sensors

Article A Hybrid Active Control System for the Attenuation of Road Noise Inside a Vehicle Cabin

Zibin Jia, Xu Zheng * , Quan Zhou, Zhiyong Hao and Yi Qiu College of Energy Engineering, Zhejiang University, Hangzhou 310027, China; [email protected] (Z.J.); [email protected] (Q.Z.); [email protected] (Z.H.); [email protected] (Y.Q.) * Correspondence: [email protected]; Tel.: +86-0571-87953286

 Received: 17 November 2020; Accepted: 14 December 2020; Published: 15 December 2020 

Abstract: This paper proposed a local active control method for the reduction of road noise inside a vehicle cabin. A multichannel simplified hybrid active (sHANC) system was first developed and applied to the rear left seat of a large sport utility vehicle (SUV). The attenuation capability of the sHANC system was investigated through simulations, using reference signals provided by accelerometers on the suspensions and bodywork of the vehicle and on the floor of cabin, respectively. It was shown that compared to the traditional feedforward system, the sHANC system using either vibrational or acoustical reference signals can produce a significant suppression of the narrowband peak noise between 75 and 80 Hz, but the system lost the control capability in a range of 100–500 Hz when the acoustic signals were used as references. To reduce the practical implementation costs while maintaining excellent reduction performance, a modified simplified hybrid ANC (msHANC) system was further proposed, in which combined vibrational and acoustical signals were used as reference signals. The off-line analyses showed that four reference accelerometers can be substituted by ten microphones without compromising attenuation performance, with 3.7 dBA overall being achieved. The effect of delays on the reduction performance of msHANC system was also investigated. The result showed that the msHANC system was more sensitive to the delays compared to the sHANC system if using only vibrational reference signals.

Keywords: active noise control; hybrid control system; road noise; reduction performances; vehicle cabin

1. Introduction Recently, the noise, vibration, and harshness (NVH) characteristics of vehicles have attracted great attention from vehicle manufacturers. Reducing the interior noise of a vehicle to improve driver and passenger comfort has become a pressing issue [1]. Since the 1980s, the application of active noise control (ANC) technology to reduce the interior noise of vehicles has been extensively investigated [2]. The first attempt at feedforward engine noise control was presented by Elliott et al. [3], and the developed engine noise control (ENC) system could reduce the pressure level at engine firing frequency between 10 and 15 dB. Due to the development of the digital signal processing (DSP), such a system has been implemented widely by a number of manufacturers [4,5]. Another application of active sound control for vehicle interior noise is road-type noise, which is a result of the interaction between the road surface and the vehicle tires. Compared to the ENC system, it is a great challenge to use a feedforward control strategy in the road noise control (RNC) system because of the broadband random nature of road excitation [6]. To achieve effective road noise control, the reference signals need to be highly correlated with the primary disturbance. However, in practice, due to the complex propagation path between structural and acoustic noise, it is difficult to obtain good coherence in the frequency band of interest.

Sensors 2020, 20, 7190; doi:10.3390/s20247190 www.mdpi.com/journal/sensors Sensors 2020, 20, 7190 2 of 15

To overcome this limitation, Duan et al. [7] proposed a combined feedforward–feedback ANC system, which was formed by adding a feedback controller to the traditional feedforward ANC (tFANC) system; the simulation results showed that the hybrid control system can suppress the narrowband road noise component, which cannot be attenuated by feedforward structure operating alone, and it achieved a further 1 dBA of overall reduction. In their research, the internal model control (IMC) structure was used to build the feedback controller, which meant that the error signals and the secondary signals filtered by the estimation of secondary path were used to synthesize the reference signals for the feedback controller. Although this structure had great stability and robustness, as discussed by Padhi et al. [8], it had a heavy computational load by filtering the secondary signals, especially when the IMC structure was combined with the feedforward structure based on the Filter-x LMS (FxLMS) algorithm, which increased the cost of hardware. To reduce the burden for a real-time controller, Wu et al. [9] proposed a simplified feedback ANC system, in which the error signals were adapted directly as the reference signals for feedback structure and had the advantages of low computational load and ease implementation. The simulation and experimental results proved that the simplified system does not degrade the steady state performance significantly compared to the IMC feedback system. A simplified hybrid ANC (sHANC) system was also developed [10]; it showed great attenuation performance for uncorrelated disturbances which cannot be controlled by a feedforward structure. Using the traditional hybrid ANC (tHANC) system based on the IMC structure as a comparison, the sHANC system can improve the convergence rate and noise reduction performance by reducing the coupling relationship between the feedforward and feedback structures. However, these conclusions were drawn under the situation that the control system was aimed at a single-input single-output (SISO) problem and the primary path was accurately modeled. Moreover, the influence of the sound field characteristic within the enclosure was not considered in the specific simulations. Another focus of implementing a road noise control system is the determination of the reference sensors. This is a common method to install accelerometers on the suspensions and bodywork of the vehicle to obtain the advance structural vibration signals caused by road excitation [11,12]. It has been shown that at least six accelerometers are needed to capture the primary components of road noise and achieve a reasonable noise reduction level [11]. Due to the high cost of accelerometers, this method shows limited commercial implementation. An attempt at reducing the cost of the control system was presented by Mohammad et al. [13], who mounted low-cost microphones on the floor of the vehicle cabin as reference sensors and proved that it had a similar attenuation capacity compared to the use of accelerometers. This system shows great application benefits, since a micro-electro-mechanical system (MEMS) is much cheaper than an accelerometer nowadays. However, in their work, since the ANC controller was formulized in the frequency domain and lacked causality constraints, the effect of advanced time provided by microphone signals was not considered. Moreover, the microphone reference signals can be affected easily by the spatial variance of the sound field within the cabin, as mentioned by Cheer et al. [14], which limits the attenuation performance of the control system. To make the sHANC system applicable in practical situations, in this paper, a multichannel sHANC system was developed to control the road noise inside a vehicle cabin. The attenuation performances of the aforementioned system, multichannel tFANC system, and tHANC system were all analyzed by using microphone or accelerometer reference signals to investigate the effects of different reference signals and control strategies. In addition, to reduce the implementation costs while maintaining the excellent noise attenuation performance of the sHANC system, a modified simplified hybrid ANC (msHANC) system using both microphone and accelerometer signals was also proposed in detail. The remaining sections of this paper are organized as follows: Section2 presents the formulation of the multichannel sHANC system in the time domain. In Section3, the disturbance and reference signals measured inside a large sport utility vehicle (SUV) cabin are used to compare the attenuation performances of the multichannel sHANC system and other traditional systems. Then, the reduction performances of a modified sHANC system using both two types of reference signals are presented in Section4. Finally, the conclusions of this work are provided in Section5. Sensors 2020, 20, x FOR PEER REVIEW 3 of 15

