:40,080
worth doing

91
00:03:36,959 --> 00:03:41,440
uh and then there is a funding issues uh

92
00:03:40,080 --> 00:03:43,440
is what i'm referring to and a couple of

93
00:03:41,440 --> 00:03:45,360
technical issues around now casting and

94
00:03:43,440 --> 00:03:47,440
qpe which stands for quantitative

95
00:03:45,360 --> 00:03:49,440
precipitation estimation

96
00:03:47,440 --> 00:03:52,640
and then we get to

97
00:03:49,440 --> 00:03:55,599
the actual idea or the prototypes i have

98
00:03:52,640 --> 00:03:56,959
or been working on for uh phased array

99
00:03:55,599 --> 00:03:58,640
radars

100
00:03:56,959 --> 00:04:01,200
and

101
00:03:58,640 --> 00:04:03,599
and this is not a brand new idea it's

102
00:04:01,200 --> 00:04:06,000
just becoming easier and easier to do

103
00:04:03,599 --> 00:04:07,200
with the advances in availability of

104
00:04:06,000 --> 00:04:10,080
hardware

105
00:04:07,200 --> 00:04:11,920
uh easier for novices and people just

106
00:04:10,080 --> 00:04:14,400
sitting in their living rooms to do i

107
00:04:11,920 --> 00:04:16,799
guess outside a lab

108
00:04:14,400 --> 00:04:19,280
and sky's ahead hewani is the project

109
00:04:16,799 --> 00:04:21,759
name hiwa means that what's in the sky

110
00:04:19,280 --> 00:04:25,199
what's the weather like in swahili

111
00:04:21,759 --> 00:04:26,840
so the sky is ahead for this project and

112
00:04:25,199 --> 00:04:30,639
all of the turbulence that would be

113
00:04:26,840 --> 00:04:33,440
experiencing or been experiencing

114
00:04:30,639 --> 00:04:35,520
so yeah so the there's a fair amount of

115
00:04:33,440 --> 00:04:38,800
slides there's a lot of words uh some

116
00:04:35,520 --> 00:04:41,199
pictures so i'll just go through it and

117
00:04:38,800 --> 00:04:42,560
i will keep an eye on the time

118
00:04:41,199 --> 00:04:44,320
so the

119
00:04:42,560 --> 00:04:46,400
microwave part of the electromagnetic

120
00:04:44,320 --> 00:04:49,040
spectrum is what we're talking about and

121
00:04:46,400 --> 00:04:52,400
it's a finite resource

122
00:04:49,040 --> 00:04:54,320
and we have governments uh

123
00:04:52,400 --> 00:04:56,400
licensing our spectrum

124
00:04:54,320 --> 00:04:58,720
at very very high rates and our modern

125
00:04:56,400 --> 00:04:59,759
civilization like me talking over wi-fi

126
00:04:58,720 --> 00:05:01,919
right now

127
00:04:59,759 --> 00:05:04,960
is tied together by this part of the

128
00:05:01,919 --> 00:05:07,039
spectrum and only very few people work

129
00:05:04,960 --> 00:05:08,800
on it understand it it's pretty black

130
00:05:07,039 --> 00:05:11,280
magic

131
00:05:08,800 --> 00:05:13,759
yeah so that's that's sort of the very

132
00:05:11,280 --> 00:05:16,000
interesting area to work in

133
00:05:13,759 --> 00:05:17,280
a bit of history of how how these things

134
00:05:16,000 --> 00:05:19,919
came about

135
00:05:17,280 --> 00:05:22,479
the radios were invented and then the

136
00:05:19,919 --> 00:05:25,039
radar itself came from

137
00:05:22,479 --> 00:05:27,039
what was a request for a death ray to

138
00:05:25,039 --> 00:05:29,440
spray something into the air and take

139
00:05:27,039 --> 00:05:32,400
down aircraft during the second world

140
00:05:29,440 --> 00:05:34,960
war uh and the uh

141
00:05:32,400 --> 00:05:37,759
the radar installations came about i

142
00:05:34,960 --> 00:05:39,759
became aware in my life journey about

143
00:05:37,759 --> 00:05:41,039
radars through reading this book the

144
00:05:39,759 --> 00:05:43,840
grand slack

145
00:05:41,039 --> 00:05:46,240
by a french free air force pilot pia

146
00:05:43,840 --> 00:05:48,800
customer who later became a politician

147
00:05:46,240 --> 00:05:50,960
who talked about the radar that was

148
00:05:48,800 --> 00:05:53,440
being used in the air force during the

149
00:05:50,960 --> 00:05:56,800
second war

150
00:05:53,440 --> 00:05:59,039
uh so yeah so that from war after the

151
00:05:56,800 --> 00:06:02,720
war ended there was a lot of radars

152
00:05:59,039 --> 00:06:06,400
available and uh there were what was the

153
00:06:02,720 --> 00:06:09,199
clutter in the wartime radar uh became

154
00:06:06,400 --> 00:06:11,039
signal in that this radar is being used

155
00:06:09,199 --> 00:06:12,319
for civilian purposes to observe

156
00:06:11,039 --> 00:06:14,800
precipitation

157
00:06:12,319 --> 00:06:18,479
rain surplus military radars form the

158
00:06:14,800 --> 00:06:20,400
seed of weather observation radars

159
00:06:18,479 --> 00:06:23,199
so there's an ocean in the sky

160
00:06:20,400 --> 00:06:25,360
uh we have so many oceans that my son

161
00:06:23,199 --> 00:06:26,960
reads about you know uh the arctic the

162
00:06:25,360 --> 00:06:29,919
antarctic the

163
00:06:26,960 --> 00:06:32,400
south the north atlantic south atlantic

164
00:06:29,919 --> 00:06:34,080
there is also an ocean in the sky

165
00:06:32,400 --> 00:06:36,240
all of the evaporation from all of the

166
00:06:34,080 --> 00:06:37,120
oceans gathers in the sky and a lot of

167
00:06:36,240 --> 00:06:39,440
our

168
00:06:37,120 --> 00:06:42,720
uh biosphere depends on that ocean in

169
00:06:39,440 --> 00:06:44,639
the sky falling down on the land

170
00:06:42,720 --> 00:06:46,479
and the

171
00:06:44,639 --> 00:06:49,039
the reason that we have things like

172
00:06:46,479 --> 00:06:51,440
wi-fi are due to the weird dams that the

173
00:06:49,039 --> 00:06:54,639
water molecule does when excited with

174
00:06:51,440 --> 00:06:56,479
frequencies at this uh this this yeah

175
00:06:54,639 --> 00:06:58,479
with these frequencies and absorbing it

176
00:06:56,479 --> 00:06:59,680
sucking it up and sometimes reflecting

177
00:06:58,479 --> 00:07:02,880
some of it

178
00:06:59,680 --> 00:07:04,720
and we end up with this unlicensed bands

179
00:07:02,880 --> 00:07:07,680
where no people don't want to use it and

180
00:07:04,720 --> 00:07:10,400
we use it for things like wi-fi and the

181
00:07:07,680 --> 00:07:12,880
ism bands

182
00:07:10,400 --> 00:07:15,759
the benefits of the weather observation

183
00:07:12,880 --> 00:07:18,639
i will go into more detail but it's it's

184
00:07:15,759 --> 00:07:21,360
uh affects both the benefit side that we

185
00:07:18,639 --> 00:07:24,319
grow crops and have drinking water

186
00:07:21,360 --> 00:07:27,360
uh to the risks we have i was living in

187
00:07:24,319 --> 00:07:28,560
canberra and uh thousands of cars were

188
00:07:27,360 --> 00:07:30,560
turned into

189
00:07:28,560 --> 00:07:33,520
what my friend kelvin refers to as golf

190
00:07:30,560 --> 00:07:36,319
balls by the meteors

191
00:07:33,520 --> 00:07:37,440
hydro meteors not that not the space

192
00:07:36,319 --> 00:07:40,639
kind

193
00:07:37,440 --> 00:07:42,400
so uh i have seen this i i learned

194
00:07:40,639 --> 00:07:44,560
french in school so i've seen this

