Ci a ion: Van Ren e ghem, T.; Le
Bescond, V.; Dekoninck, L.;
Bo eldoo en, D. Ad anced Noise
Indica o Mapping Relying on a Ci y
Mic ophone Ne wo k. Senso s 2023,
23, 5865. h ps://doi.o g/
10.3390/s23135865
Academic Edi o : Hec o
Edua do Roman
Recei ed: 25 May 2023
Re ised: 16 June 2023
Accep ed: 21 June 2023
Published: 24 June 2023
Copy igh : © 2023 by he au ho s.
Licensee MDPI, Basel, Swi ze land.
This a icle is an open access a icle
dis ibu ed unde he e ms and
condi ions o he C ea i e Commons
A ibu ion (CC BY) license (h ps://
c ea i ecommons.o g/licenses/by/
4.0/).
senso s
A icle
Ad anced Noise Indica o Mapping Relying on a Ci y
Mic ophone Ne wo k
Timo hy Van Ren e ghem 1,* , Valen in Le Bescond 2, Luc Dekoninck 1and Dick Bo eldoo en 1
1WAVES Resea ch G oup, Depa men o In o ma ion Technology, Ghen Uni e si y, Technologiepa k 126,
B 9052 Gen -Zwijnaa de, Belgium
2Join Resea ch Uni in En i onmen al Acous ics (UMRAE), Cen e o S udies on Risks, Mobili y, Land
Planning and he En i onmen (CEREMA) and Uni e si y Gus a e Ei el, F-44344 Bouguenais, F ance
*Co espondence: imo hy. an en e [email p o ec ed]
Abs ac :
In his wo k, a me hodology is p esen ed o ci y-wide oad a ic noise indica o mapping.
The need o di ec access o a ic da a is bypassed by elying on s ee ca ego iza ion and a ci y
mic ophone ne wo k. The s a ing poin o he de e minis ic modeling is a p e iously de eloped
bu simpli ied dynamic a ic model, he la e necessa y o p edic s a is ical and dynamic noise
indica o s and o es ima e he numbe o noise e en s. The sound p opaga ion module combines
aspec s o he CNOSSOS and QSIDE models. In he nex s ep, a machine lea ning echnique—an
a i icial neu al ne wo k in his wo k—is used o weigh he ou comes o he de e minis ic p edic ions
o a ious a ic pa ame e scena ios (linked o s ee ca ego ies) o app oach he measu ed indica o s
om he mic ophone ne wo k. Applica ion o he ci y o Ba celona showed ha he di e ences
be ween p edic ions and measu emen s ypically lie wi hin 2–3 dB, which should be posi ioned
ela i e o he 3 dB a ia ion in s ee -side measu emen s when mic ophone posi ioning ela i e
o he açade is no ixed. The numbe o e en s is p edic ed wi h 30% accu acy. Indica o s can be
p edic ed as a e ages o e day, e ening and nigh pe iods, bu also a an hou ly scale; sho e ime
pe iods do no seem o nega i ely a ec modeling accu acy. The cu en me hodology opens he way
o include a b oad se o noise indica o s in ci y-wide en i onmen al noise impac assessmen .
Keywo ds:
noise moni o ing ne wo ks; mic ophones; oad a ic noise; en i onmen al noise
mapping
;
noise indica o s
1. In oduc ion
Road a ic is commonly he main sou ce o exposu e o en i onmen al noise in
Eu opean ci ies [
1
]. A basic s ep in oad a ic noise mapping is gaining access o a ic
pa ame e s such as a ic in ensi y, ehicle speed, accele a ion and a ic composi ion on
each oad segmen in he ne wo k [
2
]. Howe e , mos a ic models ocus on majo oads
only o pe o m conges ion analysis du ing ush hou s [
3
]. On smalle oads, in con as ,
a ic da a a e mos o en lacking. Al hough his migh be in line wi h he En i onmen al
Noise Di ec i e [
4
] in Eu ope, s ipula ing ha noise maps should only be p oduced om
55 dB(A) Lden
on, his is ne e heless p oblema ic in iew o ci y-wide noise mapping.
When assessing human sleep dis u bance due o noise, exposu e mapping becomes e en
mo e challenging and should go down o le els as low as 40 dB(A) Lnigh .
Knowledge o less exposed zones is ele an as well since hese zones should be o
p ima y in e es o u u e esiden ial de elopmen s and a e po en ially es o a i e places
in a ci y. Fu he mo e, only mapping exposu e in pa o a ci y could in oduce bias in
en i onmen al jus ice s udies, an impo an conce n nowadays when making sus ainable
ci ies (see, e.g., [5]).
An in e es ing line o esea ch showed ha s ee ca ego iza ion in a ci y is able o
es ima e s ee -side exposu e le els easonably well [
6
–
11
], possibly accompanied wi h
limi ed se s o snapsho measu emen s, whe e e o s can be minimized by sui ed sampling
Senso s 2023,23, 5865. h ps://doi.o g/10.3390/s23135865 h ps://www.mdpi.com/jou nal/senso s
Senso s 2023,23, 5865 2 o 19
s a egies [
6
,
12
]. Simila ly, oadside noise measu emen s we e shown o be able o ade-
qua ely p edic he unde lying oad a ic pa ame e s such as ehicle speed, a ic in ensi y
and he sha e o hea y ehicles [
13
]. In he open GIS ini ia i e Open S ee Map (OSM),
e e y oad in a ci y is p esen and assigned a speci ic ca ego y. This opens possibili ies o
ull ci y noise mapping, including low(e ) exposu e zones.
Linking noise exposu e maps o human heal h e ec s is cu en ly no e y success ul.
Fo an impo an noise policy indica o such as sel - epo ed noise annoyance, less han
30% o he obse ed a iance ound in a su eyed popula ion is ac ually cap u ed [
14
].
A possible eason o his low p edic i e powe is ha cu en noise mapping ini ia i es
ocus on long- e m equi alen sound p essu e le els only. This unde mines a noise map
as an e icien u ban sound planning ins umen . Only ecen ly, he sho comings o he
commonly used ene ge ically equi alen le els ha e been o icially acknowledged [15].
The way people pe cei e en i onmen al noise is much mo e complex han wha can be
quan i ied wi h hese s anda d ene ge ically a e aged sound p essu e le els. A wide se
o noise indica o s and psycho-acous ical indica o s ha e long been used in o he con ex s,
e.g., in p oduc design [
16
] and soundscape s udies [
17
,
18
]. Essen ially, people a e e y
good lis ene s, and e en sub le changes in he spec o- empo al con en o a sound migh
impac he pe cep ion and eac ion o i . Noise indica o s o po en ial in e es a e s a is ical
sound p essu e le els, he numbe o e en s and indica o s desc ibing he dynamic na u e
o u ban sound. Cu en ly, ci y-wide mapping o such noise indica o s is e y sca ce. A ew
measu emen -based ini ia i es can be ound, whe e walke s equipped wi h mic ophones
scan a pa icula ci y qua e [
19
–
21
]. The use o hese mo e ad anced indica o s o be e
p edic he impac s o en i onmen al noise is cu en ly unde explo ed.
In his wo k, a me hodology is desc ibed o p edic bo h equi alen sound p essu e le els
and a wide ange o o he noise indica o s, by means o de e minis ic noise modeling, whe e
he accu acy o he p edic ions is imp o ed by i ing o long- e m measu emen s o a ci y noise
moni o ing ne wo k in a inal s ep. The s a e-o - he-a de e minis ic noise modeling p ocedu e,
acili a ing he calcula ion o dynamic noise indica o s, is desc ibed in b ie in Sec ion 2. The
p oposed me hodology is illus a ed o he ci y o Ba celona (Spain) in Sec ion 3, whe e a
ci y-wide mic ophone ne wo k has been ope a ional o mo e han a decade.