signals measured inside a large sport utility vehicle (SUV) cabin are used to compare the attenuation performances of the multichannel sHANC system and other traditional systems. Then, the reduction Sensorsperformances2020, 20, 7190 of a modified sHANC system using both two types of reference signals are presented3 of 15 in Section 4. Finally, the conclusions of this work are provided in Section 5.

2. The Multichannel sHANC System RoadRoad noise can be considered as stationarystationary random broadband noise, and the control systemsystem should be constrained casually. In this case, the attenuation performance of the hybrid ANC system was investigated in the time domain [15].]. A multiple multiple-input-input multiple multiple-output-output (MIMO) sHANC system was established andand thethe blockblock diagramdiagram ofof thethe systemsystem isis shownshown inin FigureFigure1 1..

Pe

Gn - d x f + Wf + Gm + v Adaptation e Ĝm algorithm u f + ut

+ Wb ub xb Adaptation Ĝm algorithm

Figure 1. Block diagram of the MIMO MIMO simplified simplified hybrid active noise control system system,, mainly including

reference signals vector x f,, errorerror signalssignals vectorvector e,, twotwo nominalnominal plantplant responsesresponses matricesmatrices GGmmandandGn ,, two control filter matrices Wf and Wb. two control filter matrices W f and Wb . In Figure1, the feedforward structure uses a number of K reference sensors, and at nth sample th In Figure 1, the feedforward structure usesh a number of K referenceiT sensors, and at n ( ) ( ) = ( ) ( ) ( ) time, the signals form a K 1 vector, x f n x f 1 n , x f 2 n ... x f K n . The MIMOT control system ×  aimssample to time, attenuate the signals the road form noise a (K component 1) vector, aroundx f(n ) the x f12 passenger’s( n ), x f ( n ) . . . head. x fK ( n ) The. The number MIMO of control error microphonessystem aims to used attenuate around the the road passenger’s noise component head is defined around as theL. Thepassenger’s(L 1) vector head. ofThe error number signals, of T × eerror(n) = microphones[e1(n), e2(n) used... eL ( aroundn)] , is giventhe passenger’s by the concatenation head is defined of L microphone as L . Thesignals, (L  1) andvectore(n ) ofcan error be expressed as: T signals, e(n )  e12 ( n ), e ( n ) . . . eL ( n ) , is given by the concatenation of L microphone signals, and e(n) = d(n) + v(n) + Gmut0(n) (1) e()n can be expressed as: T where d(n) = [d1(n), d2(n) ... dL(n)] , which is produced by reference signal vector, x f (n),  through the (L K) matrix of transfere()()()()n dresponses, n  v n  GPmte. u n The (L 1) vector of disturbances,(1) × T × v(n) = [v1(n), v2(n) ... vL(n)] , representsT uncorrelated noise in the primary path. The Gm is the matrix where d(n )  d12 ( n ), d ( n ) . . . dL ( n ) , which is produced by reference signal vector, x f ()n , through of nominal plant responses between the M secondary sources and L error microphones, and ut0(n) representsthe ()LK the vector matrix of current of transfer and past responses, control signals Pe for. Thesecondary (L  1) sources. vector of disturbances, The matrix of nominalT plant responses, G , is defined as: v(n )  v12 ( n ), v ( n ) . . . vL ( n ) , represents uncorrelatedm noise in the primary path. The Gm is the

matrix of nominal plant responses between the M secondaryT sources and L error microphones, Gm = [G1, G2 ... GL] (2) and ut()n represents the vector of current and past control signals for secondary sources. h iT G The matrix= T of nominalT T plant responses, m , is defined as: where Gl Gl1, Gl2 ... GlM , and Glm is the vector of nominal plant responses between the mth secondary source and lth error microphone, which is modeledT by an Ith order finite impulse response GGGG= , . . .  (2) (FIR) filter and written as: mL12 T h iT GGGG TTT, . . . Glm = Glm1, Glm2 ... Glm(I 1) m(3)th where l l12 l lM and Glm is the vector of nominal− plant responses between the ( )th th secondaryIn addition, source the andut0 nl is error the output microphone summation which of control is modeled signal by vector an I in feedforward order finite structure impulse andresponse feedback (FIR) structure filter and and written can be as: expressed as:

T u0(n) = u0 (n) + u0 (n) (4) GGGGlmt  lm1,f lm 2 . . . b lm ( I  1) (3)

Sensors 2020, 20, 7190 4 of 15

 T h iT u (n) = uT uT uT u (n) = uT uT uT u u where 0f f 1, f 2 ... f M and b0 b1, b2 ... bM . The f m and bm are signal vectors for mth secondary source of feedforward structure and feedback structure, respectively, which include the current and I 1 past control signals, which are defined as: − h iT u = u (n), u (n 1) ... u (n I 1) (5) f m f m f m − f m − −

u = [u (n), u (n 1) ... u (n I 1)]T (6) bm bm bm − bm − − The control signal (M 1) vector of feedforward structure at nth sample time is defined as h × iT ( ) = ( ) ( ) ( ) ( ) u f n u f 1 n , u f 2 n ... u f M n , and the u0f n can be derived from the buffer and shift operations of u f (n). As shown in the block diagram, the u f (n) is obtained from the vector of reference signals x f (n) filtered through the control filter matrix W f , which is given by:

h iT W f = w f 1, w f 2 ... w f M (7) h i where w f m = w f m1, w f m2 ... w f mK , and w f mk is a vector of control filter between mth secondary source and kth refenence signal, which is a Jth order FIR filter: h i w f mk = w f mk1, w f mk2 ... w f mk(J 1) (8) −

