Episode 30
The LoRaWAN Tsunami with Olivier Hersent
Olivier Hersent, founder and CEO of Actility, talks about the accelerating convergence of LoRaWAN and BACnet in building automation, and what that means for retrofits, logistics, and the future of wireless sensing. With BACnet still dominating 70% of the building automation market—even on brand new PLCs—Olivier explains why making LoRaWAN invisible to integrators is critical, and how Actility is bridging these two worlds one sensor at a time.
He also shares insights into the work Actility is doing with custom sensing and logistics tracking, including real-world deployments in automotive assembly lines, massive vehicle yards, and remote conservation areas. From RS485-era wired sensors to modern closed-loop logistics, Olivier points to a shift toward simplicity and targeted data collection, where most application layers are 99% noise and only one bit matters.
- Why BACnet remains the king of building automation—and what makes integrating LoRa so difficult
- How Actility is streamlining LoRaWAN-to-BACnet translation by mapping and testing each individual sensor
- The growing role of AI in simplifying sensor driver creation and semantic standards
- A new market segment in geolocation: closed-loop logistics and the return of valuable assets
- Why Africa is poised to be the next major IoT growth market, with utilities growing 50% year-over-year
- The cost advantages of LoRaWAN for emerging economies and remote infrastructure
- Why older, simpler wired protocols like RS485 are easier to port to LoRaWAN
- How LoRaWAN could reduce power grid strain by syncing appliances with real-time energy signals
- Olivier’s advice for new founders: help wired sensing companies go wireless
Transcript
Today's guest on MeteoScientific's
the Business of LoRawan is Olivier
2
:Hersent, founder and CEO of Actility,
one of the core companies
3
:behind the global LoRaWAN movement.
4
:Olivier isn't just a leader in the space,
he's a founder.
5
:He helped
write the original LoRaWAN standards
6
:and has spent the past decade driving
real world adoption across
7
:smart cities, industry
and infrastructure worldwide.
8
:In this episode, we dig into the explosive
growth of LoRaWAN in building
9
:automation, especially its integration
with the legacy BACnet protocol.
10
:Olivia explains how bridging LoRaWAN
with BACnet is unlocking massive retrofit
11
:potential in existing
12
:buildings, an overlooked
but crucial market for wireless sensing.
13
:He also walks us through emerging
geolocation use cases in closed loop
14
:logistics like automotive manufacturing
and vehicle tracking, and offers a sharp
15
:look at how AI and LoRaWAN are beginning
to work together at the edge.
16
:We covered deployments in Africa, defense
applications and Olivier's
17
:thoughts on what it would take to start
a successful LoRaWAN business today.
18
:Let's dig in. Enjoy the convo.
19
:Right on Olivier.
20
:Thanks so much for making the time to come
on. I know you're busy.
21
:It's great to have you here.
22
:Hi Nik, thanks for picking me up.
23
:Yeah, of course
it's it's an honor to have you on.
24
:I thought we'd start talking about
one of the big things that I'm hearing
25
:a ton about in LoRaWAN right now,
which is this connection with BACnet,
26
:which seems to be kind of the next big
frontier in smart buildings in general.
27
:Can you walk me through
28
:why it's such a big deal for you
as you're looking into kind of LoRaWAN
29
:in general and say, like,
this is what this will unlock for us?
30
:Well, you know what everybody says about
new technologies in new markets, right?
31
:It's very hard.
32
:So you tend to be willing to find
33
:something that new technology is useful
for in an existing market.
34
:Building
automation is probably one of the biggest,
35
:most stable markets around here.
36
:But it's very, very,
very legacy in that market.
37
:The king protocol is BACnet,
so to a certain degree, Modbus as well.
38
:But BACnet is really like 70% dominant
in that market.
39
:This protocol is so old that I don't know
if you will even remember all the times
40
:where you had to connect, coax Ethernet
and you had no, you know, it.
41
:So it was all broadcast.
42
:So, so BACnet is so old
that doesn't support routing.
43
:You know,
you can't have routed IP packets.
44
:But interestingly,
even if you buy a brand new PLC today
45
:to control the building,
it will have that BACnet.