2. De e minis ic Noise Mapping P ocedu e
2.1. Linking T a ic Da a and Open S ee Map Road Ca ego iza ion
S ee ca ego iza ion da a we e di ec ly used om Open S ee Map (OSM). Each
s ee ca ego y was assigned a se o plausible a ic pa ame e s (mo e p ecisely a ic
in ensi y, ehicle speed and sha e o hea y ehicles). This assignmen s a s om exis ing
(highway) a ic coun da abases, and i is ensu ed ha he expec ed logics such as a lowe
a ic in ensi y, lowe ehicle speed and a lowe sha e o hea y ehicles on mino s ee s
compa ed o majo s ee s a e p esen . Depending on he de e minis ically p edic ed noise
indica o s co esponding o a gi en scena io, addi ional scena ios we e manually added.
In o al, 15 scena ios we e used (see Appendix A o an o e iew o he pa ame e se ings).
In a inal s ep, he calcula ed ou comes o a wide se o noise indica o s a e weigh ed
o minimize he di e ence wi h measu emen om he mic ophone ne wo k, as will be
discussed in Sec ion 3.2.
2.2. Dynamic T a ic Model
Simpli ied ehicle mo emen s a e modeled based on he hou ly a e aged numbe o
ehicles and hei speeds. Vehicles a e launched on a oad segmen a a ixed speed. When
eaching he end o ha segmen , he ehicle is emo ed om he simula ion, meaning
he e is no ehicle ans e om one segmen o ano he . The in e - ehicle imes espec a
Poisson dis ibu ion, and ehicle speeds o he di e en ca s ollow a no mal dis ibu ion.
A he end o he simula ed hou , a each oad segmen , he (s a ic) ehicle coun s and
a e age speeds a e espec ed. Mo e in o ma ion on his simpli ied mic o-simula ion a ic
p ocedu e can be ound in [22]. A ime s ep o 1 s was conside ed in his wo k.
Senso s 2023,23, 5865 3 o 19
2.3. T a ic Noise Emission Model
Vehicle ca ego y, numbe o ehicles pe hou , and a e age speed a e used as inpu o
he CNOSSOS [
23
] oad a ic acous ic emission model. These a ic- ela ed inpu s come
om he dynamic a ic modeling p ocedu e desc ibed in Sec ion 2.2.
2.4. Sound P opaga ion Model
The sound p opaga ion modeling p ocedu e combines he CNOSSOS sound p opaga-
ion model [
23
] wi h aspec s om he QSIDE u ban sound p opaga ion model [
24
]. Vehicles
close o a ecei e (wi hin a adius o 500 m) a e ea ed di e en ly om hose u he away
(be ween 500 m and a maximum 2000 m).
A close dis ance, and when a di ec line-o -sigh p opaga ion pa h is possible be ween
a sou ce and a ecei e , geome ical di e gence, g ound e ec and a mosphe ic abso p ion
a e included ollowing he CNOSSOS sound p opaga ion model, implemen ed in he
open access NoiseModelling amewo k [
25
,
26
]. Only in he absence o a line-o -sigh
p opaga ion pa h, di ac ions a ound ho izon al edges and e lec ions on e ical objec s
a e accoun ed o . The maximum e lec ion o de is 2, and he maximum sou ce- e lec ion
dis ance is 50 m (which a e s anda d se ings; see, e.g., [
27
]). A s anda d noise mapping
ecei e heigh o 4 m is used.
Sca e ing on a mosphe ic u bulence is added o he a enua ion ac o s o a oid le els
becoming un ealis ically low, especially behind objec s. The QSIDE enginee ing sca e ing
app oach [
28
] was used, p o iding an easy- o-e alua e exp ession adap ed o he u ban
en i onmen , accoun ing o sound equency, p opaga ion dis ance, s ee canyon geome y
and u bulence s eng h. Al hough he model could include local s ee canyon geome y in
de ail, s anda d building heigh s and wid hs we e used o a oid ime-demanding e ie al
om geog aphical inpu da a. Tu bulence s uc u e pa ame e s C
2
and C
T2
(see Table 1)
depend on whe he he p opaga ion occu s in u al/subu ban se ings o in he dense u ban
ab ic, and du ing he day ime o a nigh . These u bulence pa ame e s a e based on long-
e m obse a ions o e la u al zones wi h dispe sed smalle ci ies [
29
]; in he dense u ban
en i onmen , u bulence s eng h is doubled in a simpli ied app oach.
Table 1. O e iew o he sound p opaga ion models and pa ame e se ings.
Close-by T a ic Fa T a ic
Radius a ound ecei e (in m) <500 ≥500 and <2000
T a ic (noise emission) modeling Simpli ied dynamic a ic modeling
ollowing [22], a a 1 s ime in e al.
Agg ega ed a ic a disc e e emission
poin s. Numbe o emission poin s
minimized by NoiseModelling [26]
I a di ec line-o -sigh pa h is possible CNOSSOS sound p opaga ion model [23] wi hou e lec ions on e ical objec s,
wi hou di ac ions, and in a non- e ac ing a mosphe e.
Only obs uc ed sound pa hs a e p esen
CNOSSOS sound p opaga ion model [23] including e lec ions on e ical objec s
( e lec ion o de 2, maximum sou ce- e lec ion dis ance 50 m) and including
di ac ions on ho izon al edges. Downwa d e ac ion (“ a o able condi ions”)
is assumed wi h 50% occu ence in any di ec ion.
Tu bulen sca e ing model [28]
Ru al/subu ban Dense u ban ab ic
Dis ance o açade (m) No applicable 5
Ci y canyon wid h (m) No applicable 15
Building heigh (m) 8 20
Day Nigh Day Nigh
C 2(m4/3/s2)0.4 0.2 0.8 0.4
CT2(K2/m2/3)0.7 0.04 1.4 0.08
In o de o cap u e dynamic noise indica o s and noise e en s, conside ing indi idual
nea by ehicles is essen ial. This is no he case anymo e o oad a ic u he away
Senso s 2023,23, 5865 4 o 19
con ibu ing mainly o he backg ound noise a a ecei e . This allows bundling he acous ic
ene gy o ca s in a limi ed numbe o emission poin s as op imized by he NoiseModelling
amewo k. Fo he p opaga ion simula ions, an app oach simila o ha o nea by a ic
is ollowed, so depending on whe he a line-o -sigh pa h is possible o no .
The CNOSSOS a o able sound p opaga ion app oach (i.e., downwa d e ac ion) is
only conside ed in he absence o line-o -sigh pa hs. In he CNOSSOS simpli ied cu ed
ay app oach, he di e ence be ween e ac ion/no e ac ion is mainly ele an in he case
o p opaga ion o e objec s. The p obabili y o a o able sound p opaga ion is hen se o
50% in any p opaga ion di ec ion.
3. The Ba celona Mic ophone Senso Ne wo k
3.1. Measu emen s and Da a Handling
The Ba celona mic ophone measu emen ne wo k is unique in i s kind due o i s size
( oughly 250 moni o ing poin s sp ead o e he ci y) and since i has been ope a ional
o mo e han a decade. The ne wo k con ains bo h ixed senso s and senso s ha a e
eposi ioned pe iod-wise o maximize he zone moni o ed. Toge he wi h he ac ha
indi idual senso s a e p one o acciden al ailu e, he da ase is a he discon inuous in
na u e. Ne e heless, a some ixed senso s, con inuous sound p essu e le el measu emen s
o e se e al yea s a e p esen .
The senso s a e oppo unis ically posi ioned, e.g., di ec ly a ached o window sills o
nea balconies. This means ha he ex en o which açade e lec ions impac he sound
p essu e le el measu emen s is no ixed (see Sec ion 4). Mic ophones a e always acing
he s ee s and a e ep esen a i e o he mos exposed building side.