Therefore, the u f (n) can be expressed as:

u f (n) = W f x0f (n) (9)

h iT where x (n) = x (n), x (n 1) ... x (n J + 1), x (n) ... x (n J 1) , which can be derived 0f f 1 f 1 − f 1 − f 2 f K − − from the buffer and shift operations of x f (n). For the purpose of adaptive control, the multichannel leaky FxLMS algorithm [16,17] is used to update the coefficients of the control filter matrix, W f . In order to ensure the correct dimensions for matrix multiplication, the elements in matrix W are rearranged, and the new (MKJ 1) matrix is f × defined as:  T W = wT wT wT 0f f 0, f 1 ... f (J 1) (10) − h iT where w f i = w f 11i, w f 12i ... w f 1Ki, w f 21i ... w f MKi , and the updated equation of control filter matrix can be expressed as: T W0 (n + 1) = (1 α β )W0 (n) α Rˆ (n)e(n) (11) f − f f f − f f where α f is the step size of the feedforward control, β f is the leakage factor of feedforward control, which is used to improve the robustness of the algorithm, and Rˆ f is the estimated filtered reference signal (L MKJ) matrix, which can be represented as: ×   rT (n) rT (n 1) rT (n I + 1)  f 1 f 1 f 1   − ··· −   .   rT (n) rT (n 1) .  ˆ  f 2 f 2  R f (n) =  −  (12)  . . .   . . .     rT (n) rT (n 1) rT (n I + 1)  f L f L − ··· f L − Sensors 2020, 20, 7190 5 of 15

h iT where r f l(n) = r f l11(n), r f l12(n) ... r f l1K(n), r f l21(n) ... r f lMK(n) , and r f lmk(n) represents the kth reference signal which is filtered by the estimated plant response between the mth secondary source and lth error microphone, which is equal to:

I 1 X− r (n) = Gˆ x (n i) (13) f lmk lmi f k − i=0

Then, considering the feedback structure in Figure1, the reference vector of the feedback controller, T xb(n) = [xb1(n), xb2(n) ... xbL(n)] , is derived from the error signal vector directly, i.e.,

xb(n) = e(n) (14)

Equation (14) shows that the feedforward structure will not have an effect on the feedback structure directly compared to the traditional hybrid ANC system, which will reduce the coupling between these two structures. The control signal vector ub(n) is given by:

ub(n) = Wbxb0 (n) (15)

( ) ( ) where xb0 n is derived from the buffer and shift operations of xb n . Similarly, the updated equation of the control filter matrix in the feedback structure can be expressed as: T W0 (n + 1) = (1 α β )W0 (n) α Rˆ (n)e(n) (16) b − b b b − b b ˆ T( ) where αb is the step size, βb is the leakage factor of feedback control, and Wb0 and Rb n have ˆ T( ) a similar form to W0f and R f n in Equation (11), respectively. Similarly to the feedforward structure, the adaptive feedback controller directly uses the vector e(n) as the error signal input. It should be noted that if the microphones are selected as reference sensors, a feedback path between control sources and references is needed to cancel the disturbances of the secondary sources on the reference signals, which is modeled by a (M K) transfer response matrix Gn, as shown by × the dotted line in Figure1.

3. Active Noise Control of Road Noise Based on Multichannel sHANC System

3.1. Experimental Arrangement Previous research has proven that, compared to the global control system, the local control system can extend the control frequency band while improving noise attenuation performance [15,18,19]. The local active noise control system was considered in this paper. To verify the potential performance of the simplified hybrid road noise control system as described in the previous section, the experiments were conducted in an SUV. Figure2a shows the active headrest system at the rear left seat of the SUV, which contains two headrest speakers and two error microphones representing the position of the passenger’s ears. Before the experiments, the determination of the reference sensors for feedforward structure was considered. To compare the impact of different types of reference sensor on noise attenuation and minimize implementation costs, two placement schemes were arranged initially:

1. Vibration sensors: 12 triaxial accelerometers were numbered as references. They were mounted on the four suspension systems and tires, where the characteristics of road excitation transmitted through the structure can be captured, such as the hubs of the axle, as shown in Figure2b, the bushings of the sub-frame, the joints between the bodywork and sub-frame, as shown in Figure2c, and so on. Each accelerometer had three axes, so a total of 36 vibration signals were measured. Sensors 2020, 20, x FOR PEER REVIEW 6 of 15

1. Vibration sensors: 12 triaxial accelerometers were numbered as references. They were mounted on the four suspension systems and tires, where the characteristics of road excitation transmitted through the structure can be captured, such as the hubs of the axle, as shown in Figure 2b, the bushings of the sub-frame, the joints between the bodywork and sub-frame, as shown in Figure 2c, and so on. Each accelerometer had three axes, so a total of 36 vibration signals were measured. Sensors2. Microphone2020, 20, 7190 sensors: 10 microphones located on the floor of the vehicle cabin were numbered6 of 15as references, and the exact position of each microphone is shown in Figure 2d.

2. MicrophoneThe experiments sensors: were 10 conducted microphones when located the SUV on the was floor driven of the at a vehicle constant cabin speed were of numbered 80 km/h on as a roughreferences, road, and and the the disturbance exact position signals of each from microphone error microphones is shown and in Figure the reference2d. signals of two defined schemes were obtained simultaneously.