46
:So we figured that if we would bridge
LoRaWAN and BACnet together,
47
:you would make things so much simpler
for anyone trying to retrofit
48
:a new sensor use case
in an existing building without the wires.
49
:And there are a few technologies,
wireless technologies that worked
50
:in the building automation
space and ocean being one.
51
:But nothing was really like good enough
for everything, you know.
52
:Works just as good as wired.
53
:And somehow it seems the market
54
:is not perceiving LoRaWAN
as being as good as wired.
55
:And because of the link with BACnet
56
:as easy to use as my good,
good old wired sensor.
57
:So it's been the boom surprising
even to us.
58
:You know, an example
I want to give the in in that market
59
:the dominant software is Niagara
by a company called Tritium.
60
:In the Tritium user forum,
you would have nothing about LoRa, right?
61
:Why would you.
62
:It's totally different space
in the Niagara User forum in London.
63
:That was six months ago.
64
:Each booth, like 100% of the booths,
had the LoRaWAN gateway on the stand.
65
:And all of them
were talking about wireless sensing.
66
:So it's been massive and it's going to be
even bigger in the years to come.
67
:So great. Yeah.
68
:Yeah. It's good.
And you've called it a tsunami, which is.
69
:Which is pretty
70
:is pretty pretty big word there as far
as the promise of what's to come.
71
:What do you think are the bottlenecks?
72
:Are there going to be enough
73
:people who know enough about LoRaWAN
and BACnet and integrating them?
74
:Or is it just like,
oh, there's millions of buildings
75
:and there's only a couple
hundred engineers who know how to do this.
76
:Like, what are the things that are going
to really slow this down?
77
:That's the whole point, this industry,
I mean, the building automation
78
:industry is resource constrained.
79
:There are few experts.
80
:They are used 120%.
81
:They have no will whatsoever
to learn a new technology.
82
:It's not that they're not curious.
They don't have the time.
83
:So the whole point is to make LoRa
completely transparent to them.
84
:What we do is we actually take all the
LoRa sensors, we convert them to BACnet
85
:in a way that you just connect back to
your existing layer, and it's all working.
86
:You don't need to even understand
what LoRaWAN is doing.
87
:And if you don't make it that easy,
it doesn't work.
88
:This doesn't work.
89
:Even 1% doesn't work at all.
90
:There
no way to get in this market. Got it.
91
:And so that will be an advantage
or a moat for Actility to say, hey,
92
:we can make BACnet easy.
93
:Is it as easy for someone
or some other company to do?
94
:Or is it that is a unique kind of ability
that Actility has?
95
:It's a lot of work.
96
:It’s not about being more intelligent
than others.
97
:Like many, many things, it's
about being more persistent.
98
:You need to take each and every sensor
around there in the LoRaWAN world.
99
:And as you know, LoRaWAN doesn't
have a standard for the application layer.
100
:So any sensor
is going to have a different way of,
101
:you know, different semantics,
different ways of doing what it does.
102
:You need to read it, understand it,
build a driver, map it to BACnet, test it.
103
:And and it's it's it's a lot of work.
104
:But we we've we've been doing this
for many years now.
105
:And we, we think we're getting to,
to that point where it's all,
106
:you know, smooth enough
to, to be used by BACnet professionals.
107
:And do you think that
AI will have a place in this?
108
:It seems like it's
109
:I mean, I have been super helpful to me
into LoRaWAN fairly new since:
110
:So saying like, okay,
hey, write me a decoder or fix this thing
111
:or figure this thing out.
112
:And you know, something,
even as general as GPT will will do that,
113
:I think it will actually.
114
:We are trying.
115
:But as you know, AI needs examples.
116
:So at this point we're trying to
first of all, we do it a lot manually
117
:while trying to also improve the standard,
118
:because the best way to not have to write
drivers is to make sure
119
:people, the right drivers
in the correct way up front.
120
:The LoRa alliance has progressed
121
:a little bit there,
but not enough in the actual semantics.
122
:So there is still progress to do.
123
:And once we have all these examples,
I think not now, but maybe in a year
124
:we'll be able
to, you know, have a prompt like,
125
:look at these 100 examples, just read
the manual, please write the same thing.
126
:And I think it will work.