To limi he impac o changes in he a ic ne wo k in as uc u e and i s managemen
(such as limi ing a ic in speci ic s ee s, changing he di ec ion o ci cula ion, ban o hea y
a ic), a 3-yea pe iod was selec ed, which was a comp omise be ween keeping his pe iod
as sho as possible and ha ing a su icien amoun o da a while keeping as many senso
loca ions as possible o p ocessing. Ne e heless, changes in he a ic ne wo k canno be
ully a oided wi hin his ime ame, and i his was he case, he measu emen s we e hen
he a e age be ween he wo di e en a ic si ua ions. No e ha con e gence mus s ill be
eached a such loca ions (see nex pa ag aph) o a mic ophone posi ion o be used.
The p ocessing o he measu emen senso s was pe o med as ollows. A basic ime
pe iod o 15 min was chosen o all indica o s. P e ious esea ch [
12
] showed ha his is a
sui able ime ame in oad a ic noise-domina ed u ban en i onmen s. Sho e pe iods
could lead o di icul ies in s abilizing he noise indica o s, gi ing oo much emphasis on
momen a y a ia ions. When ex ending o longe pe iods, he empo al a ia ions in he
sonic en i onmen migh no be cap u ed su icien ly.
Fo a senso loca ion o be conside ed in u he analysis, a leas 3 weeks o da a (no
necessa ily con inuous) should be a ailable. Weekends we e excluded o a oid uncommon
a ic si ua ions. As a simpli ied con e gence c i e ion, he di e ence be ween aking 80%
o he da a and all a ailable da a (in a ch onological way) should be less han 1 dB when
(linea ly) a e aging a noise indica o ha uses a decibel scale. Fo e en -based indica o s,
his c i e ion is se o i e e en s, and o he in e mi ency a io se o 5% (see u he ).
I his condi ion is no me , his senso loca ion is dis ega ded a leas o a speci ic ime
pe iod. Remo ing senso da a du ing he day pe iod, e.g., does no necessa ily mean ha
he senso loca ion is also dis ega ded du ing he e ening and nigh pe iods.
The measu emen ne wo k con ains senso s wi h wo le els o de ail. Mos mic o-
phone s a ions epo o al A-weigh ed sound p essu e le els wi h a basic in eg a ion pe iod
o 1 min. These da a we e a ailable a 93 senso s in he cu en s udy (see Sec ion 3.3), du -
ing he pe iod 2020–2021–2022. Secondly, measu emen s a ions logging 1/3-oc a e bands
wi h a basic in eg a ion pe iod o 1 s we e used du ing he pe iod 2016–2017–2018. These
mo e de ailed da a we e a ailable a 23 s a ions and allowed calcula ing mo e ad anced
noise indica o s as discussed in Sec ion 3.4.
Senso s 2023,23, 5865 5 o 19
3.2. Machine Lea ning Fi ing P ocedu e
As an example supe ised machine lea ning i ing algo i hm, an a i icial neu al
ne wo k was used, as implemen ed in Ma lab [
30
]. A s anda d spli in o aining, alida ion
and es se s using 70%, 15% and 15% o he da a, espec i ely, was chosen. The aining
algo i hm “Le enbe g–Ma qua d backp opaga ion” was used, which is ecommended as
a as , i s -choice p ocedu e [
30
]. To p e en o e i ing, only i e neu ons we e used, wi h
a single hidden laye [
31
]. Gi en he andom spli in o aining, alida ion and es da ase s,
models we e epea edly cons uc ed, allowing he use o a e aged model p edic ions and
gi ing an indica ion o con idence in e als on epea ed p edic ions.
In he case o p edic ing A-weigh ed equi alen sound p essu e le els (see Sec ion 3.3),
he inpu consis s o 93 loca ions
×
15 a ic scena ios; he e a e 93 (loca ions)
×
3 (day,
e ening and nigh ly a e aged) o 93 (loca ions)
×
24 (hou ly a e aged) ou pu s. In case
mo e ad anced noise indica o s we e included (see Sec ion 3.4), 23 (loca ions)
×
15 ( a ic
scena io)
×
29 (indica o s) inpu s we e used o p edic 23 (loca ions)
×
29 (indica o s)
×
3
(day, e ening and nigh ly a e aged) ou pu s.
This wo k does no aim a inding he mos accu a e o as es machine lea ning app oach
o his speci ic applica ion, bu a he showcases wha can be achie ed wi h a s anda d and
well-es ablished supe ised machine lea ning i ing app oach. Simila ly, u he op imiza ion
o he neu al ne wo k se ings is also beyond he scope o he cu en wo k.
3.3. P edic ing A-Weigh ed Equi alen Sound P essu e Le els
Using he basic L
Aeq,1min
alues, an in eg a ion is pe o med o 15 min. In he nex
s ep, L
Aeq,15min
da a a e linea ly a e aged o e day (7:00–19:00), e ening (19:00–23:00) and
nigh (23:00–7:00) pe iods, hus p o iding a ypical alue in each pe iod, and o m he basis
o he a i icial neu al ne wo k p edic ions. In a second se o p edic ions, hou ly a e aged
LAeq,15min da a a e used as well.
Figu es 1–3depic he 15 de e minis ic p edic ions a each senso loca ion ha o med
he basis o he weigh ing by he a i icial neu al ne wo k, oge he wi h he measu ed
alues (i.e., he g ound u h), he mean p edic ed alues and he 90 h and 10 h pe cen iles
based on epea ed model cons uc ions. On he ho izon al axis, he loca ion ID numbe is
used, which is an a bi a y numbe bu easily allows assessing changes om loca ion o
loca ion, bo h in measu emen s and p edic ions. Figu es a e shown o he daily, e ening
and nigh ly a e aged L
Aeq,15min
. No e ha senso s wi h an insu icien numbe o da a
poin s o senso s no leading o con e ged indica o s (see Sec ion 3.1) we e ob iously
no used du ing he cons uc ion o he machine lea ning model. Once he model was
cons uc ed, p edic ions wi h he model we e pe o med a all 93 senso loca ions. Clea ly,
only loca ions wi h bo h measu emen s and p edic ions we e conside ed in he subsequen
accu acy analysis.
In Figu es 4–6, he measu ed da a a e plo ed on he s ee map o Ba celona, complying
wi h he selec ion c i e ia (see Sec ion 3.1), he (mean) p edic ions a (all) senso loca ions,
and he di e ence be ween measu emen s and p edic ions whe e possible (as oo -mean-
squa e e o , RMSE). As an example, day ime da a only a e shown. A mos loca ions,
p edic ion e o s a e limi ed, al hough a ew poin s gi e ise o la ge e o s.
Senso s 2023,23, 5865 6 o 19
Senso s 2023, 23, x FOR PEER REVIEW 6 o 19
loca ions. Clea ly, only loca ions wi h bo h measu emen s and p edic ions we e consid-
e ed in he subsequen accu acy analysis.
Figu e 1. De e minis ic p edic ions o LAeq,15min o each single affic scena io, by he a i icial neu al
ne wo k (showing he mean p edic ion and he 90 h and 10 h pe cen iles, based on epea ed model
cons uc ions), oge he wi h he con e ged measu emen s, a each o he 93 loca ions whe e a sen-
so node is/was ope a ional. Da a shown he e a e he linea ly a e aged LAeq,15min du ing he day ime.
Figu e 2. De e minis ic p edic ions o LAeq,15min o each single affic scena io, by he a i icial neu al
ne wo k (showing he mean p edic ion and he 90 h and 10 h pe cen iles, based on epea ed model
cons uc ions), oge he wi h he con e ged measu emen s, a each o he 93 loca ions whe e a sen-
so node is/was ope a ional. Da a shown he e a e he linea ly a e aged LAeq,15min du ing he e ening.
Figu e 1.
