ABOUT THE SPEAKER
Bart Weetjens - Product developer
The founder of Apopo, Bart Weet­jens, is train­ing rats to detect landmine explosives in minute amounts.

Why you should listen

Dutch product designer Bart Weetjens works with locals in Morogoro, Tanzania, to trains rats to do something astonishing: sniff out land mines. (The African giant pouched rat, the species used in the project, is wide­spread in the region.) The rats that pass the train­ing -- and the same rigorous testing applied to land mine-sniffing canines -- become what Weetjens calls "HeroRATS."

Weet­jens and Apopo.org are now applying a similar approach to other fields, training rats to diagnose tuber­cu­lo­sis in hospitals.

More profile about the speaker
Bart Weetjens | Speaker | TED.com
TEDxRotterdam 2010

Bart Weetjens: How I taught rats to sniff out land mines

Filmed:
610,956 views

No one knows exactly how many landmines still litter the world, but it's safe to say: millions, waiting to kill and maim unsuspecting civilians. Clearing them is slow, expensive and dangerous. The founder of Apopo, Bart Weetjens, talks about his extraordinary project: training rats to sniff out land mines. He shows clips of his "hero rats" in action, and previews his work's next phase: teaching them to turn up tuberculosis in the lab.
- Product developer
The founder of Apopo, Bart Weet­jens, is train­ing rats to detect landmine explosives in minute amounts. Full bio

Double-click the English transcript below to play the video.