127
:And is that something that Actility
is going to make available
128
:as a tool to its customers, or
do you think that will just be available
129
:in general in the marketplace
or some hybrid?
130
:So right now what we do
is we obviously the standard itself,
131
:we push in the LoRa Alliance
and we continue to it.
132
:So that's available to everyone.
133
:And and actually most maybe
134
:60-70% of LoRaWAN sensor vendors
now adopt the standard.
135
:And we're trying to progress it
in terms of writing your own drivers.
136
:It's available in our tools.
137
:You can pick an existing driver,
tweak it, improve it, or write your own.
138
:And the next step with the AI writing
is something that we're working on.
139
:Of course, it's not yet something you can
just click on and from your driver.
140
:We're getting there.
141
:It's it's it's amazing to me
how fast it's progressing.
142
:So it's a very cool time to be alive.
143
:One of the other themes it's coming up
is this I think you're calling it
144
:or you call it closed loop logistics,
where there's this new market segment
145
:where people are just saying,
hey, I want my own piece.
146
:Can you walk me through
how you're thinking of that, that idea?
147
:Yeah, that the main point is
that it is a new market
148
:segment in geolocation and logistics.
149
:Overall.
150
:People tend to say, well,
what about tracking a truck?
151
:You know, that type of thing.
152
:So I use cellular or satellite,
and that used to be the logistics market.
153
:But today we're getting into use cases
that the most
154
:recent one I can quote
is an automotive manufacture.
155
:In the assembly chain,
they used to just have,
156
:you know, forge, walk, point, spare parts
and you know,
157
:how many spare parts
you're going to use per day.
158
:And it's it's fairly static.
159
:But now your trucks have e-fuels,
hybrid fuels.
160
:So different engines.
161
:So you get more and more options.
162
:So it's become impossible
to know in advance how many spare parts
163
:you know you need every hour.
164
:You need to measure it.
165
:You need to track your spare parts
throughout the factory.
166
:So these kind of use cases
where you need to track things inside
167
:and outside the factory,
or you build the engine in one factory
168
:and then you put the engine on the cradle,
you send it to another factory,
169
:but you want them to return the cradle
because the cradle is worth $1,000.
170
:There is a ton of use cases like this,
and it's it's fairly new.
171
:LoRa is the perfect thing because you want
something that works indoors,
172
:outdoors, very inexpensive,
and obviously the location itself
173
:is using the usual suspects
BLE sniffing, Wi-Fi sniffing GNSS outside.
174
:But but the glue the backhaul is LoRaWAN
and it's yeah, a very successful
175
:use case for us. Yeah.
176
:And this was
I know I talked to Julien over at Volvo
177
:about this a while ago
and on one of these shows, and I'm
178
:pretty sure they're using Actility.
179
:Yeah, Volvo was using it
to track the trucks throughout their
180
:assembly chain all the way to the parking,
but there are others.
181
:Finished vehicle logistics is another one.
182
:And now with all this tech stuff,
you get huge stockpiles of vehicles
183
:that were pushed through
before a tax increase
184
:or they left on the parking
because you can't sell them, you know.
185
:So there are massive parkings out there
with 20, 30,000 cars.
186
:And you need to find these cars
and you only have like what,
187
:eight different colors?
188
:You divide 20,000 by eight.
189
:It's still a lot of identical cars.
190
:So that's also a case for closed loop
logistics where LoRa is
191
:just the perfect technology.
192
:Oh, interesting.
193
:And do you think that we'll see kind of
custom tracker stuff like that, where
194
:it'll just be a magnet tracker
and you put it on the top of a car
195
:one minute,
and then as the car leaves the lot,
196
:you pull the tracker off
and put on the next thing.
197
:Or is that not really what you're saying?
198
:We do have this.
199
:We we have a subsidiary called Abeeway
that that makes obviously
200
:standard trackers.
201
:You know, I have one like here and
but indeed more and more we get customers
202
:who are obviously for a certain volume,
you know, like:
203
:They go to us with very specific
requirements.
204
:The finished vehicle logistics,
205
:for example, you want to have something
that flashes very strong
206
:so you can identify that specific car
here.