195
00:07:42,400 --> 00:07:46,479
saying in english you know one man's

196
00:07:44,560 --> 00:07:49,280
fish is another man's poison it didn't

197
00:07:46,479 --> 00:07:52,240
really make sense until i learned french

198
00:07:49,280 --> 00:07:53,919
is that poison is actually fish

199
00:07:52,240 --> 00:07:56,560
and the

200
00:07:53,919 --> 00:07:58,960
this is because what i refer to as the

201
00:07:56,560 --> 00:08:01,280
attenuation due to their resonance with

202
00:07:58,960 --> 00:08:03,680
water at these frequencies and this is

203
00:08:01,280 --> 00:08:05,120
what you can use for uh

204
00:08:03,680 --> 00:08:07,840
for the microwave observation of

205
00:08:05,120 --> 00:08:07,840
precipitation

206
00:08:08,560 --> 00:08:13,440
so the given coming from the target

207
00:08:11,199 --> 00:08:15,759
observation radars you have lots of

208
00:08:13,440 --> 00:08:17,360
other types of target observation radars

209
00:08:15,759 --> 00:08:18,560
some of them are marine radars and they

210
00:08:17,360 --> 00:08:20,319
can be

211
00:08:18,560 --> 00:08:21,680
turned into weather radars with

212
00:08:20,319 --> 00:08:23,840
modifications

213
00:08:21,680 --> 00:08:26,240
but you require specialist labs to make

214
00:08:23,840 --> 00:08:28,160
those modifications and uh yeah they're

215
00:08:26,240 --> 00:08:32,959
not quite easy to get you need a whole

216
00:08:28,160 --> 00:08:34,000
truck and it's big so this one is not uh

217
00:08:32,959 --> 00:08:36,000
like a

218
00:08:34,000 --> 00:08:37,839
small low frequency so high frequencies

219
00:08:36,000 --> 00:08:39,120
it depends a big antenna they're quite a

220
00:08:37,839 --> 00:08:40,959
bit set up

221
00:08:39,120 --> 00:08:42,399
but quite quite interesting things you

222
00:08:40,959 --> 00:08:43,839
can do with all of the target

223
00:08:42,399 --> 00:08:46,000
observation right now so that's sort of

224
00:08:43,839 --> 00:08:48,480
history of where we are

225
00:08:46,000 --> 00:08:51,519
sort of moving forward to the last 10 or

226
00:08:48,480 --> 00:08:52,800
20 years we have seen uh

227
00:08:51,519 --> 00:08:54,959
things

228
00:08:52,800 --> 00:08:57,120
become more even more compact gallium

229
00:08:54,959 --> 00:08:59,440
nitride is uh used for the power of the

230
00:08:57,120 --> 00:09:01,040
fire solid state instead of magnetrons

231
00:08:59,440 --> 00:09:03,120
which were used in the earlier radars

232
00:09:01,040 --> 00:09:06,000
and were not very disciplined in their

233
00:09:03,120 --> 00:09:07,760
frequency generation

234
00:09:06,000 --> 00:09:11,040
to have a little list of various

235
00:09:07,760 --> 00:09:13,519
military modern solid-state radars

236
00:09:11,040 --> 00:09:15,680
and then the civilian solid state radars

237
00:09:13,519 --> 00:09:17,040
are available too but they still use a

238
00:09:15,680 --> 00:09:19,200
spinning antenna

239
00:09:17,040 --> 00:09:21,279
uh so i have a few small list of

240
00:09:19,200 --> 00:09:22,720
civilian solid-state radars that are

241
00:09:21,279 --> 00:09:25,120
just been entering the market over the

242
00:09:22,720 --> 00:09:27,040
last five years or so

243
00:09:25,120 --> 00:09:29,040
um yeah

244
00:09:27,040 --> 00:09:30,720
i actually went out to the marina to

245
00:09:29,040 --> 00:09:33,040
take a picture of one

246
00:09:30,720 --> 00:09:34,880
in adelaide my yeah

247
00:09:33,040 --> 00:09:38,160
during a family beach time i'm just

248
00:09:34,880 --> 00:09:40,560
staring at radars at the marina

249
00:09:38,160 --> 00:09:43,279
so yeah so this is a project timeline of

250
00:09:40,560 --> 00:09:45,519
me being becoming involved in this

251
00:09:43,279 --> 00:09:48,839
project working through various

252
00:09:45,519 --> 00:09:51,440
challenges deploying hardware in on the

253
00:09:48,839 --> 00:09:53,519
field uh doing our first presentation on

254
00:09:51,440 --> 00:09:55,680
it so i started working on it uh with my

255
00:09:53,519 --> 00:09:57,680
friend kashini around 2018

256
00:09:55,680 --> 00:10:01,360
i visited kenya we'll talk about that in

257
00:09:57,680 --> 00:10:03,200
a minute and uh we deployed sensors uh

258
00:10:01,360 --> 00:10:05,839
we sort of i give a first public

259
00:10:03,200 --> 00:10:08,000
presentation about it in baikonu mostly

260
00:10:05,839 --> 00:10:09,760
on the devops and the high software side

261
00:10:08,000 --> 00:10:12,880
of it and the infrastructure side of it

262
00:10:09,760 --> 00:10:14,880
of getting the data together and

263
00:10:12,880 --> 00:10:16,160
sort of disseminating it

264
00:10:14,880 --> 00:10:17,519
over

265
00:10:16,160 --> 00:10:20,560
nairobi

266
00:10:17,519 --> 00:10:22,000
uh so nairobi hasn't had a weather radar

267
00:10:20,560 --> 00:10:25,440
in the last

268
00:10:22,000 --> 00:10:27,760
30 or so years before that in a place

269
00:10:25,440 --> 00:10:29,120
called current there was a big radar

270
00:10:27,760 --> 00:10:31,360
installation

271
00:10:29,120 --> 00:10:33,279
and it became such that there was no

272
00:10:31,360 --> 00:10:35,519
expertise obviously in maintaining this

273
00:10:33,279 --> 00:10:38,480
a big piece of hardware and it became

274
00:10:35,519 --> 00:10:40,480
unmaintained and there the government

275
00:10:38,480 --> 00:10:42,959
has been trying to do

276
00:10:40,480 --> 00:10:45,760
a new budgetary required request to set

277
00:10:42,959 --> 00:10:47,920
up a new one and they refund to do a one

278
00:10:45,760 --> 00:10:50,480
big lumpy one you end up with millions

279
00:10:47,920 --> 00:10:52,800
of dollars and uh sub-saharan country

280
00:10:50,480 --> 00:10:54,800
just doesn't have the room to pull it

281
00:10:52,800 --> 00:10:56,959
off

282
00:10:54,800 --> 00:10:59,200
uh so yeah so that's sort of we saw a

283
00:10:56,959 --> 00:11:01,920
need and we ideated we deployed stuff we

284
00:10:59,200 --> 00:11:04,160
talked about it i talked about it um

285
00:11:01,920 --> 00:11:05,760
vishing is a very private person uh he

286
00:11:04,160 --> 00:11:06,800
doesn't like talking about these things

287
00:11:05,760 --> 00:11:08,320
too much

288
00:11:06,800 --> 00:11:10,480
uh the

289
00:11:08,320 --> 00:11:12,399
i've been trying to stabilize stuff over

290
00:11:10,480 --> 00:11:15,120
the last year i've taken various breaks

291
00:11:12,399 --> 00:11:17,839
between jobs and written a bunch of code

292
00:11:15,120 --> 00:11:20,399
to improve the stuff from how it was run

293
00:11:17,839 --> 00:11:22,560
some serverless now i'm sort of working

294
00:11:20,399 --> 00:11:24,000
again on the hardware side

295
00:11:22,560 --> 00:11:26,320
to create

296
00:11:24,000 --> 00:11:28,640
new versions which are cheaper easier to

297
00:11:26,320 --> 00:11:30,800
deploy require less maintenance

298
00:11:28,640 --> 00:11:31,760
are more open compared to what we have