00:16
I'm here today to share with you
0
1000
2000
00:18
an extraordinary journey --
1
3000
2000
00:20
extraordinarily rewarding journey, actually --
2
5000
3000
00:23
which brought me into
3
8000
2000
00:25
training rats
4
10000
2000
00:27
to save human lives
5
12000
2000
00:29
by detecting landmines
6
14000
2000
00:31
and tuberculosis.
7
16000
2000
00:33
As a child, I had two passions.
8
18000
3000
00:36
One was a passion for rodents.
9
21000
3000
00:39
I had all kinds of rats,
10
24000
2000
00:41
mice, hamsters,
11
26000
2000
00:43
gerbils, squirrels.
12
28000
2000
00:45
You name it, I bred it, and I sold them to pet shops.
13
30000
3000
00:48
(Laughter)
14
33000
2000
00:50
I also had a passion for Africa.
15
35000
3000
00:53
Growing up in a multicultural environment,
16
38000
2000
00:55
we had African students in the house,
17
40000
2000
00:57
and I learned about their stories,
18
42000
2000
00:59
so different backgrounds,
19
44000
2000
01:01
dependency on imported know-how,
20
46000
3000
01:04
goods, services,
21
49000
2000
01:06
exuberant cultural diversity.
22
51000
3000
01:09
Africa was truly fascinating for me.
23
54000
2000
01:11
I became an industrial engineer,
24
56000
2000
01:13
engineer in product development,
25
58000
2000
01:15
and I focused on appropriate detection technologies,
26
60000
3000
01:18
actually the first appropriate technologies
27
63000
2000
01:20
for developing countries.
28
65000
3000
01:23
I started working in the industry,
29
68000
2000
01:25
but I wasn't really happy to contribute
30
70000
2000
01:27
to a material consumer society
31
72000
3000
01:30
in a linear, extracting
32
75000
3000
01:33
and manufacturing mode.
33
78000
2000
01:35
I quit my job to focus on the real world problem:
34
80000
2000
01:37
landmines.
35
82000
3000
01:40
We're talking '95 now.
36
85000
3000
01:43
Princess Diana is announcing on TV
37
88000
3000
01:46
that landmines form a structural barrier
38
91000
2000
01:48
to any development, which is really true.
39
93000
3000
01:51
As long as these devices are there,
40
96000
2000
01:53
or there is suspicion of landmines,
41
98000
2000
01:55
you can't really enter into the land.
42
100000
2000
01:57
Actually, there was an appeal worldwide
43
102000
2000
01:59
for new detectors
44
104000
3000
02:02
sustainable in the environments
45
107000
2000
02:04
where they're needed to produce,
46
109000
2000
02:06
which is mainly in the developing world.
47
111000
2000
02:08
We chose rats.
48
113000
2000
02:10
Why would you choose rats?
49
115000
2000
02:12
Because, aren't they vermin?
50
117000
2000
02:14
Well, actually rats are,
51
119000
2000
02:16
in contrary to what most people think about them,
52
121000
2000
02:18
rats are highly sociable creatures.
53
123000
4000
02:22
And actually, our product -- what you see here.
54
127000
3000
02:25
There's a target somewhere here.
55
130000
2000
02:27
You see an operator, a trained African
56
132000
2000
02:29
with his rats in front
57
134000
2000
02:31
who actually are left and right.
58
136000
2000
02:33
There, the animal finds a mine.
59
138000
2000
02:35
It scratches on the soil.
60
140000
2000
02:37
And the animal comes back for a food reward.
61
142000
3000
02:40
Very, very simple.
62
145000
2000
02:42
Very sustainable in this environment.
63
147000
3000
02:45
Here, the animal gets its food reward.
64
150000
3000
02:48
And that's how it works.
65
153000
2000
02:50
Very, very simple.
66
155000
2000
02:52
Now why would you use rats?
67
157000
2000
02:54
Rats have been used since the '50s last century,
68
159000
2000
02:56
in all kinds of experiments.
69
161000
3000
02:59
Rats have more genetic material
70
164000
3000
03:02
allocated to olfaction
71
167000
2000
03:04
than any other mammal species.
72
169000
2000
03:06
They're extremely sensitive to smell.
73
171000
3000
03:09
Moreover, they have the mechanisms to map all these smells
74
174000
3000
03:12
and to communicate about it.
75
177000
3000
03:15
Now how do we communicate with rats?
76
180000
2000
03:17
Well don't talk rat,
77
182000
3000
03:20
but we have a clicker,
78
185000
2000
03:22
a standard method for animal training,
79
187000
2000
03:24
which you see there.
80
189000
2000
03:26
A clicker, which makes a particular sound
81
191000
3000
03:29
with which you can reinforce particular behaviors.
82
194000
3000
03:32
First of all, we associate the click sound with a food reward,
83
197000
3000
03:35
which is smashed banana and peanuts together in a syringe.
84
200000
3000
03:39
Once the animal knows click, food,
85
204000
2000
03:41
click, food, click, food --
86
206000
2000
03:43
so click is food --
87
208000
2000
03:45
we bring it in a cage with a hole,
88
210000
2000
03:47
and actually the animal learns
89
212000
2000
03:49
to stick the nose in the hole
90
214000
2000
03:51
under which a target scent is placed,
91
216000
2000
03:53
and to do that for five seconds --
92
218000
2000
03:55
five seconds, which is long for a rat.
93
220000
2000
03:57
Once the animal knows this, we make the task a bit more difficult.
94
222000
3000
04:00
It learns how to find the target smell
95
225000
3000
04:03
in a cage with several holes, up to 10 holes.
96
228000
3000
04:06
Then the animal learns
97
231000
2000
04:08
to walk on a leash in the open
98
233000
2000
04:10
and find targets.
99
235000
2000
04:12
In the next step, animals learn
100
237000
3000
04:15
to find real mines in real minefields.
101
240000
2000
04:17
They are tested and accredited