(a) (b)

8 6 4 10 2

Floor

1 3 5 9 7 — Microphone

(c) (d)

FigureFigure 2. 2.( a()a The) The active active headrest headrest system system at the at rear the leftrear seat left of seat the SUVof the containing SUV containing two headrest two speakersheadrest andspeakers two error and microphones;two error microphones; (b) The installation (b) The installation location of location accelerometer: of accelerometer: the hubs of the axle; hubs (c) of A axle; few installation(c) A few locationsinstallation of accelerometers:locations of accelerometers: the bushings ofthe sub-frame, bushings theof sub-frame, joints between the bodyworkjoints between and sub-frame;bodywork (dand) The sub-frame; schematic diagram(d) The ofschematic the numbered diagram reference of the microphone numbered placement.reference microphone placement. The experiments were conducted when the SUV was driven at a constant speed of 80 km/h on a3.2. rough Selection road, of and Reference the disturbance Signals signals from error microphones and the reference signals of two defined schemes were obtained simultaneously. To ensure that the feedforward structure of the hybrid system can achieve good controlling 3.2.performance, Selection of it Reference is necessary Signals to keep the reference signals correlated with disturbances. Theoretically, better coherence between the disturbances and reference signals can be achieved by using more referenceTo ensure sensors. that Therefore, the feedforward for a low-cost structure microp of thehone hybrid placement system scheme, can achieve firstly, all good ten microphone controlling performance,inputs were used it isnecessary as reference to keepsignals. the However, reference signalsfor the accelerometer correlated with placement disturbances. scheme, Theoretically, due to the betterhigh coherencecost of accelerometers, between the disturbances it is necessary and to reference select the signals minimum can be number achieved of by sensors using more to maintain reference a sensors.relatively Therefore, strong correlation for a low-cost with microphone disturbances. placement Moreover, scheme, in firstly,actual all implementation, ten microphone inputstriaxial wereaccelerometers used as reference will be signals. replaced However, by uniaxial for the accelerometers, accelerometer placement so the goal scheme, became due the to the selection high cost of ofdominant accelerometers, vibration it issignals necessary among to selectall signals. the minimum As suggested number by of[7,11], sensors the tovibration maintain signals a relatively can be strongranked correlation by calculating with disturbances.the average coherence Moreover, between in actual the implementation, vibration signals triaxial and accelerometerseach error signal will in bethe replaced control byrange, uniaxial and accelerometers,the flowchart is soshown the goal in Fi becamegure 3. the At selectionthe beginning, of dominant all the vibrationreference signalssignals amongwere calculated all signals. respectively As suggested to by find [7,11 the], thesignal vibration having signals the highest can be rankedaverage by coherence calculating with the the average error coherencesignals, which between was thedefined vibration as the signals first dominant and each signal. error signal Then, inthe the reference control range,signal was and removed the flowchart from is shown in Figure3. At the beginning, all the reference signals were calculated respectively to find the signal having the highest average coherence with the error signals, which was defined as the first dominant signal. Then, the reference signal was removed from all the reference signals and stored in the buffer. In the next iteration, along with reference signals in the buffer, the remaining reference Sensors 2020, 20, x FOR PEER REVIEW 7 of 15 all the reference signals and stored in the buffer. In the next iteration, along with reference signals in the buffer, the remaining reference signals were calculated again to find the second dominant signal based on the average multiple coherence function, which is defined as:

L fupper = 1 γ 2 Sensors 2020, 20, 7190 J  l ()f (17)7 of 15 = L lf1 lower where f = 20 Hz , f = 500 Hz , representing the frequency band of interest. The multiple signals werelower calculated againupper to find the second dominant signal based on the average multiple th γ 2 coherence function,function at which the l is definederror microphone, as: l , is defined as:

−1 H LS fupperSS γ 2 X= xXexxxell l1 2 (18) J = S γl ( f ) (17) L eell l=1 flower Similarly, the second dominant signal was stored in the buffer by order. The above iteration wherecontinuedflower until= 20all Hzthe, referencefupper = 500 signalsHz, were representing ranked, and the frequencythe corresponding band of ranking interest. list The of reference multiple 2 signalscoherence was function output. atThe the abovelth error rank microphone, operation wasγl , conducted is defined as:based on the data when the SUV was driven at a constant speed of 80 km/h on a rough road.1 H Through analysis, it was found that the Sxe S− S vibration signals had a higher correlation with2 the errorl xx signalsxel located in the rear suspension system γl = (18) of the SUV rather than the front suspension (for simplicity,Selel the ranking list is not shown in this section).

Figure 3. TheThe flowchart flowchart of ranking the vi vibrationalbrational reference signals.

Then,Similarly, to remove the second some dominant redundant signal accelerometers was stored from in the pre-set buffer positions, by order. the The simulations above iteration based oncontinued the feedforward until all thestructure reference (Figure signals 1) alone were were ranked, operated and the multiple corresponding times wi rankingth a decreasing list of reference number ofsignals references was output. to observe The how above the rank attenuation operation perf wasormance conducted dropped. based onAccording the data to when the ranking the SUV lists, was thedriven lowest at a two constant reference speed channels of 80 km were/h on removed a rough road. when Through a certain analysis, simulation it was was found finished. that theFor vibration instance, thesignals first had simulation a higher contained correlation all with 36 references, the error signals the second located simulation in the rear contained suspension the systemtop 34 references, of the SUV andrather the than last thesimulation front suspension only contained (for simplicity, the top four the reference ranking list signals. is not The shown ANC in thisperformance section). (overall reduction)Then, with to remove the number some redundantof vibrational accelerometers reference channels from pre-set decreasing positions, is shown the simulationsin Figure 4. basedIt can beon theseen feedforward that the ANC structure performanc (Figuree 1dropped) alone were significantly operated when multiple the times number with of a decreasingreferences was number less thanof references 10; hence, to the observe top 10 how vibrational the attenuation reference performance signals were dropped. chosen for According the following to the analysis. ranking lists, the lowest two reference channels were removed when a certain simulation was finished. For instance, the first simulation contained all 36 references, the second simulation contained the top 34 references, and the last simulation only contained the top four reference signals. The ANC performance (overall reduction) with the number of vibrational reference channels decreasing is shown in Figure4. It can be seen that the ANC performance dropped significantly when the number of references was less than 10; hence, the top 10 vibrational reference signals were chosen for the following analysis. The multiple coherences between each error signal and selected reference signals are shown in Figure5; when (a) ten vibration signals were used as reference signals, (b) ten acoustic signals were used as reference signals. From the results, it is interesting to observe that these two different types of reference signals were both highly correlated with each error signal until the frequency reached around 200 Hz, while at the higher frequencies, they lost the coherence, which limited the attenuation performance of the feedforward controller. These results also indicated that if the causality constraint of the control system was not considered, using acoustic signals as references could achieve a similar level of control effect as using vibration signals. Sensors 2020, 20, x FOR PEER REVIEW 8 of 15

Sensors 2020, 20, 7190 8 of 15 Sensors 2020, 20, x FOR PEER REVIEW 8 of 15

Figure 4. The predicted performance (overall reduction) of pure feedforward control system using different numbers of vibration channels.