299
00:11:30,800 --> 00:11:34,640
now

300
00:11:31,760 --> 00:11:36,880
uh and sort of uh maybe in a few years

301
00:11:34,640 --> 00:11:39,200
maybe hardware is hard

302
00:11:36,880 --> 00:11:40,720
someone will pay us for it or

303
00:11:39,200 --> 00:11:42,560
there will be a big market where

304
00:11:40,720 --> 00:11:43,839
everyone is making them and we can buy

305
00:11:42,560 --> 00:11:46,399
them cheaper than we can make them

306
00:11:43,839 --> 00:11:48,640
ourselves which is sort of even though

307
00:11:46,399 --> 00:11:51,519
the gray market for arduinos is huge

308
00:11:48,640 --> 00:11:53,839
this day is in aliexpress

309
00:11:51,519 --> 00:11:56,399
it's great for people in other parts of

310
00:11:53,839 --> 00:11:58,480
the world like say india or kenya

311
00:11:56,399 --> 00:12:00,639
is a financially really good

312
00:11:58,480 --> 00:12:03,360
option for a cheap hardware to become

313
00:12:00,639 --> 00:12:04,880
available that people can learn outside

314
00:12:03,360 --> 00:12:08,399
outside

315
00:12:04,880 --> 00:12:08,399
of what they can afford yeah

316
00:12:08,560 --> 00:12:14,079
so uh so this is sort of the loop

317
00:12:11,440 --> 00:12:17,200
again i'm reminded of john's talk of uh

318
00:12:14,079 --> 00:12:19,440
prototyping trying something out

319
00:12:17,200 --> 00:12:21,360
sharing about it getting some learning

320
00:12:19,440 --> 00:12:24,240
new things getting feedback

321
00:12:21,360 --> 00:12:25,040
refining your ideas and trying going

322
00:12:24,240 --> 00:12:27,200
again

323
00:12:25,040 --> 00:12:30,000
just keep doing it the cost of iterating

324
00:12:27,200 --> 00:12:30,880
on hardware is higher time costs are

325
00:12:30,000 --> 00:12:33,200
higher

326
00:12:30,880 --> 00:12:35,200
it's a slow cycle so you make small

327
00:12:33,200 --> 00:12:37,120
changes and you put on

328
00:12:35,200 --> 00:12:39,040
yeah instead of getting a new stencil

329
00:12:37,120 --> 00:12:41,040
made you put on captain tape where you

330
00:12:39,040 --> 00:12:42,880
need to hack things together have

331
00:12:41,040 --> 00:12:45,600
botched wires

332
00:12:42,880 --> 00:12:48,079
it's a it's how hardware is done uh

333
00:12:45,600 --> 00:12:50,160
it's not pretty i was iterating on my

334
00:12:48,079 --> 00:12:51,920
background this morning to get it in the

335
00:12:50,160 --> 00:12:54,000
good white balance it's a real world

336
00:12:51,920 --> 00:12:55,680
changing the real world is hard

337
00:12:54,000 --> 00:12:58,639
so i have to run off and get some you

338
00:12:55,680 --> 00:13:00,320
know uh dark paper to fix it and i could

339
00:12:58,639 --> 00:13:03,360
have done two in software but uh i'm a

340
00:13:00,320 --> 00:13:03,360
hardware person i guess

341
00:13:03,839 --> 00:13:08,079
so yeah so one of the challenges we had

342
00:13:05,600 --> 00:13:09,519
was uh getting started bootstrapping and

343
00:13:08,079 --> 00:13:11,519
getting a little bit of money to get

344
00:13:09,519 --> 00:13:13,600
started uh so

345
00:13:11,519 --> 00:13:15,760
the top picture is me and my friend

346
00:13:13,600 --> 00:13:16,880
kishini and my tattoo which are like

347
00:13:15,760 --> 00:13:19,600
local

348
00:13:16,880 --> 00:13:21,680
small minivans typically

349
00:13:19,600 --> 00:13:24,000
toyota toyotas

350
00:13:21,680 --> 00:13:26,800
uh that take people around at the back

351
00:13:24,000 --> 00:13:28,800
of it we are going going to some place

352
00:13:26,800 --> 00:13:30,880
called lamu which is like a few

353
00:13:28,800 --> 00:13:33,040
kilometers from the somali border and is

354
00:13:30,880 --> 00:13:34,560
famous for kidnapping tourists but

355
00:13:33,040 --> 00:13:36,800
anyway

356
00:13:34,560 --> 00:13:39,199
and then we sort of put this idea

357
00:13:36,800 --> 00:13:42,079
together so yushini is super smart he

358
00:13:39,199 --> 00:13:43,199
went to mit he got into some got some

359
00:13:42,079 --> 00:13:46,480
contacts

360
00:13:43,199 --> 00:13:50,160
uh the major of them was greg greg

361
00:13:46,480 --> 00:13:52,399
runs the microwave sensing course at mit

362
00:13:50,160 --> 00:13:56,480
i've got a few props here so one of the

363
00:13:52,399 --> 00:13:59,279
courses that greg runs uses containers

364
00:13:56,480 --> 00:14:01,519
to build a small short range

365
00:13:59,279 --> 00:14:03,440
radar which you can generate images from

366
00:14:01,519 --> 00:14:05,360
using synthetic apertures you should

367
00:14:03,440 --> 00:14:07,920
check out it's an open course you should

368
00:14:05,360 --> 00:14:10,880
check out that stuff that greg has

369
00:14:07,920 --> 00:14:13,440
um yeah and he advised us

370
00:14:10,880 --> 00:14:14,240
we needed some money to build some stuff

371
00:14:13,440 --> 00:14:17,440
so

372
00:14:14,240 --> 00:14:20,959
ray stata who is the founder uh and

373
00:14:17,440 --> 00:14:24,560
for a while the ceo of analog devices

374
00:14:20,959 --> 00:14:26,000
uh a very high-speed analog sensing uh

375
00:14:24,560 --> 00:14:28,160
instruments maker

376
00:14:26,000 --> 00:14:30,160
he talked to us said what are you

377
00:14:28,160 --> 00:14:32,320
planning to do in three years we wrote

378
00:14:30,160 --> 00:14:33,600
down on one page he spoke to us for five

379
00:14:32,320 --> 00:14:35,360
minutes

380
00:14:33,600 --> 00:14:38,000
and he gave us

381
00:14:35,360 --> 00:14:40,720
a bit of money to do do some crazy

382
00:14:38,000 --> 00:14:42,399
things try out some crazy things

383
00:14:40,720 --> 00:14:44,639
um so yeah so that was how we got

384
00:14:42,399 --> 00:14:48,000
started on this idea

385
00:14:44,639 --> 00:14:51,920
uh so the upsides that we did sell to

386
00:14:48,000 --> 00:14:55,440
people was uh you know the rain is uh

387
00:14:51,920 --> 00:14:57,519
uh is i is vital for how we grow food

388
00:14:55,440 --> 00:14:58,880
the upsides of knowing predicting when

389
00:14:57,519 --> 00:15:00,639
it will be

390
00:14:58,880 --> 00:15:02,399
are you going to get decent amounts of

391
00:15:00,639 --> 00:15:04,399
it so you can see the graph of the

392
00:15:02,399 --> 00:15:07,120
rainfall versus food production here due

393
00:15:04,399 --> 00:15:09,199
to rainfall agriculture

394
00:15:07,120 --> 00:15:11,440
and then on the flip side the damages

395
00:15:09,199 --> 00:15:12,560
caused by unexpected precipitation

396
00:15:11,440 --> 00:15:14,560
events

397
00:15:12,560 --> 00:15:16,480
are there there's a very good site from

398
00:15:14,560 --> 00:15:18,320
noaa called billions

399
00:15:16,480 --> 00:15:20,720
and you can see billions of dollars of

400
00:15:18,320 --> 00:15:26,079
losses due to weather-related events

401
00:15:20,720 --> 00:15:28,959
flooding hail that the us experiences

402
00:15:26,079 --> 00:15:30,480
yeah so there is a good case for doing

403