De e minis ic p edic ions o L
Aeq,15min
o each single a ic scena io, by he a i icial neu al
ne wo k (showing he mean p edic ion and he 90 h and 10 h pe cen iles, based on epea ed model
cons uc ions), oge he wi h he con e ged measu emen s, a each o he 93 loca ions whe e a senso
node is/was ope a ional. Da a shown he e a e he linea ly a e aged L
Aeq,15min
du ing he day ime.
Senso s 2023, 23, x FOR PEER REVIEW 6 o 19
loca ions. Clea ly, only loca ions wi h bo h measu emen s and p edic ions we e consid-
e ed in he subsequen accu acy analysis.
Figu e 1. De e minis ic p edic ions o LAeq,15min o each single affic scena io, by he a i icial neu al
ne wo k (showing he mean p edic ion and he 90 h and 10 h pe cen iles, based on epea ed model
cons uc ions), oge he wi h he con e ged measu emen s, a each o he 93 loca ions whe e a sen-
so node is/was ope a ional. Da a shown he e a e he linea ly a e aged LAeq,15min du ing he day ime.
Figu e 2. De e minis ic p edic ions o LAeq,15min o each single affic scena io, by he a i icial neu al
ne wo k (showing he mean p edic ion and he 90 h and 10 h pe cen iles, based on epea ed model
cons uc ions), oge he wi h he con e ged measu emen s, a each o he 93 loca ions whe e a sen-
so node is/was ope a ional. Da a shown he e a e he linea ly a e aged LAeq,15min du ing he e ening.
Figu e 2.
De e minis ic p edic ions o L
Aeq,15min
o each single a ic scena io, by he a i icial neu al
ne wo k (showing he mean p edic ion and he 90 h and 10 h pe cen iles, based on epea ed model
cons uc ions), oge he wi h he con e ged measu emen s, a each o he 93 loca ions whe e a senso
node is/was ope a ional. Da a shown he e a e he linea ly a e aged LAeq,15min du ing he e ening.
Senso s 2023,23, 5865 7 o 19
Senso s 2023, 23, x FOR PEER REVIEW 7 o 19
Figu e 3. De e minis ic p edic ions o L
Aeq,15min
o each single affic scena io, by he a i icial neu al
ne wo k (showing he mean p edic ion and he 90
h
and 10
h
pe cen iles, based on epea ed model
cons uc ions), oge he wi h he con e ged measu emen s, a each o he 93 loca ions whe e a sen-
so node is/was ope a ional. Da a shown he e a e he linea ly a e aged L
Aeq,15min
du ing he nigh .
In Figu es 4–6, he measu ed da a a e plo ed on he s ee map o Ba celona, com-
plying wi h he selec ion c i e ia (see Sec ion 3.1), he (mean) p edic ions a (all) senso
loca ions, and he diffe ence be ween measu emen s and p edic ions whe e possible (as
oo -mean-squa e e o , RMSE). As an example, day ime da a only a e shown. A mos
loca ions, p edic ion e o s a e limi ed, al hough a ew poin s gi e ise o la ge e o s.
Figu e 4. Linea ly a e aged L
Aeq,15min
om measu emen s du ing day ime. Only senso loca ions
wi h con e ged measu emen s and da a alling wi hin he p e-selec ed ime ame a e shown.
Figu e 3.
De e minis ic p edic ions o L
Aeq,15min
o each single a ic scena io, by he a i icial neu al
ne wo k (showing he mean p edic ion and he 90 h and 10 h pe cen iles, based on epea ed model
cons uc ions), oge he wi h he con e ged measu emen s, a each o he 93 loca ions whe e a senso
node is/was ope a ional. Da a shown he e a e he linea ly a e aged LAeq,15min du ing he nigh .
Senso s 2023, 23, x FOR PEER REVIEW 7 o 19
Figu e 3. De e minis ic p edic ions o L
Aeq,15min
o each single affic scena io, by he a i icial neu al
ne wo k (showing he mean p edic ion and he 90
h
and 10
h
pe cen iles, based on epea ed model
cons uc ions), oge he wi h he con e ged measu emen s, a each o he 93 loca ions whe e a sen-
so node is/was ope a ional. Da a shown he e a e he linea ly a e aged L
Aeq,15min
du ing he nigh .
In Figu es 4–6, he measu ed da a a e plo ed on he s ee map o Ba celona, com-
plying wi h he selec ion c i e ia (see Sec ion 3.1), he (mean) p edic ions a (all) senso
loca ions, and he diffe ence be ween measu emen s and p edic ions whe e possible (as
oo -mean-squa e e o , RMSE). As an example, day ime da a only a e shown. A mos
loca ions, p edic ion e o s a e limi ed, al hough a ew poin s gi e ise o la ge e o s.
Figu e 4. Linea ly a e aged L
Aeq,15min
om measu emen s du ing day ime. Only senso loca ions
wi h con e ged measu emen s and da a alling wi hin he p e-selec ed ime ame a e shown.
Figu e 4.
Linea ly a e aged L
Aeq,15min
om measu emen s du ing day ime. Only senso loca ions
wi h con e ged measu emen s and da a alling wi hin he p e-selec ed ime ame a e shown.
Senso s 2023,23, 5865 8 o 19
Senso s 2023, 23, x FOR PEER REVIEW 8 o 19
Figu e 5. Mean L
Aeq,15min
p edic ions du ing day ime, a 93 spo s whe e a senso node is/was ope a-
ional.
Figu e 6. Roo -mean-squa e e o (RMSE) be ween measu ed and (mean) p edic ed L
Aeq,15min
du ing
day ime. Only senso loca ions wi h con e ged measu emen s and da a alling wi hin he p e-se-
lec ed ime ame we e used o his analysis.
The his og ams in Figu e 7 depic he ac ual diffe ences be ween measu emen s and
p edic ions, showing ha he ze o e o class is mos popula ed in all ime pe iods con-
side ed. Du ing he nigh , he dis ibu ion is s ill symme ical, bu he sp ead seems some-
wha la ge . The RMSEs a e 2.0 dB(A) du ing he day ime, 2.1 dB(A) du ing he e ening
and 3.3 dB(A) du ing he nigh .
Figu e 5.
Mean L
Aeq,15min
p edic ions du ing day ime, a 93 spo s whe e a senso node is/was
ope a ional.
Senso s 2023, 23, x FOR PEER REVIEW 8 o 19
Figu e 5. Mean L
Aeq,15min
p edic ions du ing day ime, a 93 spo s whe e a senso node is/was ope a-
ional.
Figu e 6. Roo -mean-squa e e o (RMSE) be ween measu ed and (mean) p edic ed L
Aeq,15min
du ing
day ime. Only senso loca ions wi h con e ged measu emen s and da a alling wi hin he p e-se-
lec ed ime ame we e used o his analysis.
The his og ams in Figu e 7 depic he ac ual diffe ences be ween measu emen s and
p edic ions, showing ha he ze o e o class is mos popula ed in all ime pe iods con-
side ed. Du ing he nigh , he dis ibu ion is s ill symme ical, bu he sp ead seems some-
wha la ge . The RMSEs a e 2.0 dB(A) du ing he day ime, 2.1 dB(A) du ing he e ening
and 3.3 dB(A) du ing he nigh .
Figu e 6.
Roo -mean-squa e e o (RMSE) be ween measu ed and (mean) p edic ed L
Aeq,15min
du ing
day ime. Only senso loca ions wi h con e ged measu emen s and da a alling wi hin he p e-selec ed
ime ame we e used o his analysis.
The his og ams in Figu e 7depic he ac ual di e ences be ween measu emen s and
p edic ions, showing ha he ze o e o class is mos popula ed in all ime pe iods consid-
e ed. Du ing he nigh , he dis ibu ion is s ill symme ical, bu he sp ead seems somewha
la ge . The RMSEs a e 2.0 dB(A) du ing he day ime, 2.1 dB(A) du ing he e ening and
3.3 dB(A) du ing he nigh .