207
:Recently we've been working
with an animal reserve in Africa
208
:where they want to get a sort of emergency
SMS like signal. You.
209
:You pair your phone with the tracker
and if you lose the cellular signal,
210
:you can still have a primitive way of SMS
through the tracker.
211
:This kind of thing, we see more and more,
and it's actually easy to do
212
:because once you get the core
obviously sensor and LoRaWAN part,
213
:adding a few IOs here
and there is not so difficult.
214
:And in the end
215
:it's mostly the cost is mostly plastics
if you need new form factors.
216
:Yeah. And you spent time
growing up in Africa, right.
217
:So it must be nice for you
to kind of return back
218
:to, to that a little bit
and have a reason to pay attention to it.
219
:It is great.
220
:And there are two things in Africa
that are great,
221
:of course,
this kind of tracking use cases.
222
:The other thing is running
water is coming to Africa in the in
223
:it was of course in the main big cities,
but now it's coming to cities
224
:below 10,000 inhabitants
and it's a revolution.
225
:Here we'll talk about AMI
when when you and I talk around in Europe
226
:or the US about what it's all about,
or do you have that smart meter?
227
:But in Africa it's more interesting.
228
:It's, oh, do you have water?
229
:And there are utilities there
that grow 50% a year.
230
:And again,
it's a great use case for LoRaWAN there
231
:because most cellular networks
are still 2G very expensive.
232
:So utility don't bother
they just deploy their own infrastructure.
233
:Interesting.
234
:It's I mean I always remember this story
about Africa going straight
235
:to cell towers and never putting in,
you know, hard lines for phones.
236
:They may just go straight to LoRaWAN.
237
:It's kind of the sensorial net
for everything across the continent.
238
:They do.
239
:And Ivory Coast
is one of the leading countries.
240
:They're very interesting
with what they're doing. It's
241
:and yeah, we, we see lots of, you know,
it's it would be an economic podcast.
242
:Everyone says Africa is probably
the next high growth economy in IoT.
243
:It's also true
244
:I think actually we spend a lot of time in
Africa, a lot of attention, because it's
245
:I think it's going to be fairly
explosive there on basic infrastructure.
246
:I interesting
because it always seemed to me
247
:like they just that the capital depth
isn't there to manage
248
:or to support that, but it sounds like the
maybe the raw resources.
249
:But you know what, LoRa is inexpensive.
250
:The budget to do a nationwide network
is, what, 15-20 million.
251
:It's yeah.
252
:For utilities
nothing. It's it's pocket money.
253
:So it's not the issue
actually that the cost point or LoRa makes
254
:it really suitable for,
emerging economies.
255
:Oh very cool.
256
:So you I mean, you've been in LoRaWAN
since the beginning.
257
:You helped write
the standard and founded the whole thing.
258
:If you could maybe rewind the clock
a little bit or think of it,
259
:is some 20 year
260
:old kid comes up to you and says, hey,
I want to start a company in LoRaWAN.
261
:What should I be looking at
or paying attention to that
262
:I should build in the next couple of years
where would you direct them?
263
:Oh, that's a tough one.
264
:I think now it's mature enough
that you no longer
265
:really talk about LoRaWAN in general.
266
:You need to
you need to focus on your use case.
267
:Okay.
268
:So my yeah, my advice would probably pick
269
:a problem
that LoRaWAN can solve that you can solve.
270
:No. Together with a sensor.
271
:And I'm getting very excited
on this custom sensing thing.
272
:Actually, if I would start something.
273
:No, that's probably what I would do,
because there is an enormous legacy
274
:industry of sensing all wired,
275
:and they're all coming to wireless,
so they all need help.
276
:So I think there is a huge market
to to just go around,
277
:visit existing sensing companies
and help them go wireless.
278
:That's probably
one of the things I would push for.
279
:And is that something
where you would basically say like, hey,
280
:you cut the old wires and you add
in this LoRaWAN radio with the battery
281
:or are you saying like, hey,
282
:you would just create an entirely
new sensor to replace the old one?
283
:If you take metering, it's a good example.
284
:You take an existing protocol like DLMS.
285
:It's really more the application layer.
286
:That's the problem.
287
:Changing just the PHY layer
is really not the problem at all.