00:15:28,959 --> 00:15:32,959
it we can't really control the weather

404
00:15:30,480 --> 00:15:34,639
we can observe and move people to safety

405
00:15:32,959 --> 00:15:37,519
uh so that's one of the things you can

406
00:15:34,639 --> 00:15:40,800
do or you can estimate insurance losses

407
00:15:37,519 --> 00:15:42,800
or risk of looking forward insurance

408
00:15:40,800 --> 00:15:44,560
works on the principle of the insurance

409
00:15:42,800 --> 00:15:48,880
industry knowing the risks slightly

410
00:15:44,560 --> 00:15:50,720
better than the people being insured and

411
00:15:48,880 --> 00:15:53,279
benefiting of that gap

412
00:15:50,720 --> 00:15:56,240
um yeah so that's sort of the benefits

413
00:15:53,279 --> 00:15:59,040
of better precipitation observation

414
00:15:56,240 --> 00:16:01,759
so that's the first part of getting

415
00:15:59,040 --> 00:16:03,360
going and getting some funding in and

416
00:16:01,759 --> 00:16:05,440
convincing people that this is worth

417
00:16:03,360 --> 00:16:07,920
doing maybe uh

418
00:16:05,440 --> 00:16:10,320
it's already being done by states like

419
00:16:07,920 --> 00:16:11,199
the nation states uh in the western

420
00:16:10,320 --> 00:16:12,639
world

421
00:16:11,199 --> 00:16:14,560
it's just uh

422
00:16:12,639 --> 00:16:16,399
it's worth doing for a place like africa

423
00:16:14,560 --> 00:16:19,839
as well

424
00:16:16,399 --> 00:16:21,360
so the next challenge which i guess

425
00:16:19,839 --> 00:16:23,120
came as

426
00:16:21,360 --> 00:16:25,199
creating something cool

427
00:16:23,120 --> 00:16:27,920
the project started we got the idea

428
00:16:25,199 --> 00:16:29,680
initially when i was like watching

429
00:16:27,920 --> 00:16:32,399
soccer it's uh premier league is pretty

430
00:16:29,680 --> 00:16:33,440
big and in kenya i was watching soccer

431
00:16:32,399 --> 00:16:35,680
at a pub

432
00:16:33,440 --> 00:16:38,160
and whenever the rain would come the

433
00:16:35,680 --> 00:16:40,399
satellite broadcast would like fade away

434
00:16:38,160 --> 00:16:42,240
due to rain fade obviously and i was

435
00:16:40,399 --> 00:16:44,480
like oh it's gonna i know where the

436
00:16:42,240 --> 00:16:46,240
satellite is it's probably gonna rain in

437
00:16:44,480 --> 00:16:48,160
five minutes and other people watching

438
00:16:46,240 --> 00:16:50,480
that thing with me at the bar were like

439
00:16:48,160 --> 00:16:52,000
how are you doing that

440
00:16:50,480 --> 00:16:54,000
so that sort of

441
00:16:52,000 --> 00:16:55,920
that sort of five minute prediction is

442
00:16:54,000 --> 00:16:57,440
easier but uh if you want to do 90

443
00:16:55,920 --> 00:16:58,800
minutes and so you'll take years of

444
00:16:57,440 --> 00:17:00,079
research to do it so first of the

445
00:16:58,800 --> 00:17:02,320
challenge one of the challenges we

446
00:17:00,079 --> 00:17:05,600
looked at was now casting

447
00:17:02,320 --> 00:17:08,480
and that requires a large sampling

448
00:17:05,600 --> 00:17:10,240
window and uh creating a grid of this

449
00:17:08,480 --> 00:17:12,880
sensors

450
00:17:10,240 --> 00:17:15,679
uh i have in here uh radar data

451
00:17:12,880 --> 00:17:17,839
simulation and now casting paper from

452
00:17:15,679 --> 00:17:19,439
one of the many research papers going

453
00:17:17,839 --> 00:17:21,520
around this one in particular one is in

454
00:17:19,439 --> 00:17:24,319
catalonia

455
00:17:21,520 --> 00:17:26,799
uh so we got to some stage we got to uh

456
00:17:24,319 --> 00:17:28,559
sort of 20 minute window now casting

457
00:17:26,799 --> 00:17:32,160
approach i'm just gonna talk a little

458
00:17:28,559 --> 00:17:34,720
bit more about the the way it works in

459
00:17:32,160 --> 00:17:36,799
uh in a properly deployed scenario so

460
00:17:34,720 --> 00:17:39,360
you need typically you lean big lumpy

461
00:17:36,799 --> 00:17:40,160
radars like uh the bureau of meteorology

462
00:17:39,360 --> 00:17:42,080
has

463
00:17:40,160 --> 00:17:44,080
in various different wavelengths each

464
00:17:42,080 --> 00:17:46,480
with their own advantages

465
00:17:44,080 --> 00:17:48,640
uh bureau of met in australia has things

466
00:17:46,480 --> 00:17:50,320
with up to 200 kilometers in coverage

467
00:17:48,640 --> 00:17:52,480
after that the curvature of the earth

468
00:17:50,320 --> 00:17:54,559
again as we will discuss um you know

469
00:17:52,480 --> 00:17:57,280
just a slight couple of slides down uh

470
00:17:54,559 --> 00:17:58,960
starts affecting the big lumpy radar

471
00:17:57,280 --> 00:18:02,000
uh you can also have what i call the pie

472
00:17:58,960 --> 00:18:04,480
in the sky very high high risk

473
00:18:02,000 --> 00:18:06,559
systems which can blow up regularly so

474
00:18:04,480 --> 00:18:08,480
one of the some of these meteorological

475
00:18:06,559 --> 00:18:10,000
satellites have blown up while being

476
00:18:08,480 --> 00:18:12,640
launched

477
00:18:10,000 --> 00:18:14,080
so you put you make a baby for years and

478
00:18:12,640 --> 00:18:15,760
it costs millions and then you put it on

479
00:18:14,080 --> 00:18:17,760
a firecracker and you try to send it to

480
00:18:15,760 --> 00:18:20,160
space it's it's a high risk venture but

481
00:18:17,760 --> 00:18:23,360
no it's still doing it

482
00:18:20,160 --> 00:18:26,080
and then uh sort of uh the or the whole

483
00:18:23,360 --> 00:18:27,440
big picture uh requires some of this for

484
00:18:26,080 --> 00:18:29,120
stitching the things you're seeing on

485
00:18:27,440 --> 00:18:30,559
the ground it gives us a good big

486
00:18:29,120 --> 00:18:32,400
picture of you

487
00:18:30,559 --> 00:18:34,240
uh this is sort of the history of the

488
00:18:32,400 --> 00:18:35,840
various meteorological satellites that

489
00:18:34,240 --> 00:18:38,000
are going have been up

490
00:18:35,840 --> 00:18:40,720
my i love this last one it's a very uh

491
00:18:38,000 --> 00:18:43,600
the severi uh has a spinning

492
00:18:40,720 --> 00:18:47,039
uh approach with a mirror for observing

493
00:18:43,600 --> 00:18:48,880
the hemisphere getting a disc image uh

494
00:18:47,039 --> 00:18:51,200
there is an animation on the umed sat

495
00:18:48,880 --> 00:18:52,960
website of how severe composites this

496
00:18:51,200 --> 00:18:56,080
image is really really cool it's really

497
00:18:52,960 --> 00:18:58,160
cool with a spinning mirror

498
00:18:56,080 --> 00:18:59,919
and then you can have this sort of

499
00:18:58,160 --> 00:19:03,120
honeycomb approach which is what we

500
00:18:59,919 --> 00:19:06,240
aimed for in our project uh in putting

501
00:19:03,120 --> 00:19:09,840
together cell of uh networks depending

502
00:19:06,240 --> 00:19:11,840
on density depending on the

503
00:19:09,840 --> 00:19:14,080
importance i guess from a human

504
00:19:11,840 --> 00:19:17,039