Resul s o hou ly p edic ions a e depic ed in Figu e 8, shown as empo al pa e ns o e
24 h pe iods a each senso loca ion. The measu ed empo al pa e ns a e shown as well. This
igu e does no allow he compa ison o measu emen s and p edic ions a any indi idual
senso loca ion, bu i nicely shows ha he bulk o he empo al pa e ns a e well p edic ed.
Bo h loca ions wi h a a he la pa e n and hose wi h s onge le el d ops du ing he nigh
Senso s 2023,23, 5865 9 o 19
hou s can be dis inguished, bo h in he measu emen s and p edic ions. Di ec ly ela ed o
Figu e 8, Figu e 9shows he hou ly RMSEs. Minimum alues a e ound a ound noon, nea
2 dB(A), and inc ease sligh ly abo e 3 dB(A) be ween 3 and 4 o’clock a nigh .
Senso s 2023, 23, x FOR PEER REVIEW 9 o 19
Figu e 7. His og ams showing he diffe ence be ween he (mean) p edic ed and measu ed LAeq,15min,
linea ly a e aged o e day ime, e ening and nigh hou s.
Resul s o hou ly p edic ions a e depic ed in Figu e 8, shown as empo al pa e ns
o e 24 h pe iods a each senso loca ion. The measu ed empo al pa e ns a e shown as
well. This igu e does no allow he compa ison o measu emen s and p edic ions a any
indi idual senso loca ion, bu i nicely shows ha he bulk o he empo al pa e ns a e
well p edic ed. Bo h loca ions wi h a a he la pa e n and hose wi h s onge le el d ops
du ing he nigh hou s can be dis inguished, bo h in he measu emen s and p edic ions.
Di ec ly ela ed o Figu e 8, Figu e 9 shows he hou ly RMSEs. Minimum alues a e ound
a ound noon, nea 2 dB(A), and inc ease sligh ly abo e 3 dB(A) be ween 3 and 4 o’clock
a nigh .
Figu e 8. Hou ly empo al pa e ns o LAeq,15min a all 93 measu emen loca ions. The measu emen s
a e shown oge he wi h he mean p edic ions based on epea ed model cons uc ion. Da a shown
he e a e he linea ly a e aged LAeq,15min du ing a speci ic hou . In e up ed lines indica e hou s whe e
measu emen s a e no con e ged due o an insufficien amoun o da a.
Figu e 7.
His og ams showing he di e ence be ween he (mean) p edic ed and measu ed L
Aeq,15min
,
linea ly a e aged o e day ime, e ening and nigh hou s.
Senso s 2023, 23, x FOR PEER REVIEW 9 o 19
Figu e 7. His og ams showing he diffe ence be ween he (mean) p edic ed and measu ed LAeq,15min,
linea ly a e aged o e day ime, e ening and nigh hou s.
Resul s o hou ly p edic ions a e depic ed in Figu e 8, shown as empo al pa e ns
o e 24 h pe iods a each senso loca ion. The measu ed empo al pa e ns a e shown as
well. This igu e does no allow he compa ison o measu emen s and p edic ions a any
indi idual senso loca ion, bu i nicely shows ha he bulk o he empo al pa e ns a e
well p edic ed. Bo h loca ions wi h a a he la pa e n and hose wi h s onge le el d ops
du ing he nigh hou s can be dis inguished, bo h in he measu emen s and p edic ions.
Di ec ly ela ed o Figu e 8, Figu e 9 shows he hou ly RMSEs. Minimum alues a e ound
a ound noon, nea 2 dB(A), and inc ease sligh ly abo e 3 dB(A) be ween 3 and 4 o’clock
a nigh .
Figu e 8. Hou ly empo al pa e ns o LAeq,15min a all 93 measu emen loca ions. The measu emen s
a e shown oge he wi h he mean p edic ions based on epea ed model cons uc ion. Da a shown
he e a e he linea ly a e aged LAeq,15min du ing a speci ic hou . In e up ed lines indica e hou s whe e
measu emen s a e no con e ged due o an insufficien amoun o da a.
Figu e 8.
Hou ly empo al pa e ns o L
Aeq,15min
a all 93 measu emen loca ions. The measu emen s
a e shown oge he wi h he mean p edic ions based on epea ed model cons uc ion. Da a shown
he e a e he linea ly a e aged L
Aeq,15min
du ing a speci ic hou . In e up ed lines indica e hou s
whe e measu emen s a e no con e ged due o an insu icien amoun o da a.
Senso s 2023,23, 5865 16 o 19
ob ained, in excess o he de ia ion amongs e e ence mic ophones hemsel es [
49
]. Mo e
ecen de elopmen s and expe ience wi h MEMS mic ophones [
50
–
52
] could u he boos
he deploymen o ci y-wide mic ophone ne wo ks. The implici a ic da a e ie al in
he cu en wo k could u he bene i om including compu e ision echnologies [
53
].
In [54], e.g., came a images we e di ec ly used o noise mapping using machine lea ning.
A ele an ques ion is whe he he ained i ing ne wo k could be ans e able o
o he ci ies. I is expec ed ha his is unlikely, since he link be ween s ee ca ego ies and
a ic pa ame e s could be s ongly locally dependen . In addi ion, as discussed be o e,
he ne wo k migh no only weigh he a ic scena ios, bu also a ic and p opaga ion-
ela ed aspec s a e co ec ed o . Examples a e he ypical s ee wid hs and numbe o
lanes co esponding o a speci ic s ee ca ego y in a ci y, a ic managemen policy, and
p e e ed oad su aces and hei main enance.
5. Conclusions
The p oposed ad anced noise indica o mapping p ocedu e, using a se o de e min-
is ic p edic ions combined wi h da a om a ci y mic ophone measu emen ne wo k, has
been shown o be an app oach wi h high po en ial. Bo h equi alen sound p essu e le els
and mo e ad anced noise indica o s exp essed in decibel uni s lead o RMSEs be ween 2
and 3 dB. These de ia ions should be posi ioned ela i e o he 3 dB a ia ion in s ee -side
u ban oad a ic noise exposu e measu emen s when he mic ophone posi ioning ela i e
o he açade is no ixed. The cu en wo k u he shows ha ci y-wide noise mapping
wi hou access o di ec a ic da a is easible on he condi ion ha a mic ophone ne wo k
is a ailable, and a he same ime, sys ema ic inaccu acies occu ing a any s age du ing
he de e minis ic modeling p ocess migh be implici ly co ec ed o , a leas o some ex en .
Con inued esea ch and mo e case s udies a e needed o see whe he he cu en concep
can g ow o a ma u e u ban a ic noise mapping me hodology.
Au ho Con ibu ions:
Concep ualiza ion, D.B. and T.V.R.; me hodology, D.B., T.V.R., L.D. and
V.L.B.; so wa e, V.L.B., T.V.R., D.B. and L.D.; alida ion, T.V.R.; o mal analysis, T.V.R.; in es iga ion,
V.L.B., D.B., T.V.R. and L.D.; esou ces, D.B. and T.V.R.; da a cu a ion, T.V.R.; w i ing—o iginal d a
p epa a ion, T.V.R.; w i ing— e iew and edi ing, T.V.R., V.L.B., D.B. and L.D.; isualiza ion, T.V.R.;
supe ision, T.V.R. and D.B.; p ojec adminis a ion, D.B.; unding acquisi ion, D.B. All au ho s ha e
ead and ag eed o he published e sion o he manusc ip .
Funding:
We acknowledge he unding om he Eu opean Union’s Ho izon 2020 Resea ch and
Inno a ion P og amme o he p ojec “Equal li e”, as pa o he Eu opean Human Exposome
Ne wo k, unde g an ag eemen No. 87474.