The multiple coherences between each error signal and selected reference signals are shown in Figure 5; when (a) ten vibration signals were used as reference signals, (b) ten acoustic signals were used as reference signals. From the results, it is interesting to observe that these two different types of reference signals were both highly correlated with each error signal until the frequency reached around 200 Hz, while at the higher frequencies, they lost the coherence, which limited the attenuation performanceFigure 4. ofThe the predicted feedforward performance controller. (overall These reduction) results alsoof pure indicated feedforward that if control the causality system constraintusing of thedifferentdi controlfferent numbers system was of vibration not considered, channels. using acoustic signals as references could achieve a similar level of control effect as using vibration signals. The multiple coherences between each error signal and selected reference signals are shown in Figure 5; when (a) ten vibration signals were used as reference signals, (b) ten acoustic signals were used as reference signals. From the results, it is interesting to observe that these two different types of reference signals were both highly correlated with each error signal until the frequency reached around 200 Hz, while at the higher frequencies, they lost the coherence, which limited the attenuation performance of the feedforward controller. These results also indicated that if the causality constraint of the control system was not considered, using acoustic signals as references could achieve a similar level of control effect as using vibration signals.

(a) (b)

Figure 5. The multiple coherence betweenbetween eacheach of of the the error error microphones microphones and and all all selected selected vibrational vibrational or oracoustical acoustical reference reference signals signals when wh theen the SUV SUV was was driven driven at a at constant a constant speed speed of 80 of km 80/h km/h on a roughon a rough road. road.(a) Ten (a vibrational) Ten vibrational reference reference signals; signals; (b) Ten (b acoustical) Ten acoustical reference reference signals. signals.

3.3. Plant Plant Response Response Measurement

In order to implement the hybrid control system, the estimated plant responses matrix, Gˆ m, shouldˆ In order to implement the hybrid control system, the estimated plant responses matrix, G m , be determined. In practice, the individual plant response between each headrest speaker and each should be determined. In practice, the individual plant response between each headrest speaker and microphone was modeled(a) by an adaptive 128-tap FIR filter. Each headrest(b speaker) was driven with each microphone was modeled by an adaptive 128-tap FIR filter. Each headrest speaker was driven a sweep signal independently to obtain the responses of the microphones. Along with the drive with Figurea sweep 5. Thesignal multiple independently coherence betweento obtain each the of responses the error microphonesof the microphones. and all selected Along vibrational with the drive signals, these response signals were used to train the FIR filters. After the filters were converged, signals,or acoustical these response reference signals signals were when used the SUVto train was the driven FIR atfilters. a constant After speed the filters of 80 werekm/h converged,on a rough the the coefficients of each FIR filter were considered as the estimated impulse response function (IRF). coefficientsroad. (a )of Ten each vibrational FIR filter reference were considered signals; (b) asTen the acoustical estimated reference impu lsesignals. response function (IRF). The The IRFs between left speaker 1 and two error microphones are presented as an example in Figure6. IRFs between left speaker 1 and two error microphones are presented as an example in Figure 6. 3.3.Moreover, Plant Response the plots Measurement showed that the IRFs between the secondary speakers and the error microphones can be well modeled by 128-tap filters. ˆ In order to implement the hybrid control system, the estimated plant responses matrix, G m , should be determined. In practice, the individual plant response between each headrest speaker and each microphone was modeled by an adaptive 128-tap FIR filter. Each headrest speaker was driven with a sweep signal independently to obtain the responses of the microphones. Along with the drive signals, these response signals were used to train the FIR filters. After the filters were converged, the coefficients of each FIR filter were considered as the estimated impulse response function (IRF). The IRFs between left speaker 1 and two error microphones are presented as an example in Figure 6.

Sensors 2020, 20, x FOR PEER REVIEW 9 of 15

Moreover,Sensors 2020, the20, 7190 plots showed that the IRFs between the secondary speakers and the error microphones9 of 15 can be well modeled by 128-tap filters.

FigureFigure 6. 6. TheThe impulse impulse response response (IRF) (IRF) between between left left speaker speaker 1 1 and and two two error error microphones. microphones.