perspective so you don't want to have

505
00:19:14,080 --> 00:19:19,520
lots and lots of dense expand radars in

506
00:19:17,039 --> 00:19:22,400
an area where there are not many people

507
00:19:19,520 --> 00:19:24,160
a 200 kilometer one would be fine so

508
00:19:22,400 --> 00:19:25,760
this sort of approach is quite common in

509
00:19:24,160 --> 00:19:27,360
weather simulations there is one of

510
00:19:25,760 --> 00:19:28,400
those models i was looking at called

511
00:19:27,360 --> 00:19:30,480
empass

512
00:19:28,400 --> 00:19:32,480
uses a multi-grid approach rather than

513
00:19:30,480 --> 00:19:35,039
an even grid it uses a multiple

514
00:19:32,480 --> 00:19:36,799
different density grids to create create

515
00:19:35,039 --> 00:19:39,280
a

516
00:19:36,799 --> 00:19:40,720
lower data volume but still quite

517
00:19:39,280 --> 00:19:42,960
accurate with the model with a

518
00:19:40,720 --> 00:19:45,760
simulation model

519
00:19:42,960 --> 00:19:47,679
uh so many many years ago uh well i was

520
00:19:45,760 --> 00:19:50,400
still doing my phd one of the other

521
00:19:47,679 --> 00:19:51,360
projects also being run out of my lab by

522
00:19:50,400 --> 00:19:53,760
my

523
00:19:51,360 --> 00:19:55,120
pg supervisor doug gray was this thing

524
00:19:53,760 --> 00:19:56,640
called the ren

525
00:19:55,120 --> 00:19:58,880
uh so this was run out of the adelaide

526
00:19:56,640 --> 00:20:00,480
university uh i'm not quite sure where

527
00:19:58,880 --> 00:20:01,919
it is at i've been out of the lab for

528
00:20:00,480 --> 00:20:04,320
many years now

529
00:20:01,919 --> 00:20:07,200
but it sort of also inspired the work we

530
00:20:04,320 --> 00:20:09,280
did was to wear the radar experimental

531
00:20:07,200 --> 00:20:11,200
network which was essentially the same

532
00:20:09,280 --> 00:20:13,120
idea taking small radars and then

533
00:20:11,200 --> 00:20:16,080
joining them together into a into a

534
00:20:13,120 --> 00:20:17,919
network to get rid of what the what

535
00:20:16,080 --> 00:20:19,840
bureau of metrology talks about the

536
00:20:17,919 --> 00:20:20,799
earth curvature effect of the big lumpy

537
00:20:19,840 --> 00:20:22,400
radar

538
00:20:20,799 --> 00:20:24,880
so it gives you better microclimate

539
00:20:22,400 --> 00:20:26,080
estimations higher resolutions temporal

540
00:20:24,880 --> 00:20:28,480
spatial

541
00:20:26,080 --> 00:20:31,440
and uh yeah takes care of takes care of

542
00:20:28,480 --> 00:20:34,640
the curvature the earth is not flat

543
00:20:31,440 --> 00:20:37,440
uh the rend project was supported by

544
00:20:34,640 --> 00:20:38,799
raytheon's phased array radar

545
00:20:37,440 --> 00:20:39,919
which is not again commercially

546
00:20:38,799 --> 00:20:41,760
available

547
00:20:39,919 --> 00:20:42,640
uh so yeah so that

548
00:20:41,760 --> 00:20:45,919
again

549
00:20:42,640 --> 00:20:49,120
my attempt at this ongoing is to develop

550
00:20:45,919 --> 00:20:51,679
something that is not a military system

551
00:20:49,120 --> 00:20:54,640
and is fairly open so that we can

552
00:20:51,679 --> 00:20:54,640
create more of them

553
00:20:55,440 --> 00:21:01,200
uh so the other challenge is uh what i

554
00:20:58,000 --> 00:21:03,760
had in my uh stream as a qpe

555
00:21:01,200 --> 00:21:07,600
quantitative precipitation estimation

556
00:21:03,760 --> 00:21:10,480
uh the qp challenge for is addressed in

557
00:21:07,600 --> 00:21:12,559
say i have a paper this uh this as a

558
00:21:10,480 --> 00:21:14,559
reference paper relatively recent a year

559
00:21:12,559 --> 00:21:15,520
or so ago from taiwan

560
00:21:14,559 --> 00:21:18,720
um

561
00:21:15,520 --> 00:21:20,720
is a you need lots of ground truth

562
00:21:18,720 --> 00:21:23,600
because the rain moves around by the

563
00:21:20,720 --> 00:21:26,320
wind uh and where you observe it in in

564
00:21:23,600 --> 00:21:28,960
the space may not be where it lands

565
00:21:26,320 --> 00:21:31,919
uh but the source data available to

566
00:21:28,960 --> 00:21:34,000
perform qpe over africa is very sparse

567
00:21:31,919 --> 00:21:35,600
and they don't agree with each other

568
00:21:34,000 --> 00:21:37,520
there are various data sources that

569
00:21:35,600 --> 00:21:39,679
tahmo temps at

570
00:21:37,520 --> 00:21:41,280
the a couple of weather models guests

571
00:21:39,679 --> 00:21:43,360
ecmwf

572
00:21:41,280 --> 00:21:46,320
their resolutions are very different to

573
00:21:43,360 --> 00:21:47,280
the sensor that we are using

574
00:21:46,320 --> 00:21:49,200
and

575
00:21:47,280 --> 00:21:51,679
the inversion problem

576
00:21:49,200 --> 00:21:54,159
becomes interactable without good data

577
00:21:51,679 --> 00:21:57,120
to tie to the taiwan paper which was

578
00:21:54,159 --> 00:21:57,840
fairly successful uh is

579
00:21:57,120 --> 00:22:00,720
uh

580
00:21:57,840 --> 00:22:02,880
uses a network of thousand stations

581
00:22:00,720 --> 00:22:04,480
uh for automated weather stations so

582
00:22:02,880 --> 00:22:06,080
there is another piece of hardware that

583
00:22:04,480 --> 00:22:07,840
needs to be built and

584
00:22:06,080 --> 00:22:09,760
made much more commodity which is

585
00:22:07,840 --> 00:22:12,080
automated weather stations

586
00:22:09,760 --> 00:22:14,159
um um this is an open hardware mini

587
00:22:12,080 --> 00:22:15,679
console it's another another prop for

588
00:22:14,159 --> 00:22:18,960
another piece of hardware that needs to

589
00:22:15,679 --> 00:22:21,280
exist to solve this problem better

590
00:22:18,960 --> 00:22:23,280
the another fun fact about taiwan

591
00:22:21,280 --> 00:22:25,039
apparently part of the

592
00:22:23,280 --> 00:22:27,360
silicon supply chain issue is due to

593
00:22:25,039 --> 00:22:29,840
reduced rain in the mountains of taiwan

594
00:22:27,360 --> 00:22:30,880
reducing supply of fresh water

595
00:22:29,840 --> 00:22:32,880
so

596
00:22:30,880 --> 00:22:34,799
if you are not seeing chip shortages

597
00:22:32,880 --> 00:22:37,600
there is the rain is causing it less

598
00:22:34,799 --> 00:22:37,600
rain is causing

599
00:22:38,159 --> 00:22:44,640
uh so the the the

600
00:22:41,120 --> 00:22:46,720
in uh absence of the ground rules so one

601
00:22:44,640 --> 00:22:48,880
of the things i looked at is and i would

602
00:22:46,720 --> 00:22:51,520
like to introduce here is this concept

603
00:22:48,880 --> 00:22:53,360
of precipitation air rcs

604
00:22:51,520 --> 00:22:56,400
oh that's a duplication

605
00:22:53,360 --> 00:22:58,320
radar rcs it's just rcs

606
00:22:56,400 --> 00:23:00,720
radar cross section

607
00:22:58,320 --> 00:23:03,280
i have to fix up the slide but anyway

608
00:23:00,720 --> 00:23:04,159
the teardrop shape that you have in all

609
00:23:03,280 --> 00:23:06,640
of the

610
00:23:04,159 --> 00:23:09,440
emojis of rain is not true the rain