Ins i u ional Re iew Boa d S a emen : No applicable.
In o med Consen S a emen : No applicable.
Da a A ailabili y S a emen :
This s udy made use o hi d-pa y da a and so wa e ha a e pa ly
open. The new da a c ea ed in his wo k is a ailable upon eques .
Acknowledgmen s:
We a e g a e ul o Ja ie Casado No as and Julia Camps Fa es o enabling
access o he his o ical da a om Ba celona’s mic ophone ne wo k.
Con lic s o In e es : The au ho s decla e no con lic o in e es .
Appendix A
In Table A1, he link be ween he Open S ee Map ca ego ies and he a ic in ensi y,
ehicle speed and sha e o hea y ehicles is shown, o he 15 scena ios ha we e used in
his wo k.
Senso s 2023,23, 5865 17 o 19
Table A1.
T a ic pa ame e s assigned o he Open S ee Map s ee ca ego ies ha we e explici ly
calcula ed wi h he de e minis ic noise mapping me hodology. Vehicle in ensi ies (VIs) a e exp essed
in ca s pe hou , he sha e o hea y ehicles (SHV) in %.
Pe iod Road Type Vehicle
Speed (km/h)
VI (SHV)
Scena io 1
VI (SHV)
Scena io 2
VI (SHV)
Scena io 3
VI (SHV)
Scena io 4
VI (SHV)
Scena io 5
VI (SHV)
Scena io 6
Day
mo o way 130 20,400 (15%) 10,200 (15%) 5100 (15%) 20,400 (15%) 20,400 (20%) 20,400 (15%)
unk 110 8400 (15%) 4200 (15%) 2100 (15%) 8400 (15%) 33,600 (20%) 16,800 (15%)
p ima y 80 4800 (0%) 2400 (0%) 1200 (0%) 4800 (0%) 19,200 (5%) 9600 (0%)
seconda y 80 3300 (0%) 3300 (0%) 3300 (0%) 1750 (0%) 26,400 (5%) 6600 (0%)
e ia y 50 350 (0%) 350 (0%) 350 (0%) 175 (0%) 8400 (0%) 2100 (0%)
esiden ial 30 175 (0%) 175 (0%) 175 (0%) 85 (0%) 350 (0%) 1400 (0%)
se ice 30 80 (0%) 80 (0%) 80 (0%) 42 (0%) 175 (0%) 175 (0%)
E ening
mo o way 130 20,400 (11%) 10,200 (11%) 5100 (11%) 20,400 (11%) 20,400 (16%) 20,400 (11%)
unk 110 1600 (11%) 800 (11%) 400 (11%) 1600 (11%) 12,800 (16%) 3200 (11%)
p ima y 80 1000 (0%) 500 (0%) 250 (0%) 1000 (0%) 8000 (5%) 2000 (0%)
seconda y 80 600 (0%) 600 (0%) 600 (0%) 300 (0%) 9600 (5%) 1200 (0%)
e ia y 50 100 (0%) 100 (0%) 100 (0%) 50 (0%) 2400 (0%) 600 (0%)
esiden ial 30 50 (0%) 50 (0%) 50 (0%) 25 (0%) 100 (0%) 400 (0%)
se ice 30 25 (0%) 25 (0%) 25 (0%) 12 (0%) 50 (0%) 50 (0%)
Nigh
mo o way 130 20,400 (32%) 10,200 (32%) 5100 (32%) 20,400 (32%) 20,400 (37%) 20,400 (32%)
unk 110 800 (32%) 400 (32%) 200 (32%) 800 (32%) 6400 (37%) 1600 (32%)
p ima y 80 640 (0%) 320 (0%) 160 (0%) 640 (0%) 5120 (5%) 1280 (0%)
seconda y 80 360 (0%) 180 (0%) 180 (0%) 160 (0%) 5760 (0.5%) 720 (0%)
e ia y 50 50 (0%) 50 (0%) 50 (0%) 25 (0%) 1200 (0%) 300 (0%)
esiden ial 30 25 (0%) 25 (0%) 25 (0%) 12 (0%) 50 (0%) 50 (0%)
se ice 30 12 (0%) 12 (0%) 12 (0%) 6 (0%) 25 (0%) 100 (0%)
VI (SHV)
Scena io 7
VI (SHV)
Scena io 8
VI (SHV)
Scena io 9
VI (SHV)
Scena io 10
VI (SHV)
Scena io 11
VI (SHV)
Scena io 12
VI (SHV)
Scena io 13
VI (SHV)
Scena io 14
VI (SHV)
Scena io 15
33,300 (12.9%) 28,404 (16.2%) 34,315 (11.8%)
35,481 (7.7%) 20,400 (15%) 20,400 (20%) 20,400 (15%) 20,400 (15%) 20,400 (15%)
26,928 (5.4%) 21,012 (7.3%) 26,794 (4.4%) 27,705 (2.3%) 8400 (15%) 8400 (20%) 16,800 (15%) 8400 (15%) 33,600 (15%)
26,928 (5.4%) 21,012 (7.3%) 26,794 (4.4%) 27,705 (2.3%) 4800 (0%) 4800 (5%) 9600 (0%) 4800 (0%) 19,200 (0%)
18,192 (10.7%) 14,652 (13.6%)
18,061 (9.8%) 19,562 (5.7%) 6600 (0%) 6600 (5%) 13,200 (0%) 13,200 (0%) 26,400 (0%)
8928 (6.2%) 7476 (7.6%) 9050 (5.1%) 9717 (2.6%) 2100 (0%) 2100 (5%) 2100 (0%) 4200 (0%) 8400 (0%)
3216 (3.5%) 2400 (4.4%) 3062 (3.1%) 3404 (1.6%) 350 (0%) 350 (5%) 350 (0%) 700 (0%) 1400 (0%)
1098 (2.7%) 768 (3.6%) 1059 (2.4%) 1110 (1.4%) 175 (0%) 175 (5%) 175 (0%) 350 (0%) 700 (0%)
20,400 (11%) 20,400 (11%) 20,400 (11%) 20,400 (11%) 20,400 (11%) 20,400 (16%) 20,400 (11%) 20,400 (11%) 20,400 (11%)
3200 (11%) 3200 (11%) 3200 (11%) 3200 (11%) 1600 (11%) 1600 (16%) 3200 (11%) 1600 (11%) 12,800 (11%)
2000 (0%) 2000 (0%) 2000 (0%) 2000 (0%) 1000 (0%) 1000 (5%) 2000 (0%) 1000 (0%) 8000 (0%)
1200 (0%) 2400 (0%) 2400 (0%) 2400 (0%) 1200 (0%) 1200 (5%) 2400 (0%) 4800 (0%) 9600 (0%)
600 (0%) 600 (0%) 600 (0%) 600 (0%) 600 (0%) 600 (5%) 600 (0%) 1200 (0%) 2400 (0%)
400 (0%) 100 (0%) 100 (0%) 100 (0%) 100 (0%) 100 (5%) 100 (0%) 200 (0%) 400 (0%)
50 (0%) 50 (0%) 50 (0%) 50 (0%) 50 (0%) 50 (5%) 50 (0%) 100 (0%) 200 (0%)
20,400 (32%) 20,400 (32%) 20,400 (32%) 20,400 (32%) 20,400 (32%) 20,400 (37%) 20,400 (32%) 20,400 (32%) 20,400 (32%)
1600 (32%) 1600 (32%) 1600 (32%) 1600 (32%) 800 (32%) 800 (37%) 1600 (32%) 800 (32%) 6400 (32%)
1280 (0%) 1280 (0%) 1280 (0%) 1280 (0%) 640 (0%) 640 (5%) 1280 (0%) 640 (0%) 5120 (0%)
1440 (0%) 1440 (0%) 1440 (0%) 1440 (0%) 720 (0%) 720 (5%) 1440 (0%) 2880 (0%) 5760 (0%)
300 (0%) 300 (0%) 300 (0%) 300 (0%) 300 (0%) 300 (5%) 300 (0%) 600 (0%) 1200 (0%)
50 (0%) 50 (0%) 50 (0%) 50 (0%) 50 (0%) 50 (5%) 50 (0%) 100 (0%) 200 (0%)
25 (0%) 25 (0%) 25 (0%) 25 (0%) 25 (0%) 25 (5%) 25 (0%) 50 (0%) 100 (0%)
Re e ences
1.