288
:But when you're wired,
you tend to be talkative.
289
:Not if you have a very old sensor
and actually many of those belong
290
:to the very old category
where they are still using RS485,
291
:you know, very slow connections for these.
292
:The switch to LoRaWAN
is actually fairly easy,
293
:but if you were used to cellular,
for example, the switch
294
:from cellular to LoRaWAN
is a bit more painful.
295
:Street
lighting is a perfect example of this,
296
:but most of the time when you look
at these very fat application layers,
297
:a lot of XML, a lot of JSON, it's 99% fat.
298
:You know, sometimes you just get one bit
in the middle of a big JSON
299
:so that the consulting part of
it is just say, okay, well, you're
300
:actually just interested in that one bit.
301
:Well guess what, just transmit that bit,
get rid of the rest,
302
:you reconstruct the rest.
303
:So it's a bit of consultative usually not
not too difficult most of the time.
304
:And it's it's part of the adaptation work.
305
:Very cool.
306
:And I think we, we end up with this thing
that I don't hear a bunch about.
307
:Every so often I'll get a hint of it
308
:on LinkedIn or something,
which is LoRaWAN and the military.
309
:What are you seeing with with military
as far as them using LoRaWAN?
310
:Like how are they using it?
What are they thinking of?
311
:What does that look like?
312
:We've seen obviously this is not
the mainstream technology in the military,
313
:but we see it more and more.
314
:Perhaps the most logical use case
that I can't disclose the where and who,
315
:but I can tell you it's active together
in live Military bases is way to truck
316
:the people,
the vehicle and the tools on an air base.
317
:I have been in the military, I think
you have to, if you're on a big air base,
318
:there are thousands of people
going around.
319
:And today
that the security, when a visitor gets
320
:in, you get the paper badge
and you feel secure.
321
:Okay? He gets his badge right.
322
:He can go anywhere.
323
:Obviously, no real security behind it.
324
:So now in some cases, they actually get
the badge that tracks them around.
325
:Of course you can take it off, but
if you take it off, it's stopped moving.
326
:And guess what? That's an alarm.
327
:If you get to a place where you shouldn't
be going, you have an alarm.
328
:Same for vehicles.
You get to bus in the base.
329
:The bus is supposed to go to that place,
330
:or this car to that place,
and back at that place and nowhere else.
331
:So more and more we see this kind of
332
:additional security layers
come up with electronics.
333
:The other thing is finding stuff around,
and you get great videos
334
:of Cisco on that, because obviously
there are a lot in the US Army base.
335
:And you would expect that if certifies or,
you know, it's all sci fi around it,
336
:but that descriptions
of how in the real world
337
:there are still pliers,
there are still tools.
338
:And it's like, you know, very, very human.
339
:If you forget the screwdriver
and you might have forgotten it
340
:in an airplane engine,
that's worth a billion.
341
:Some of these air bases
they get shut entirely, no one gets out,
342
:no one gets in
until you find that damn screwdriver.
343
:So more and more you get tracking for
in also in these bases to
344
:to make sure you get less and less of
this lost item could be a weapon,
345
:could be a $100,000 tool, or
it could be a $10 tool in the wrong place.
346
:And you just want to find it.
347
:Yeah.
348
:Oh interesting. Okay.
349
:So it's really the application of of
wireless, IoT, you know across the board.
350
:Okay.
351
:If we then kind of finish up
by this idea of
352
:LoRaWAN is is still growing
IoT is still growing.
353
:We're seeing this pretty amazing
coming together of AI and IoT,
354
:where IoT is providing
basically like the nervous network,
355
:the nervous system of the world
where AI can say, okay, now
356
:I can see this thing,
I can sense this thing.
357
:They can start to really interact
with the physical world.
358
:What do you think it looks like
when we're done?
359
:What do you think it looks like
when kind of the IoT
360
:is really fully developed
across the world?
361
:How does that change our lives?
362
:Obviously, there is a lot of optimization
that will go on.
363
:I can give you an example
about integration of renewable energy.
364
:We all hear that, okay.
365
:You get a lot of wind
farms, a lot of solar.
366
:And that's a problem because, you know,
you get more energy, less energy.
367
:How do you adapt to this randomness.