611
00:23:06,640 --> 00:23:11,440
doesn't form a teardrop it forms this

612
00:23:09,440 --> 00:23:12,640
thing which is like an inverted goblet

613
00:23:11,440 --> 00:23:14,720
which comes to

614
00:23:12,640 --> 00:23:16,960
closest to the teardrop shape

615
00:23:14,720 --> 00:23:18,880
uh when it gets pretty big but it never

616
00:23:16,960 --> 00:23:20,960
is a teardrop it looks like a flattened

617
00:23:18,880 --> 00:23:22,720
ball it looks more like an orange when

618
00:23:20,960 --> 00:23:25,120
it's falling down due to the pressure of

619
00:23:22,720 --> 00:23:26,960
the air on on the raindrop only when

620
00:23:25,120 --> 00:23:29,280
it's very small it's a sphere because

621
00:23:26,960 --> 00:23:31,440
the surface tension of the water but as

622
00:23:29,280 --> 00:23:33,520
it as it gets bigger and due to the wind

623
00:23:31,440 --> 00:23:36,400
velocity it starts distorting

624
00:23:33,520 --> 00:23:38,400
the its shape is never a teardrop which

625
00:23:36,400 --> 00:23:40,960
is quite interesting

626
00:23:38,400 --> 00:23:42,799
um and then the the other type of rcs

627
00:23:40,960 --> 00:23:46,320
you have to get is for the hailstones i

628
00:23:42,799 --> 00:23:48,240
was discussing about in canberra uh the

629
00:23:46,320 --> 00:23:50,400
this is a pretty funky one i like it

630
00:23:48,240 --> 00:23:52,480
because it looks almost like a virus

631
00:23:50,400 --> 00:23:54,559
this one is created by by the super

632
00:23:52,480 --> 00:23:57,440
cooled water spinning and pushing off

633
00:23:54,559 --> 00:24:00,400
little cones as it spins and those cones

634
00:23:57,440 --> 00:24:04,159
freeze into place and uh

635
00:24:00,400 --> 00:24:06,320
but usually it's like a water ice mixed

636
00:24:04,159 --> 00:24:08,400
with sometimes air bubbles in it so

637
00:24:06,320 --> 00:24:10,480
you're creating the dielectric for

638
00:24:08,400 --> 00:24:12,559
hydrometeors is quite hard

639
00:24:10,480 --> 00:24:14,320
um anyway so it's a it's another

640
00:24:12,559 --> 00:24:17,760
interesting precipitation if you want to

641
00:24:14,320 --> 00:24:22,000
compute rcs uh using simulations instead

642
00:24:17,760 --> 00:24:24,240
of using uh ground station data

643
00:24:22,000 --> 00:24:26,159
and the other one type of rcs

644
00:24:24,240 --> 00:24:29,039
precipitation form is fox which you can

645
00:24:26,159 --> 00:24:30,559
just use small spheres uh to calculate

646
00:24:29,039 --> 00:24:33,279
i've been experimenting with another

647
00:24:30,559 --> 00:24:34,799
open source software project called meep

648
00:24:33,279 --> 00:24:37,840
from mit also

649
00:24:34,799 --> 00:24:40,720
which looks at this uh quite cool uh it

650
00:24:37,840 --> 00:24:42,960
has a python api you can create objects

651
00:24:40,720 --> 00:24:44,480
and then throw radar at it to see what

652
00:24:42,960 --> 00:24:46,240
the backscatter looks like just click

653
00:24:44,480 --> 00:24:48,880
cool

654
00:24:46,240 --> 00:24:51,679
and this is my favorite good looking one

655
00:24:48,880 --> 00:24:54,000
uh that i have added here a sample for a

656
00:24:51,679 --> 00:24:57,039
concocts uh

657
00:24:54,000 --> 00:24:59,200
fractal based uh

658
00:24:57,039 --> 00:25:01,120
notebook for creating ice crystals that

659
00:24:59,200 --> 00:25:03,279
grop learning has it's very cool i had a

660
00:25:01,120 --> 00:25:06,640
lot of fun playing around with it

661
00:25:03,279 --> 00:25:08,240
uh yeah so sometimes it's just fun uh

662
00:25:06,640 --> 00:25:11,840
doing some of these things and looking

663
00:25:08,240 --> 00:25:14,480
at the shapes that water forms

664
00:25:11,840 --> 00:25:17,679
so this is the open hardware radar

665
00:25:14,480 --> 00:25:20,400
presentation i'm fairly into my time

666
00:25:17,679 --> 00:25:23,600
but i will i'll go over it

667
00:25:20,400 --> 00:25:25,600
i have bits of kit to show

668
00:25:23,600 --> 00:25:27,679
so the idea is essentially instead of

669
00:25:25,600 --> 00:25:30,480
having a big antenna that spins like we

670
00:25:27,679 --> 00:25:31,520
saw in the other one of the historical

671
00:25:30,480 --> 00:25:34,080
notes

672
00:25:31,520 --> 00:25:36,640
you create a panel with lots of antennas

673
00:25:34,080 --> 00:25:39,360
in it and you can steer in one direction

674
00:25:36,640 --> 00:25:41,760
not physically but electronically

675
00:25:39,360 --> 00:25:44,559
so it requires a bunch of software

676
00:25:41,760 --> 00:25:46,320
defined radios frequency doublers

677
00:25:44,559 --> 00:25:48,320
in my case because i only have like five

678
00:25:46,320 --> 00:25:51,679
six gigahertz radios

679
00:25:48,320 --> 00:25:53,600
uh power amplifiers antenna arrays those

680
00:25:51,679 --> 00:25:56,480
are the basic building blocks there's a

681
00:25:53,600 --> 00:25:58,720
nice write up this image is from toshiba

682
00:25:56,480 --> 00:26:00,880
research on this again as i said most of

683
00:25:58,720 --> 00:26:02,640
this stuff happens in the lab and it's i

684
00:26:00,880 --> 00:26:04,880
would love to see it moving towards open

685
00:26:02,640 --> 00:26:06,960
hardware community and becoming much

686
00:26:04,880 --> 00:26:08,480
more commonplace

687
00:26:06,960 --> 00:26:10,640
so these are the couple of things i've

688
00:26:08,480 --> 00:26:12,320
tried there's an fpga here an ultra

689
00:26:10,640 --> 00:26:14,480
cyclone four

690
00:26:12,320 --> 00:26:16,640
lime sdr i have a few of these

691
00:26:14,480 --> 00:26:17,679
uh i'm trying to get one of these blade

692
00:26:16,640 --> 00:26:19,760
rf's

693
00:26:17,679 --> 00:26:22,400
they're quite expensive this also has an

694
00:26:19,760 --> 00:26:25,919
fpga but next generation up from the

695
00:26:22,400 --> 00:26:28,159
lime sdr the second five fp here

696
00:26:25,919 --> 00:26:31,039
um there is a frequency doublers which

697
00:26:28,159 --> 00:26:33,440
uh i contacted lyme sdr about and they

698
00:26:31,039 --> 00:26:36,159
said tire one customer has bought all of

699
00:26:33,440 --> 00:26:38,720
them so it's not available so i love to

700
00:26:36,159 --> 00:26:42,000
see what this tier one customer is doing

701
00:26:38,720 --> 00:26:43,760
so i have a pcb here

702
00:26:42,000 --> 00:26:46,320
which is essentially the pc we have in

703
00:26:43,760 --> 00:26:48,400
the slide um with feed it has frequency

704
00:26:46,320 --> 00:26:50,720
doubling output feeds and

705
00:26:48,400 --> 00:26:54,799
transmit and receive channels i'll be

706
00:26:50,720 --> 00:26:54,799
quite interesting to see how how we go

707
00:26:56,080 --> 00:26:59,200
uh the next one is the

708
00:26:58,320 --> 00:27:01,360
uh

709
00:26:59,200 --> 00:27:03,200
like a passive frequency doubler so this

710
00:27:01,360 --> 00:27:04,720
one has an active chip on it

711
00:27:03,200 --> 00:27:06,000