102
242000
3000
04:20
according to International Mine Action Standards,
103
245000
2000
04:22
just like dogs have to pass a test.
104
247000
3000
04:25
This consists of 400 square meters.
105
250000
2000
04:27
There's a number of mines
106
252000
3000
04:30
placed blindly,
107
255000
2000
04:32
and the team of trainer and their rat
108
257000
3000
04:35
have to find all the targets.
109
260000
3000
04:39
If the animal does that, it gets a license
110
264000
3000
04:42
as an accredited animal
111
267000
2000
04:44
to be operational in the field --
112
269000
2000
04:46
just like dogs, by the way.
113
271000
2000
04:48
Maybe one slight difference:
114
273000
2000
04:50
we can train rats at a fifth of the price
115
275000
3000
04:53
of training the mining dog.
116
278000
2000
04:55
This is our team in Mozambique:
117
280000
2000
04:57
one Tanzanian trainer,
118
282000
2000
04:59
who transfers the skills
119
284000
2000
05:01
to these three Mozambican fellows.
120
286000
2000
05:03
And you should see the pride in the eyes of these people.
121
288000
3000
05:06
They have a skill,
122
291000
2000
05:08
which makes them much less dependent
123
293000
2000
05:10
on foreign aid.
124
295000
2000
05:12
Moreover, this small team
125
297000
3000
05:15
together with, of course, you need the heavy vehicles
126
300000
3000
05:18
and the manual de-miners to follow-up.
127
303000
3000
05:21
But with this small investment in a rat capacity,
128
306000
3000
05:24
we have demonstrated in Mozambique
129
309000
3000
05:27
that we can reduce the cost-price per square meter
130
312000
3000
05:30
up to 60 percent
131
315000
2000
05:32
of what is currently normal --
132
317000
2000
05:34
two dollars per square meter, we do it at $1.18,
133
319000
2000
05:36
and we can still bring that price down.
134
321000
2000
05:38
Question of scale.
135
323000
2000
05:40
If you can bring in more rats,
136
325000
2000
05:42
we can actually make the output even bigger.
137
327000
2000
05:44
We have a demonstration site in Mozambique.
138
329000
3000
05:47
Eleven African governments
139
332000
3000
05:50
have seen that they can become less dependent
140
335000
3000
05:53
by using this technology.
141
338000
2000
05:55
They have signed the pact for peace
142
340000
2000
05:57
and treaty in the Great Lakes region,
143
342000
3000
06:00
and they endorse hero rats
144
345000
3000
06:03
to clear their common borders of landmines.
145
348000
3000
06:06
But let me bring you to a very different problem.
146
351000
3000
06:09
And there's about 6,000 people last year
147
354000
2000
06:11
that walked on a landmine,
148
356000
2000
06:13
but worldwide last year,
149
358000
2000
06:15
almost 1.9 million died from tuberculosis
150
360000
2000
06:17
as a first cause of infection.
151
362000
3000
06:21
Especially in Africa
152
366000
2000
06:23
where T.B. and HIV are strongly linked,
153
368000
3000
06:26
there is a huge common problem.
154
371000
4000
06:31
Microscopy, the standard WHO procedure,
155
376000
3000
06:34
reaches from 40 to 60 percent reliability.
156
379000
3000
06:38
In Tanzania -- the numbers don't lie --
157
383000
3000
06:41
45 percent of people -- T.B. patients --
158
386000
3000
06:44
get diagnosed with T.B. before they die.
159
389000
3000
06:48
It means that, if you have T.B.,
160
393000
3000
06:51
you have more chance that you won't be detected,
161
396000
2000
06:53
but will just die from T.B. secondary infections and so on.
162
398000
3000
07:00
And if, however,
163
405000
2000
07:02
you are detected very early, diagnosed early,
164
407000
2000
07:04
treatment can start,
165
409000
2000
07:06
and even in HIV-positives, it makes sense.
166
411000
3000
07:09
You can actually cure T.B.,
167
414000
2000
07:11
even in HIV-positives.
168
416000
3000
07:14
So in our common language, Dutch,
169
419000
3000
07:17
the name for T.B.
170
422000
2000
07:19
is "tering,"
171
424000
2000
07:21
which, etymologically,
172
426000
2000
07:23
refers to the smell of tar.
173
428000
3000
07:26
Already the old Chinese
174
431000
2000
07:28
and the Greek, Hippocrates,
175
433000
3000
07:31
have actually published,
176
436000
2000
07:33
documented, that T.B. can be diagnosed
177
438000
3000
07:36
based on the volatiles
178
441000
2000
07:38
exuding from patients.
179
443000
3000
07:41
So what we did is we collected some samples --
180
446000
2000
07:43
just as a way of testing --
181
448000
2000
07:45
from hospitals,
182
450000
2000
07:47
trained rats on them
183
452000
3000
07:50
and see if this works,
184
455000
2000
07:52
and wonder, well,
185
457000
2000
07:54
we can reach 89 percent sensitivity,
186
459000
2000
07:56
86 percent specificity
187
461000
2000
07:58
using multiple rats in a row.
188
463000
2000
08:00
This is how it works,
189
465000
2000
08:03
and really, this is a generic technology.
190
468000
3000
08:06
We're talking now explosives, tuberculosis,
191
471000
3000
08:09
but can you imagine,
192
474000
2000
08:11
you can actually put anything under there.
193
476000
2000
08:13
So how does it work?
194
478000
2000
08:15
You have a cassette with 10 samples.
195
480000
2000
08:17
You put these 10 samples at once in the cage.
196
482000
3000
08:20
An animal only needs two hundredths of a second
197
485000
2000
08:22
to discriminate the scent, so it goes extremely fast.
198
487000
3000
08:25
Here it's already at the third sample.
199
490000
3000
08:28
This is a positive sample.
200
493000
3000
08:32