3.4.3.4. The The Attenuation Attenuation Performance Performance of of Multichannel Multichannel sHANC sHANC System System InIn this this section, section, the the off-line off-line model model is is built built to to analyze analyze the the attenuation attenuation performance performance of of the the proposed proposed MIMOMIMO sHANC sHANC system. system. As As a a comparison, comparison, the the noise noise attenuation attenuation capability capability of of tFANC tFANC and and tHANC tHANC systemsystem are alsoalso presented. presented. These These systems systems were were built built using using disturbance disturbance signals signals and the and reference the reference signals signalsof two definedof two defined schemes, schemes, which were which obtained were obtain simultaneouslyed simultaneously when the SUVwhen was the drivenSUV was at 80 driven km/h onat 80a roughkm/h road.on a rough In the road. simulations, In the simulations, to guarantee to that guarantee the systems that the were systems effective were and effective stable, the and step stable, size theand step leakage size factorand leakage of the adaptivefactor of feedforwardthe adaptive (feedback)feedforward controller (feedback) were controller chosen bywere trial chosen and error by trialempirically. and error Figure empirically.7 compares Figure the e7 ffcomparesects of di fftheerent effects reference of different signals reference and control signals systems and oncontrol noise systemsreduction on performance. noise reduction Figure performance.7a shows theFigure performance 7a shows ofthe the performance sHANC system of the compared sHANC system to two comparedtraditional to ANC two traditional systems using ANC vibration systems reference using vibr signals.ation reference From the signals. sound From pressure the spectrum,sound pressure it can spectrum,be seen that it whencan be vibration seen that signals when were vibration used as references,signals were more used significant as references, suppression more of significant the narrow suppressionpeak from 70 of tothe 85 narrow Hz could peak be from achieved 70 to 85 by Hz using could the be sHANC achieved and by tHANCusing the systems, sHANC comparedand tHANC to systems,tFANC system. compared This to phenomenon tFANC system. indicated This thatphenom the 70–85enon Hzindicated uncorrelated that the narrow 70–85 disturbance Hz uncorrelated cannot narrowbe controlled disturbance by the cannot pure feedforwardbe controlled controller,by the pure while feedforward it can be controller, reduced bywhile adding it can the be feedbackreduced bystructure. adding Forthe thefeedback broadband structure. range For 100–500 the broadban Hz, it cand berange observed 100–500 that Hz, all theit can control be observed systems showedthat all thea relatively control goodsystems reduction showed capability, a relatively though good better reduction reduction capability, performance though was limitedbetter becausereduction of performancethe poor coherence was limited between because the reference of the poor signals coherence and error between signals the at higher reference frequencies, signals and as shown error signalsin Figure at5 a.higher These frequencies, results also showedas shown that in theFigu sHANCre 5a. These system results can achieve also showed similar noisethat reductionthe sHANC to systemthe tHANC can achieve system. similar Compared noise to thereduction tFANC to system, the tHANC the sHANC system. system Compared not only to achieved the tFANC 3 dBA system, more theattenuation sHANC insystem the narrow not only band achieved of 70–85 3 Hz dBA but more also achievedattenuation a further in the 0.3narrow dBA ofband the overallof 70–85 reduction. Hz but also achievedThe same a reductionfurther 0.3 performancesdBA of the overall of control reduction. systems using all acoustical reference signals are shownThe in same Figure reduction7b. An interestingperformances observation of control was systems that all using control all acoustical systems were reference e fficient signals in noise are shownreduction in Figure of the 70–857b. An Hz interesting narrow peak, observation while the was performances that all control of the systems sHANC were and efficient tHANC in system noise reductionimproved of slightly. the 70–85 It canHz narrow be inferred peak, that while the the reference performances microphones of the onsHANC the cabin and tHANC floor can system detect improvedmore information slightly. aboutIt can thebe inferred narrow peakthat the disturbance, reference andmicrophones this conclusion on the cancabin also floor be verifiedcan detect by morethe coherence information results about in the Figure narrow5b. peak However, disturbance, for the and 100–500 this conclusion Hz broadband can also range, be verified the tHANC by the coherencesystem made results only in a slightFigure di 5b.fference However, compared for th toe 100–500 the system Hz whenbroadband acceleration range, signalsthe tHANC were system used as madereferences only (Figurea slight7a). difference This may compared be caused byto the unsatisfiedsystem when causality acceleration of the feedforwardsignals were controller, used as referencesi.e., the microphone (Figure 7a). signals This may cannot be caused provide by enough the unsatisfied advanced causality time. Moreover, of the feedforward within this controller, frequency i.e.,band, the themicrophone other two signals hybrid cannot systems provide seemed enough to demonstrate advanced time. no improvementMoreover, within in noise this frequency reduction. band,This was the becauseother two the hybrid delays systems in the plantseemed responses to demonstrate limit the no performance improvement and in achievable noise reduction. attenuation This wasof the because feedback the controller,delays in the as discussedplant responses in [6,13 limit]. Therefore, the performance based on and the achievable above findings, attenuation to reduce of thethe feedback implementation controller, costs as bydiscusse substitutingd in [6,13]. a few Therefore, accelerometers based withon the microphones above findings, while to maintaining reduce the

Sensors 20202020,, 2020,, 7190x FOR PEER REVIEW 1010 of of 15

implementation costs by substituting a few accelerometers with microphones while maintaining the thegood good noise noise attenuation attenuation performa performancence of ofthe the sHANC sHANC system, system, the the combination combination of of acceleration and acoustic signals as reference signalssignals cancan bebe considered.considered.

(a) (b)

Figure 7.7. A-weighted averaged sound pressure level ov overer two error microphones before and after control byby applyingapplying di differentfferent control control systems, systems, when when the the SUV SUV was was driven driven at a constantat a constant speed speed of 80 kmof 80/h onkm/h a rough on a rough road, (aroad,) using (a) ten using vibrational ten vibrational reference reference signals; signals; (b) using (b ten) using acoustical ten acoustical reference reference signals. signals. 4. Optimization of Attenuation Performance for Road Noise Control 4. Optimization of Attenuation Performance for Road Noise Control 4.1. Proposed Modified System 4.1. ProposedDue to the Modified difference System in the property of the vibration signals and acoustic signals, if these two types x of signalDue are to composedthe difference directly in the into property the reference of the vector, vibrationf , as signals shown and in Figure acoustic1, it willsignals, distort if these the input two signals and further bias the control effect of the controller. Therefore, these two parts of reference signals types of signal are composed directly into the reference vector, x f , as shown in Figure 1, it will need to be input independently. To overcome this phenomenon, a modified MIMO simplified hybrid distort the input signals and further bias the control effect of the controller. Therefore, these two parts ANC (msHANC) system was proposed, and the block diagram is shown in Figure8. Compared to of reference signals need to be input independently. To overcome this phenomenon, a modified the sHANC system, the msHANC system employs two feedforward structures containing filter matrix MIMO simplified hybrid ANC (msHANC) system was proposed, and the block diagram is shown in W f m and W f a, respectively. The former is called a microphone-based feedforward structure, which is Figure 8. Compared to the sHANC system, the msHANC system employs two feedforward mainly devoted to controlling the 70–85 Hz narrow peak response using microphone signals. The latter structures containing filter matrix W and W , respectively. The former is called a microphone- is an accelerometer-based feedforwardfm structure,fa which uses vibration signals to control the residual noisebased spectrumfeedforward between structure, 100 and which 500 Hz is thatmainly cannot devoted be controlled to controlling by the formerthe 70–85 structure. Hz narrow The K mpeak1 × vectorresponse of microphoneusing microphone signals andsignals. the ( KThea 1latter) vector is an of vibrationaccelerometer-based signals give thefeedforward full reference structure, signal which uses vibrationh iT signals to control the× residual noise spectrum between 100 and 500 Hz that x = x x d x vector, f f m f a , and the correlated disturbance is× produced by f via the matrix of transfer cannot be controlled by the former structure. The Km 1 vector of microphone signals and the responses, Pe. As mentioned in the previous section, it will be a feedback path Gn between control T sources(1)K × andvector reference of vibration microphones, signals andgive itthe is assumedfull reference to be signal modeled vector, perfectly xxx= in this paper,, and so thatthe a ffmfa the x f m obtained from experiments can be considered as the “true” reference acoustic signals. correlated disturbance d is produced by x f via the matrix of transfer responses, Pe . As In the msHANC system, the control signal vector ut0(n) can be expressed as: mentioned in the previous section, it will be a feedback path Gn between control sources and reference microphones, and it is assumed to be modeled perfectly in this paper, so that the x ut0(n) = u0f m(n) + u0f a(n) + ub0 (n) (19)fm obtained from experiments can be considered as the “true” reference acoustic signals. where u0 (n), u0 (n), and u0 (n) are control signal vectors′ derived from two feedforward structures In thef m msHANCf a system,b the control signal vector ut ()n can be expressed as: and one feedback structure, respectively. The signal generation in the msHANC system, including uu′′()=nnnn ()++ u ′ () u ′ () the updated equation of three controltfmfab filter matrixes, is similar to the sHANC system; for simplicity,(19) it is not re-explained′ ′ in this section.′ where u fm ()n , u fa ()n , and ub ()n are control signal vectors derived from two feedforward structures and one feedback structure, respectively. The signal generation in the msHANC system, including the updated equation of three control filter matrixes, is similar to the sHANC system; for simplicity, it is not re-explained in this section.