but this one is a passive frequency

712
00:27:04,720 --> 00:27:07,600
doubler which

713
00:27:06,000 --> 00:27:09,520
which you can get as well but you have a

714
00:27:07,600 --> 00:27:10,960
bit of signal loss i have ordered a

715
00:27:09,520 --> 00:27:12,240
couple of these to experiment with as

716
00:27:10,960 --> 00:27:14,559
well

717
00:27:12,240 --> 00:27:15,919
uh so then the next part is the hardest

718
00:27:14,559 --> 00:27:18,480
to get stuff because you will be

719
00:27:15,919 --> 00:27:20,799
spraying real power into the air as the

720
00:27:18,480 --> 00:27:21,840
power amplifiers so stay within low

721
00:27:20,799 --> 00:27:22,880
powers

722
00:27:21,840 --> 00:27:25,279
um

723
00:27:22,880 --> 00:27:27,200
60 watt amps you cannot get because you

724
00:27:25,279 --> 00:27:29,600
need to sign 10 000 forms to say you're

725
00:27:27,200 --> 00:27:32,240
not gonna do anything naughty with them

726
00:27:29,600 --> 00:27:34,159
uh but 35 watt or so maybe you'll still

727
00:27:32,240 --> 00:27:38,000
have to sign forms and you can get those

728
00:27:34,159 --> 00:27:38,000
uh again power amplifiers

729
00:27:38,080 --> 00:27:42,240
and then this is a fun bit like i was

730
00:27:39,840 --> 00:27:44,320
talking about the in the can array so i

731
00:27:42,240 --> 00:27:46,320
was been messing around creating antenna

732
00:27:44,320 --> 00:27:49,120
arrays there are existing designs you

733
00:27:46,320 --> 00:27:51,520
can get of the mimo antenna arrays uh

734
00:27:49,120 --> 00:27:53,440
you can at 5.8 gigahertz you can pick up

735
00:27:51,520 --> 00:27:55,279
your standard wi-fi antenna arrays and

736
00:27:53,440 --> 00:27:57,279
work work with those

737
00:27:55,279 --> 00:27:59,919
uh so this is like

738
00:27:57,279 --> 00:28:02,320
created using uh freecad and i'm

739
00:27:59,919 --> 00:28:05,039
planning to put some antenna designs uh

740
00:28:02,320 --> 00:28:08,480
off the shelf and my own into fico which

741
00:28:05,039 --> 00:28:10,080
is a tool i used uh for uh mems based

742
00:28:08,480 --> 00:28:11,600
antenna simulation to see how they

743
00:28:10,080 --> 00:28:13,520
behave

744
00:28:11,600 --> 00:28:16,159
uh this is what a realistic antenna

745
00:28:13,520 --> 00:28:19,200
would look like uh for sending and

746
00:28:16,159 --> 00:28:19,200
receiving and so on

747
00:28:19,279 --> 00:28:24,799
uh so yeah so that's sort of uh the last

748
00:28:22,720 --> 00:28:26,960
part of the open hardware approach that

749
00:28:24,799 --> 00:28:28,880
i'm working on i don't have completed

750
00:28:26,960 --> 00:28:31,200
hardware i have got like lots of parts

751
00:28:28,880 --> 00:28:32,799
i've got like a bunch of cables to put

752
00:28:31,200 --> 00:28:35,679
into my rubber ducks

753
00:28:32,799 --> 00:28:37,840
uh but i haven't got a full built yet

754
00:28:35,679 --> 00:28:40,080
but i would love to see people contact

755
00:28:37,840 --> 00:28:42,960
me and collaborate on that

756
00:28:40,080 --> 00:28:45,679
so the skies ahead are pretty turbulent

757
00:28:42,960 --> 00:28:48,080
for our project we have sort of run

758
00:28:45,679 --> 00:28:49,919
through the angel funds that we had

759
00:28:48,080 --> 00:28:51,440
doing learning various things

760
00:28:49,919 --> 00:28:53,120
experimenting

761
00:28:51,440 --> 00:28:55,039
uh funding for building something in

762
00:28:53,120 --> 00:28:57,120
africa is challenging people don't seek

763
00:28:55,039 --> 00:28:59,360
any clear path to revenue

764
00:28:57,120 --> 00:29:00,720
operating with these things outside of

765
00:28:59,360 --> 00:29:02,799
cleared lab

766
00:29:00,720 --> 00:29:04,159
in a western world is challenging a

767
00:29:02,799 --> 00:29:06,480
little bit

768
00:29:04,159 --> 00:29:09,679
so that's that's quite interesting in

769
00:29:06,480 --> 00:29:11,600
balance in africa we can reach out to

770
00:29:09,679 --> 00:29:14,080
people i have seen the head of the

771
00:29:11,600 --> 00:29:16,480
meteorological department uh

772
00:29:14,080 --> 00:29:18,640
assistant director or in kenya and

773
00:29:16,480 --> 00:29:21,039
pitched our idea and got some support of

774
00:29:18,640 --> 00:29:21,840
some some push back and corruption as

775
00:29:21,039 --> 00:29:24,240
well

776
00:29:21,840 --> 00:29:26,320
but yeah so it's that will be a higher

777
00:29:24,240 --> 00:29:28,159
bar to do in australia

778
00:29:26,320 --> 00:29:29,679
the operating environment in africa is

779
00:29:28,159 --> 00:29:31,679
difficult building hardware getting

780
00:29:29,679 --> 00:29:33,360
stuff shipped in is difficult

781
00:29:31,679 --> 00:29:35,679
and i would love to see open source

782
00:29:33,360 --> 00:29:38,720
hardware and software community jump in

783
00:29:35,679 --> 00:29:41,039
with their resources to put together a

784
00:29:38,720 --> 00:29:43,679
good set of things that can be that can

785
00:29:41,039 --> 00:29:46,080
be relatively affordably built into a

786
00:29:43,679 --> 00:29:47,840
large array so to in order to build a

787
00:29:46,080 --> 00:29:50,399
antenna array like this you would

788
00:29:47,840 --> 00:29:52,159
require hundreds of fpr's

789
00:29:50,399 --> 00:29:52,960
not hundreds yeah perhaps under this

790
00:29:52,159 --> 00:29:55,760
yeah

791
00:29:52,960 --> 00:29:59,039
when in my lab also in the lab we built

792
00:29:55,760 --> 00:30:02,080
a 100 antenna gps passive radar so you

793
00:29:59,039 --> 00:30:04,640
could bounce gps of an airplane and see

794
00:30:02,080 --> 00:30:06,240
use that as a radar so it requires a lot

795
00:30:04,640 --> 00:30:09,039
of a lot of

796
00:30:06,240 --> 00:30:10,559
rf rf signal chain stuff

797
00:30:09,039 --> 00:30:12,000
yes i would love to see the open source

798
00:30:10,559 --> 00:30:15,200
hardware and software community dive

799
00:30:12,000 --> 00:30:17,679
deeper into the rf land

800
00:30:15,200 --> 00:30:20,159
and thank you for listening i'm happy to

801
00:30:17,679 --> 00:30:22,559
take questions uh i think that went went

802
00:30:20,159 --> 00:30:24,559
well for a reasonable amount of time

803
00:30:22,559 --> 00:30:26,640
i was i've been making this presentation

804
00:30:24,559 --> 00:30:28,640
after john asked me to do it over a few

805
00:30:26,640 --> 00:30:30,240
few weeks and i was hanging out whenever

806
00:30:28,640 --> 00:30:31,520
i would have free moment i would hang

807
00:30:30,240 --> 00:30:33,520
out

808
00:30:31,520 --> 00:30:35,200
and one of the people who saw me writing

809
00:30:33,520 --> 00:30:37,840
this presentation wrote me a poem so i

810
00:30:35,200 --> 00:30:37,840
have it in there

811
00:30:38,320 --> 00:30:41,919
oh that's great thank you tish that was

812
00:30:40,559 --> 00:30:43,360
very interesting

813
00:30:41,919 --> 00:30:44,880
now we're pretty much on the limit for

814