Eu opean En i onmen al Agency. En i onmen al Noise in Eu ope 2020; EEA epo No 22/2019; Publica ions O ice o he Eu opean
Union: Copenhagen, Denma k, 2020.
2.
Lici a, G. Noise Mapping in he EU: Models and P ocedu es; CRC P ess: Boca Ra on, FL, USA; Taylo and F ancis G oup: Ge man-
own, NY, USA, 2013.
3.
Kessels, F. EURO Ad anced Tu o ials on Ope a ional Resea ch. In T a ic Flow Modelling: In oduc ion o T a ic Flow Theo y Th ough
a Genealogy o Models; Sp inge : Be lin/Heidelbe g, Ge many, 2018.
4.
END. Di ec i e 2002/49/EC o he Eu opean Pa liamen and o he Council o 25 June 2002 Rela ing o he Assessmen and Managemen o
En i onmen al Noise; Eu opean Commission: B ussels, Belgium, 2002.
5.
Rickenbacke , H.; B own, F.; Bilec, M. C ea ing en i onmen al consciousness in unde se ed communi ies: Implemen a ion and
ou comes o communi y-based en i onmen al jus ice and ai pollu ion esea ch. Sus . Ci ies Soc. 2019,47, 101473. [C ossRe ]
6.
Ba igón Mo illas, J.M.; Gómez Escoba , V.; Méndez Sie a, J.; Vílchez-Gómez, R.; Vaque o Ma ínez, J.; T ujillo Ca mona, J. A
ca ego iza ion me hod applied o he s udy o u ban oad a ic noise. J. Acous . Soc. Am. 2005,117, 2844–2852. [C ossRe ]
Senso s 2023,23, 5865 18 o 19
7.
Rey Gozalo, G.; Ba igón Mo illas, J.M.; Gómez Escoba , V. U ban s ee s unc ionali y as a ool o u ban pollu ion managemen .
Sci. To al En i on. 2013,461–462, 453–461. [C ossRe ] [PubMed]
8.
Zambon, G.; Benocci, R.; B ambilla, G. S a is ical Road Classi ica ion Applied o S a i ied Spa ial Sampling o Road T a ic Noise
in U ban A eas. In . J. En i on. Res. 2016,10, 411–420.
9.
Zambon, G.; Benocci, R.; B ambilla, G. Clus e ca ego iza ion o u ban oads o op imize hei noise moni o ing. En i on. Mon.
Assess. 2016,188, 26. [C ossRe ] [PubMed]
10.
Ba igón Mo illas, J.M.; Mon es González, D.; Gómez Escoba , V.; Rey Gozalo, G.; Vílchez-Gómez, R. A p oposal o p oducing
calcula ed noise mapping de ining he sound powe le els o oads by s ee s a i ica ion. En i on. Pollu .
2021
,270, 116080.
[C ossRe ] [PubMed]
11.
S aab, J.; Schady, A.; Weigand, M.; Lakes, T.; Taubenböck, H. P edic ing a ic noise using land-use eg ession—A scalable
app oach. J. Exp. Sci. En i on. Epidem. 2022,32, 232–243. [C ossRe ]
12.
Can, A.; Van Ren e ghem, T.; Rademake , M.; Dauwe, S.; Thomas, P.; De Bae s, B.; Bo eldoo en, D. Sampling app oaches o
p edic u ban s ee noise le els using ixed and empo a y mic ophones. J. En i on. Moni . 2011,13, 2710–2719. [C ossRe ]
13.
Can, A.; Dekoninck, L.; Rademake , M.; Van Ren e ghem, T.; De Bae s, B.; Bo eldoo en, D. Noise measu emen s as p oxies o
a ic pa ame e s in moni o ing ne wo ks. Sci. To al En i on. 2011,410, 198–204. [C ossRe ]
14.
B ink, M. A Re iew o Explained Va iance in Exposu e-Annoyance Rela ionships in Noise Annoyance Su eys. In P oceedings o
he In e na ional Commission on Biological E ec s o Noise (ICBEN), Na a, Japan, 18–22 June 2014.
15. WHO. En i onmen al Noise Guidelines o he Eu opean Region; WHO Regional O ice o Eu ope: Gene a, Swi ze land, 2018.
16. Spence, C.; Zampini, M. Audi o y con ibu ions o mul isenso y p oduc pe cep ion. Ac . Acus . Acus . 2006,92, 1009–1025.
17.
Kang, J.; Ale a, F.; Gjes land, T.; B own, L.; Bo eldoo en, D.; Schul e-Fo kamp, B.; Le che , P.; an Kamp, I.; Genui , K.; Fiebig,
A.; e al. Ten ques ions on he soundscapes o he buil en i onmen . Build. En i on. 2016,108, 284–294. [C ossRe ]
18.
Lionello, M.; Ale a, F.; Kang, J. A sys ema ic e iew o p edic ion models o he expe ience o u ban soundscapes. Appl. Acous .
2020,170, 107479. [C ossRe ]
19.
Can, A.; Gau eau, B. Desc ibing and classi ying u ban sound en i onmen s wi h a ele an se o physical indica o s. J. Acous .
Soc. Am. 2015,137, 208–218. [C ossRe ]
20.
Aumond, P.; Can, A.; De Coensel, B.; Bo eldoo en, D.; Ribei o, C.; La andie , C. Modeling Soundscape Pleasan ness Using
pe cep ual Assessmen s and Acous ic Measu emen s Along Pa hs in U ban Con ex . Ac . Acus . Acus .
2017
,103, 430–443.
[C ossRe ]
21.
Van Ren e ghem, T.; Thomas, P.; Dekoninck, L.; Bo eldoo en, D. Ge ing insigh in he pe o mance o noise in e en ions by
mobile sound le el measu emen s. Appl. Acous . 2022,185, 108385. [C ossRe ]
22.
De Coensel, B.; B own, A.L.; Tome ini, D. A oad a ic noise pa e n simula ion model ha includes dis ibu ions o ehicle
sound powe le els. Appl. Acous . 2016,111, 170–178. [C ossRe ]
23.
Kephalopoulos, S.; Pa io i, M.; An osso-Lédée, F. Common Noise Assessmen Me hods in Eu ope (CNOSSOS-EU); Publica ions
O ice o he Eu opean Union: Luxembou g, 2012; 180p.
24.
Wei, W.; Bo eldoo en, D.; Van Ren e ghem, T.; Ho nikx, M.; Fo ssén, J.; Salomons, E.; Ög en, M. U ban backg ound noise
mapping: The gene al model. Ac . Acus . Acus . 2014,100, 1098–1111. [C ossRe ]
25.
Aumond, P.; Fo in, N.; Can, A. O e iew o he NoiseModelling Open-Sou ce So wa e Ve sion 3 and i s Applica ions. In
P oceedings o he NOISE-CON Cong ess (261, 4, 2005–2011), Seoul, Republic o Ko ea, 12 Oc obe 2012.
26.
Boche , E.; Guillaume, G.; Picau , J.; Pe i , G.; Fo in, N. NoiseModelling: An Open Sou ce GIS Based Tool o P oduce En i on-
men al Noise Maps. ISPRS In . J. Geo. In o m. 2019,8, 130. [C ossRe ]
27.