SensorsSensors 20202020,, 2020,, x 7190 FOR PEER REVIEW 1111 of of 15 15 Sensors 2020, 20, x FOR PEER REVIEW 11 of 15

x f x f Pe Pe Gn Gn d x fm - +d x ∑ Wfm + fm - + Wfm Gm + ∑ + Gm e v Adaptation +e v Ĝm algorithm Adaptation u fm Ĝm algorithm u+fm x fa + + ut u x fa Wfa +∑ t u fa Wfa + u fa + u Wb Ĝm b Wb Ĝm ub Adaptation Adaptation Ĝm algorithmAdaptation algorithmAdaptation Ĝm algorithm algorithm

Figure 8. Block diagram of the MIMO modified simplified hybrid active noise control system (the two FigureFigure 8.8. Block diagram of the MIMO modified modified simplifiedsimplified hybrid active noise control system (the(the twotwo feedforward structures are indicated by the dashed rectangles), mainly including two reference feedforwardfeedforward structures structures are are indicated indicated by theby dashedthe dashed rectangles), rectangles) mainly, mainly including including two reference two reference signals signals vectors x fm and x fa , error signals vector e , nominal plant responses matrices Gm and Gn , vectorssignals x vectorfm ands xfax,fm errorand x signalsfa , error vector signalse, nominal vector e plant, nominal responses plant matrices responsesGm matriand Gcesn, threeGm and control Gn , three control filter matrices W , W and Wb . filter matrices Wfm, Wfa and Wfmb. fa three control filter matrices W fm , W fa and Wb . 4.2.4.2. Performance Performance Assessment Assessment of of the the msHANC msHANC System System 4.2. Performance Assessment of the msHANC System TheThe spectrum ofof thethe averaged averaged sound sound pressure pressure level level measured measured at two at two error error microphones microphones after beingafter The spectrum of the averaged sound pressure level measured at two error microphones after beingcontrolled controlled by the by proposed the proposed msHANC msHANC system (usingsystem ten (using vibrational ten vibr referenceational reference signals and signals ten acoustical and ten being controlled by the proposed msHANC system (using ten vibrational reference signals and ten acousticalreference signals)reference is signals) shown inis Figureshown9 in, and Figure the results9, and the controlled results bycontrolled the sHANC by the system sHANC using system two acoustical reference signals) is shown in Figure 9, and the results controlled by the sHANC system usingtypes two of reference types of signalsreference independently signals independently are re-plotted are re-plotted for comparison. for compar Fromison. these From plots, these it canplots, be using two types of reference signals independently are re-plotted for comparison. From these plots, itseen can that be theseen proposed that the msHANC proposed system msHANC combined system the combined advantages the of usingadvantages vibration of referenceusing vibration signals it can be seen that the proposed msHANC system combined the advantages of using vibration referenceand acoustic signals reference and signalsacoustic alone; reference thatis, signals it can achievealone; that both is, high it can levels achieve of narrowband both high attenuation levels of reference signals and acoustic reference signals alone; that is, it can achieve both high levels of narrowbandbetween 70 andattenuation 85 Hz and between broadband 70 and attenuation 85 Hz and betweenbroadband 100 attenuation and 500 Hz. between The msHANC 100 and 500 system Hz. narrowband attenuation between 70 and 85 Hz and broadband attenuation between 100 and 500 Hz. Theachieved msHANC a maximum system overall achieved reduction a maximum of 4.1 dBA,overall which reduction is 0.5 dBA of 4.1 more dBA, overall which reduction is 0.5 dBA compared more The msHANC system achieved a maximum overall reduction of 4.1 dBA, which is 0.5 dBA more overallto the sHANC reduction using compared vibration to referencethe sHANC signals using alone. vibration reference signals alone. overall reduction compared to the sHANC using vibration reference signals alone.

FigureFigure 9. A-weightedA-weighted averaged averaged sound sound pressure pressure level level ov overer two two error error microphones microphones before before and and after after Figure 9. A-weighted averaged sound pressure level over two error microphones before and after controlcontrol byby sHANCsHANC system system using using two two types types of reference of reference signals signals independently, independently, the proposed the msHANCproposed control by sHANC system using two types of reference signals independently, the proposed msHANCsystem using system both using reference both signals,reference when signals, the when SUV wasthe SUV driven was at driven a constant at a speedconstant of speed 80 km /ofh 80 on msHANC system using both reference signals, when the SUV was driven at a constant speed of 80 km/ha rough on road.a rough road. km/h on a rough road. AsAs mentioned mentioned in in the the previous previous section, section, to to reduce reduce the the implementation implementation costs, costs, a a few few accelerometers accelerometers As mentioned in the previous section, to reduce the implementation costs, a few accelerometers were supposed to remove from selected reference sensors. Thus, Thus, it it is is interesting interesting to to predict predict the the noise noise were supposed to remove from selected reference sensors. Thus, it is interesting to predict the noise attenuation performance of the msHANC system as the number of vibration references decreases. To attenuation performance of the msHANC system as the number of vibration references decreases. To