00:30:43,360 --> 00:30:47,360
time here but we do have a couple of

815
00:30:44,880 --> 00:30:48,640
questions which hopefully

816
00:30:47,360 --> 00:30:50,799
will be brief

817
00:30:48,640 --> 00:30:52,159
and the first one is calling back to

818
00:30:50,799 --> 00:30:52,960
what you were just mentioning a moment

819
00:30:52,159 --> 00:30:55,520
ago

820
00:30:52,960 --> 00:30:57,840
um you were talking about gps

821
00:30:55,520 --> 00:30:59,360
reflectometry and one of the questions

822
00:30:57,840 --> 00:31:01,039
from in the chat is do you do any work

823
00:30:59,360 --> 00:31:03,279
with passive reception of reflected

824
00:31:01,039 --> 00:31:04,399
noise the rf component of rain noise can

825
00:31:03,279 --> 00:31:07,519
be received but i don't know how

826
00:31:04,399 --> 00:31:11,039
directional it is fairly broadband

827
00:31:07,519 --> 00:31:13,679
yeah yeah so some of the ideas are there

828
00:31:11,039 --> 00:31:16,960
for using rain fade as well

829
00:31:13,679 --> 00:31:19,440
so if you have a good number of cell

830
00:31:16,960 --> 00:31:21,760
towers or things like that or you can

831
00:31:19,440 --> 00:31:23,760
measure the rain fade using passive

832
00:31:21,760 --> 00:31:25,760
received mode only and then you have to

833
00:31:23,760 --> 00:31:27,919
again go back and invert all of that in

834
00:31:25,760 --> 00:31:30,880
the actual body of the rain and the

835
00:31:27,919 --> 00:31:33,039
reflectivity is scatters so you have to

836
00:31:30,880 --> 00:31:35,039
do it's a quite a difficult inversion

837
00:31:33,039 --> 00:31:37,360
problem if you have control over the

838
00:31:35,039 --> 00:31:38,480
radiator it's an easier easier problem

839
00:31:37,360 --> 00:31:39,360
to solve

840
00:31:38,480 --> 00:31:41,279
okay

841
00:31:39,360 --> 00:31:44,640
and another question is do any of the

842
00:31:41,279 --> 00:31:46,399
marine radars work with linux

843
00:31:44,640 --> 00:31:49,360
ah that's that's a that's a good

844
00:31:46,399 --> 00:31:51,440
question so uh the marine readers ship

845
00:31:49,360 --> 00:31:53,279
with their own hardware uh so some of

846
00:31:51,440 --> 00:31:56,399
the you have to break into the signal

847
00:31:53,279 --> 00:32:00,399
chain i spent an amount of time sitting

848
00:31:56,399 --> 00:32:02,799
on um wireshark uh they have ethernet

849
00:32:00,399 --> 00:32:04,960
going over the protocols and stuff uh i

850
00:32:02,799 --> 00:32:06,000
haven't opened uh any of that yet it's

851
00:32:04,960 --> 00:32:07,600
not very

852
00:32:06,000 --> 00:32:09,919
i'm not sure what the

853
00:32:07,600 --> 00:32:12,000
legal ramifications of reverse

854
00:32:09,919 --> 00:32:14,960
engineering these things is yet i don't

855
00:32:12,000 --> 00:32:14,960
have i'm

856
00:32:15,360 --> 00:32:19,039
yeah so

857
00:32:16,480 --> 00:32:21,279
they work over ethernet so once a driver

858
00:32:19,039 --> 00:32:23,760
is built and the ethernet protocol is

859
00:32:21,279 --> 00:32:25,200
reverse engineered you could make them

860
00:32:23,760 --> 00:32:27,519
work

861
00:32:25,200 --> 00:32:29,519
okay i would love help there sen if

862
00:32:27,519 --> 00:32:32,080
please privately send me something else

863
00:32:29,519 --> 00:32:32,880
and do some wireshark capture packets

864
00:32:32,080 --> 00:32:34,080
yeah

865
00:32:32,880 --> 00:32:35,360
um

866
00:32:34,080 --> 00:32:37,120
there was

867
00:32:35,360 --> 00:32:38,640
we are really pushing for time but no

868
00:32:37,120 --> 00:32:40,080
two more quick questions one is would

869
00:32:38,640 --> 00:32:43,279
these radars be useful for people

870
00:32:40,080 --> 00:32:45,600
running solar power plants

871
00:32:43,279 --> 00:32:48,080
interesting uh so if you have want to

872
00:32:45,600 --> 00:32:50,000
see real time quick response so if you

873
00:32:48,080 --> 00:32:52,960
want to see rain clouds arriving over

874
00:32:50,000 --> 00:32:54,960
the horizon beyond the predictions that

875
00:32:52,960 --> 00:32:56,640
bureau of med provides you

876
00:32:54,960 --> 00:32:58,880
then you would want to do this yourself

877
00:32:56,640 --> 00:33:00,399
bureau of met i have spoken to jaya from

878
00:32:58,880 --> 00:33:03,039
after my previous presentation at

879
00:33:00,399 --> 00:33:05,120
birofmet and bureau of met is about to

880
00:33:03,039 --> 00:33:08,720
release after many years of research uh

881
00:33:05,120 --> 00:33:10,480
like 90 minute now casting uh service

882
00:33:08,720 --> 00:33:12,240
which they'll it takes them five minutes

883
00:33:10,480 --> 00:33:14,080
to assimilate and run the radar and then

884
00:33:12,240 --> 00:33:16,559
they can look forward 90 minutes which

885
00:33:14,080 --> 00:33:18,640
is fairly accurate so maybe rather than

886
00:33:16,559 --> 00:33:21,120
hosting your own uh in australia

887
00:33:18,640 --> 00:33:23,200
especially you could use the bombs bomb

888
00:33:21,120 --> 00:33:24,960
is an excellent science agency you could

889
00:33:23,200 --> 00:33:27,039
use bomb service

890
00:33:24,960 --> 00:33:28,559
yeah okay um well we better wrap it up

891
00:33:27,039 --> 00:33:30,159
but just before you go

892
00:33:28,559 --> 00:33:31,519
if people want to follow up on this and

893
00:33:30,159 --> 00:33:33,279
they want to learn more about your

894
00:33:31,519 --> 00:33:34,880
project is there a particular place they

895
00:33:33,279 --> 00:33:37,840
should look

896
00:33:34,880 --> 00:33:39,919
um there is a twitter bot that we built

897
00:33:37,840 --> 00:33:42,159
as part of this uh called heroin aka

898
00:33:39,919 --> 00:33:44,720
which has been gone silent i haven't

899
00:33:42,159 --> 00:33:46,480
been able to maintain it um so you can

900
00:33:44,720 --> 00:33:49,120
look at hawaiian k's previous updates

901
00:33:46,480 --> 00:33:52,159
over nairobi so hewani is what is in the

902
00:33:49,120 --> 00:33:54,559
air k in kenya so it gives tweets out

903
00:33:52,159 --> 00:33:56,240
the rain in nairobi

904
00:33:54,559 --> 00:33:58,240
and you can also reach me on twitter on

905
00:33:56,240 --> 00:33:59,440
what nick is the social media i'm most

906
00:33:58,240 --> 00:34:01,919
active in

907
00:33:59,440 --> 00:34:03,360
yeah great that's fantastic well thanks

908
00:34:01,919 --> 00:34:04,640
tish i really appreciate you coming

909
00:34:03,360 --> 00:34:06,080
along and

910
00:34:04,640 --> 00:34:08,000
this has been a very interesting thing

911
00:34:06,080 --> 00:34:10,240
and a couple of people in the chat also

912
00:34:08,000 --> 00:34:11,839
mentioned i really liked the imagery of

913
00:34:10,240 --> 00:34:14,000
the ocean in the sky that you mentioned

914
00:34:11,839 --> 00:34:15,040
at the start

915
00:34:14,000 --> 00:34:17,599
thank you

916
00:34:15,040 --> 00:34:20,599
okay thank you see ya

917
00:34:17,599 --> 00:34:20,599
bye