368
:And very often it's presented
like an almost unsolvable problem.
369
:But guess what?
370
:If tomorrow
all your fridges, all our fridges,
371
:all our electric cars receive a signal?
372
:Okay, now energy is a lot cheaper.
373
:Just go on.
374
:I, you know, turn on the compressor right
now, charge the car a little faster. Now
375
:the problem is gone.
376
:There is no investment behind it.
377
:You know, you don't need to invest
in a ton of new batteries or whatever.
378
:You just need to make the existing cars,
the existing charging systems of cars
379
:smarter or AC,
you know, air conditioning is coming.
380
:In the US It's everywhere already.
381
:But in the rest of the world, France
included.
382
:It's coming
now. All these heat pumps, same.
383
:You know,
you control that heat pump by half degree.
384
:Up or down.
385
:You get an enormous energy buffer
that can valorize
386
:additional energy or vice versa, release
more energy to the grid when needed.
387
:There are tons of examples like this
where it's just way more efficient,
388
:a lot less weight,
a lot less costs than than today.
389
:And the other thing
is, when it comes to AI today,
390
:setting up an alarm, things like,
you know, if the temperature becomes
391
:more than 25 degrees for more
and so on, it's an extra
392
:thing and people are lazy
and you just don't do it.
393
:I think tomorrow you'll just be able
and actually we'll doing some trials here.
394
:You just talk to and I say, okay,
my idea of an alarm is that,
395
:you know, it's more or less this and,
and it creates the rules by itself.
396
:I think it will also make it easier
and more massive to,
397
:to just create the smart things
by talking to some AI system.
398
:Yeah.
399
:It is super cool to think that, you know,
we started the conversation
400
:talking about BACnet and LoRaWAN
why and how there's
401
:this kind of limited amount of people
that understand the entirety of it.
402
:And it's it's becoming easier is really
these LLMs are allowing us
403
:just to, to start to speak to our world
and to speak to the world of IoT.
404
:And it's a pretty, pretty cool time
to be alive and to be experiencing that.
405
:Indeed.
406
:And it's, you know,
when you put the BACnet and the age of it
407
:and AI and you can bridge this, this.
408
:to, it's an interesting perspective
through time.
409
:Yeah. Yeah, it's very cool.
410
:Hey, I know you're super busy.
411
:I really appreciate you
making the time to come on.
412
:Thank you so much for coming on
and talking with us
413
:on The Business of LoRaWAN.
It was great talking to you Nik.
414
:That's it for
this episode of The Business of LoRaWAN.
415
:I built this for you.
416
:So whether you're a business owner,
a lawyer, one professional or a hobbyist,
417
:the intent is to give you great LoRaWAN
information.
418
:Of course,
the best information doesn't come from me.
419
:It comes from the conversations
420
:we have with the people building
and deploying this tech in the real world.
421
:And that's where you come in.
422
:LoRaWAN is a global
patchwork of talent and ideas.
423
:And ironically,
for a globally connected network,
424
:most of the brilliant folks working on it
aren’t connected yet.
425
:Help me change that.
426
:Introduce me
to someone awesome in your network,
427
:someone doing meaningful work in LoRaWAN,
or just shoot me a name.
428
:I'll take it from there
429
:and get them on the show
so we can share their work with the world.
430
:You can always find me at metsci.show.
431
:That's M-E-T-S-C-I dot
432
:S-H-O-W, metsci.show.
433
:If you want to support the show
in other ways, you can subscribe,
434
:leave a review,
share it with your corner of the world.
435
:All those are super helpful.
436
:If you'd like to support financially,
you can go to support.metsci.show
437
:for both one time and recurring options.
438
:We're also open to sponsors.
439
:If your company serves
the LoRaWAN community
440
:and you want to reach this dedicated
audience, let's talk.
441
:If you want to try LoRaWAN for yourself,
create a MeteoScientific account
442
:at console.meteoscientific.com
and get your first 400 DC for free,
443
:which is enough to run a device
sending hourly for about a year.
444
:This show is supported
445
:by a grant from the Helium Foundation
and produced by Gristle King, Inc..
446
:I'm Nik Hawks.
I'll see you on the next show.