Le Bescond, V.; Can, A.; Aumond, P.; Gas ineau, P. Open-sou ce modeling chain o he dynamic assessmen o oad a ic noise
exposu e. T ansp. Res. Pa D T ansp. En i on. 2021,94, 102793. [C ossRe ]
28. Fo ssén, J.; Ho nikx, M.; Bo eldoo en, D.; Wei, W.; Van Ren e ghem, T.; Ög en, M. A model o sound sca e ing by a mosphe ic
u bulence o use in noise mapping calcula ions. Ac . Acus . Acus . 2014,100, 810–815. [C ossRe ]
29.
Van Ren e ghem, T.; Ho oshenko , K.; Pa y, J.; Williams, D. S a is ical analysis o sound le el p edic ions in e ac ing and
u bulen a mosphe es. Appl. Acous . 2022,185, 108426. [C ossRe ]
30.
Ma lab. The Ma hWo ks Inc., e sion: 9.13.0 (R2022b); The Ma hWo ks Inc.: Na ick, MA, USA, 2022. A ailable online: h ps:
//www.ma hwo ks.com (accessed on 1 Decembe 2022).
31. Hagan, M.; Demu h, H.; Beale, M.; De Jesus, O. Neu al Ne wo k Design, 2nd ed.; Ma in Hagan: S illwa e , OK, USA, 2014.
32.
Wunde li, J.-M.; Pie en, R.; Habe mache , M.; Vienneau, D.; Cajochen, C.; P obs -Hensch, N.; Röösli, M.; B ink, M. In e mi ency
a io: A me ic e lec ing sho - e m empo al a ia ions o anspo a ion noise exposu e. J. Exp. Sci. En i on. Epidemiol.
2016
,
26, 575–585. [C ossRe ] [PubMed]
33.
Hall, F.L.; Papaky iakou, M.J.; Qui , J.D. Compa ison o ou doo mic ophone loca ions o measu ing sound insula ion o
building acades. J. Sound Vib. 1984,92, 559–567. [C ossRe ]
34.
Memoli, G.; Pa io i, M.; Kephalopoulos, S.; Lici a, G. Tes ing he acous ical co ec ions o e lec ions on a acade. Appl. Acous .
2008,69, 479–495. [C ossRe ]
35.
Ma eus, M.; Ca ilho, J.D.; Da Sil a, M.G. An expe imen al analysis o he co ec ion ac o s adop ed on en i onmen al noise
measu emen s pe o med wi h window moun ed mic ophones. Appl. Acous . 2015,87, 212–218. [C ossRe ]
Senso s 2023,23, 5865 19 o 19
36.
Ba igón Mo illas, J.M.; Mon es González, D.; Rey Gozalo, G. A e iew o he measu emen p ocedu e o he ISO 1996 s anda d.
Rela ionship wi h he Eu opean Noise Di ec i e. Sci. To al En i on. 2016,565, 595–606. [C ossRe ] [PubMed]
37.
Heu schi, K. A simple me hod o e alua e he inc ease o a ic noise emission le el due o buildings o a long s aigh s ee .
Appl. Acous . 1995,44, 259–274. [C ossRe ]
38.
Mon es González, D.; Ba igón Mo illas, J.M.; Rey Gozalo, G. The in luence o mic ophone loca ion on he esul s o u ban noise
measu emen s. Appl. Acous . 2015,90, 64–73. [C ossRe ]
39.
ISO 9613-2; Acous ics-A enua ion o Sound P opaga ion Ou doo s, Pa 2: Gene al Me hod o Calcula ion. In e na ional
O ganiza ion o S anda diza ion: Gene a, Swi ze land, 1996; e ised in 2017.
40.
Salomons, E.; Polinde , H.; Lohman, W.; Zhou, H.; Bo s , H.; Miedema, H. Enginee ing modeling o a ic noise in shielded a eas
in ci ies. J. Acous . Soc. Am. 2009,126, 2340–2349. [C ossRe ]
41.
Thomas, P.; Van Ren e ghem, T.; De Boeck, E.; D agone i, L.; Bo eldoo en, D. Re e be a ion-based u ban s ee sound le el
p edic ion. J. Acous . Soc. Am. 2013,133, 3929–3939. [C ossRe ]
42. Jonasson, H. Acous ical Sou ce Modelling o Road Vehicles. Ac . Acus . Acus . 2007,93, 173–184.
43.
Hadden, W.; Pie ce, A. Sound di ac ion a ound sc eens and wedges o a bi a y poin sou ce loca ions. J. Acous . Soc. Am.
1981
,
69, 1266–1276. [C ossRe ]
44.
Öh s öm, E.; Skånbe g, A.; S ensson, H.; Gidlö -Gunna sson, A. E ec s o oad a ic noise and he bene i o access o quie ness.
J. Sound Vib. 2006,295, 40–59. [C ossRe ]
45.
Fo ssén, J.; Ho nikx, M. S a is ics o A-weigh ed oad a ic noise le els in shielded u ban a eas. Ac . Acus . Acus .
2006
,
92, 998–1008.
46.
Fa es, J.C. Ba celona Noise Moni o ing Ne wo k. In P oceedings o he Eu onoise 2015, Maas ich , The Ne he lands, 31 May–3
June 2015; pp. 2315–2320.
47.
Mydla z, C.; Sha ma, M.; Locke man, Y.; S ee s, B.; Sil a, C.; Bello, J.P. The Li e o a New Yo k Ci y Noise Senso Ne wo k. Senso s
2019,19, 1415. [C ossRe ]
48.
Mie licki, F.; Mie licki, C.; Sineau, M. An Inno a i e App oach o Long Te m En i onmen al Noise Measu emen : RUMEUR Ne -
wo k in he Pa is Region. In P oceedings o he Eu onoise 2015, Maas ich , The Ne he lands, 31 May–3 June 2015;
pp. 2315–2320
.
49.
Van Ren e ghem, T.; Thomas, P.; Dominguez, F.; Dauwe, S.; Touha i, A.; Dhoed , B.; Bo eldoo en, D. On he abili y o consume
elec onics mic ophones o en i onmen al noise moni o ing. J. En i on. Mon. 2011,13, 544–552. [C ossRe ]
50.
Mydla z, C.; Salamon, J.; Bello, J.P. The implemen a ion o low-cos u ban acous ic moni o ing de ices. Appl. Acous .
2017
,
117, 207–218. [C ossRe ]
51.
Quin e o, G.; Balas egui, A.; Romeu, J. A low-cos noise measu emen de ice o noise mapping based on mobile sampling.
Measu emen 2019,148, 106894. [C ossRe ]
52. Yang, D.; Zhao, J. Acous ic Wake-Up Technology o Mic osys ems: A Re iew. Mic omachines 2023,14, 129. [C ossRe ]
53.
Buch, N.; Velas in, S.; O well, J. A Re iew o Compu e Vision Techniques o he Analysis o U ban T a ic. IEEE T ans. In ell.
T anspo . Sys . 2011,12, 920–939. [C ossRe ]
54.
F edianelli, L.; Ca pi a, S.; Be na dini, M.; Del Pizzo, L.; B occhi, F.; Bianco, F.; Lici a, G. T a ic Flow De ec ion Using Came a
Images and Machine Lea ning Me hods in ITS o Noise Map and Ac ion Plan Op imiza ion. Senso s
2022
,22, 1929. [C ossRe ]
[PubMed]
Disclaime /Publishe ’s No e:
The s a emen s, opinions and da a con ained in all publica ions a e solely hose o he indi idual
au ho (s) and con ibu o (s) and no o MDPI and/o he edi o (s). MDPI and/o he edi o (s) disclaim esponsibili y o any inju y o
people o p ope y esul ing om any ideas, me hods, ins uc ions o p oduc s e e ed o in he con en .