It gets a click sound and comes for the food reward.
201
497000
3000
08:37
And by doing so, very fast,
202
502000
2000
08:39
we can have like a second-line opinion
203
504000
3000
08:42
to see which patients are positive,
204
507000
2000
08:44
which are negative.
205
509000
3000
08:47
Just as an indication,
206
512000
2000
08:49
whereas a microscopist can process
207
514000
2000
08:51
40 samples in a day,
208
516000
2000
08:53
a rat can process
209
518000
2000
08:55
the same amount of samples
210
520000
2000
08:57
in seven minutes only.
211
522000
2000
08:59
A cage like this --
212
524000
2000
09:01
(Applause)
213
526000
5000
09:06
A cage like this -- provided that you have rats,
214
531000
3000
09:09
and we have now currently
215
534000
2000
09:11
25 tuberculosis rats --
216
536000
2000
09:13
a cage like this, operating throughout the day,
217
538000
3000
09:16
can process 1,680 samples.
218
541000
3000
09:21
Can you imagine the potential offspring applications --
219
546000
3000
09:24
environmental detection
220
549000
2000
09:26
of pollutants in soils,
221
551000
2000
09:28
customs applications,
222
553000
2000
09:30
detection of illicit goods in containers and so on.
223
555000
3000
09:34
But let's stick first to tuberculosis.
224
559000
2000
09:36
I just want to briefly highlight,
225
561000
2000
09:38
the blue rods
226
563000
2000
09:40
are the scores of microscopy only
227
565000
2000
09:42
at the five clinics in Dar es Salaam
228
567000
3000
09:45
on a population of 500,000 people,
229
570000
2000
09:47
where 15,000 reported to get a test done.
230
572000
3000
09:50
Microscopy for 1,800 patients.
231
575000
3000
09:53
And by just presenting the samples once more to the rats
232
578000
4000
09:57
and looping those results back,
233
582000
3000
10:00
we were able to increase case detection rates
234
585000
2000
10:02
by over 30 percent.
235
587000
2000
10:04
Throughout last year,
236
589000
2000
10:06
we've been -- depending on which intervals you take --
237
591000
2000
10:08
we've been consistently
238
593000
2000
10:10
increasing case detection rates
239
595000
2000
10:12
in five hospitals in Dar es Salaam
240
597000
2000
10:14
between 30 and 40 percent.
241
599000
3000
10:17
So this is really considerable.
242
602000
2000
10:19
Knowing that a missed patient by microscopy
243
604000
2000
10:21
infects up to 15 people,
244
606000
2000
10:23
healthy people, per year,
245
608000
2000
10:25
you can be sure
246
610000
2000
10:27
that we have saved lots of lives.
247
612000
2000
10:29
At least our hero rats have saved lots of lives.
248
614000
3000
10:32
The way forward for us
249
617000
2000
10:34
is now to standardize this technology.
250
619000
2000
10:36
And there are simple things
251
621000
2000
10:38
like, for instance, we have a small laser in the sniffer hole
252
623000
4000
10:42
where the animal has to stick for five seconds.
253
627000
2000
10:44
So, to standardize this.
254
629000
2000
10:46
Also, to standardize the pellets,
255
631000
2000
10:48
the food rewards,
256
633000
2000
10:50
and to semi-automate this
257
635000
2000
10:52
in order to replicate this on a much larger scale
258
637000
3000
10:55
and affect the lives of many more people.
259
640000
3000
10:58
To conclude, there are also other applications at the horizon.
260
643000
3000
11:01
Here is a first prototype
261
646000
2000
11:03
of our camera rat,
262
648000
2000
11:05
which is a rat with a rat backpack
263
650000
2000
11:07
with a camera that can go under rubble
264
652000
2000
11:09
to detect for victims
265
654000
2000
11:11
after earthquake and so on.
266
656000
2000
11:13
This is in a prototype stage.
267
658000
2000
11:15
We don't have a working system here yet.
268
660000
2000
11:18
To conclude, I would actually like to say,
269
663000
3000
11:21
you may think this is about rats, these projects,
270
666000
2000
11:23
but in the end it is about people.
271
668000
2000
11:25
It is about empowering vulnerable communities
272
670000
2000
11:27
to tackle difficult, expensive
273
672000
3000
11:30
and dangerous humanitarian detection tasks,
274
675000
3000
11:33
and doing that with a local resource,
275
678000
3000
11:36
plenty available.
276
681000
2000
11:38
So something completely different
277
683000
3000
11:41
is to keep on challenging your perception
278
686000
3000
11:44
about the resources surrounding you,
279
689000
3000
11:47
whether they are environmental,
280
692000
3000
11:50
technological, animal, or human.
281
695000
3000
11:55
And to respectfully harmonize with them
282
700000
3000
11:58
in order to foster a sustainable work.
283
703000
3000
12:01
Thank you very much.
284
706000
2000
12:03
(Applause)
285
708000
2000

▲Back to top

ABOUT THE SPEAKER
Bart Weetjens - Product developer
The founder of Apopo, Bart Weet­jens, is train­ing rats to detect landmine explosives in minute amounts.

Why you should listen

Dutch product designer Bart Weetjens works with locals in Morogoro, Tanzania, to trains rats to do something astonishing: sniff out land mines. (The African giant pouched rat, the species used in the project, is wide­spread in the region.) The rats that pass the train­ing -- and the same rigorous testing applied to land mine-sniffing canines -- become what Weetjens calls "HeroRATS."

Weet­jens and Apopo.org are now applying a similar approach to other fields, training rats to diagnose tuber­cu­lo­sis in hospitals.

More profile about the speaker
Bart Weetjens | Speaker | TED.com