Sensors 2020, 20, x FOR PEER REVIEW 12 of 15 demonstrate the effectiveness of the proposed system, the performance of the sHANC system using vibration reference signals was also predicted as the number of vibration references increased. The two control systems both contained ten vibration signals at the first simulation, and the results are shown in Figure 10. It is worth highlighting that the selection of vibration references in each adjustment depended on the ranking list obtained previously. From the results in Figure 10, it can be seen that when ten microphone signals and the top ten vibration signals were integrated into the full reference signal vector x f for the msHANC system (Figure 8), the system could achieve slightly more noise reduction compared with the sHANC system using the top twenty vibration signals. This meant that if the number of reference signal channels was the same, ten accelerometers could be replaced by ten microphones among twenty accelerometers to obtain similar attenuation performance. Moreover, it can be seen that when ten microphoneSensors 2020, 20 ,signals 7190 and the top six vibration signals were integrated into the full reference signal12 of 15 vector, the msHANC system achieved the same overall noise reduction compared with the sHANC system using the top ten vibration signals, which obtained an overall reduction of approximately 3.7 dBA.attenuation In this performancecase, this meant of thethat msHANC four reference system accelerometers as the number among of vibration ten essential references accelerometers decreases. couldTo demonstrate be substituted the e byffectiveness ten microphones of the proposed while maintaining system, the the performance same attenuation of the sHANC performance. system Based using onvibration the comparison, reference signalsit can be was concluded also predicted that the as msHANC the number system of vibration using both references microphone increased. signals The and two vibrationcontrol systems reference both signals contained can improve ten vibration the attenuat signals ation the performance first simulation, significantly. and theresults It also areprovided shown thein Figurepotential 10. to It isfurther worth reduce highlighting the required that the number selection of of accelerometers vibration references by using in eacha few adjustment low-cost microphones,depended on thesuch ranking as MEMS list microphones. obtained previously.

FigureFigure 10. 10. TheThe predicted predicted performance performance (overall (overall reduct reduction)ion) of two of twocontrol control systems systems with withthe increasing/decreasingthe increasing/decreasing of the of number the number of vibration of vibration references references (the (thetwo twocontrol control systems systems both both contain contain ten vibrationten vibration signals signals at the at beginning). the beginning).

4.3. EffectFrom of theTime results Delays in on Figure Performance 10, it can be seen that when ten microphone signals and the top ten vibration signals were integrated into the full reference signal vector x f for the msHANC system For the practical implementation of the msHANC system, the delays of the real-time system (Figure8), the system could achieve slightly more noise reduction compared with the sHANC system need to be considered, which mainly consist of two types. As discussed [15,20], one type is caused by using the top twenty vibration signals. This meant that if the number of reference signal channels was electronic, such as analog-to-digital/ digital-to- analog (AD/DA) conversions, antialiasing/smoothing the same, ten accelerometers could be replaced by ten microphones among twenty accelerometers to filters, and reference sensors, and it has been included in the obtained secondary transfer responses obtain similar attenuation performance. Moreover, it can be seen that when ten microphone signals and reference signals. The other type of delay is caused by digital signal processing, especially in the and the top six vibration signals were integrated into the full reference signal vector, the msHANC multichannel signal convolution operations. To investigate the effect of processing delays on the system achieved the same overall noise reduction compared with the sHANC system using the top msHANC system, a few additional delays were added into the reference signals, and the performance ten vibration signals, which obtained an overall reduction of approximately 3.7 dBA. In this case, this of the msHANC system using microphone signals and the top six vibration signals, which were meant that four reference accelerometers among ten essential accelerometers could be substituted by verified to have the same attenuation capacity as the sHANC system using the top ten vibration ten microphones while maintaining the same attenuation performance. Based on the comparison, signals, was analyzed off-line, as shown in Figure 11. The performance of the sHANC system using it can be concluded that the msHANC system using both microphone signals and vibration reference vibration signals was also demonstrated for comparison. From these results, it can be seen that the signals can improve the attenuation performance significantly. It also provided the potential to msHANC system was more sensitive to the delays, which may be due to the microphone signals used further reduce the required number of accelerometers by using a few low-cost microphones, such as MEMS microphones.

4.3. Effect of Time Delays on Performance For the practical implementation of the msHANC system, the delays of the real-time system need to be considered, which mainly consist of two types. As discussed [15,20], one type is caused by electronic, such as analog-to-digital/digital-to-analog (AD/DA) conversions, antialiasing/smoothing filters, and reference sensors, and it has been included in the obtained secondary transfer responses and reference signals. The other type of delay is caused by digital signal processing, especially in the multichannel signal convolution operations. To investigate the effect of processing delays on the msHANC system, a few additional delays were added into the reference signals, and the performance of the msHANC system using microphone signals and the top six vibration Sensors 2020, 20, 7190 13 of 15 signals, which were verified to have the same attenuation capacity as the sHANC system using the top ten vibration signals, was analyzed off-line, as shown in Figure 11. The performance of the sHANC system using vibration signals was also demonstrated for comparison. From these results, it can be Sensors 2020, 20, x FOR PEER REVIEW 13 of 15 seen that the msHANC system was more sensitive to the delays, which may be due to the microphone insignals the msHANC used in the system msHANC being system collected being from collected the floor from inside the floor the inside cabin, the while cabin, the while sHANC the sHANCsystem usedsystem only used the onlyvibration the vibration signals obtained signals obtainedfrom suspen fromsions suspensions or wheels, or which wheels, lost which the causal lost theconstraint causal moreconstraint easily more as delays easily increased. as delays In increased.this case, the In overall this case, reduction the overall of the msHANC reduction system of the msHANCdecreased bysystem approximately decreased by1 dB approximately when additional 1 dB delays when additionalexceeded 10 delays ms. Moreover, exceeded 10 it ms.was Moreover,also shown it that was whenalso shown the additional that when delays the additional exceeded delays20 ms, exceededboth control 20 ms, systems both controlseemed systemsto lose the seemed capacity to lose to controlthe capacity the disturbance to control theand disturbance the overall andreduction the overall was reduced reduction to was1 dBA. reduced Therefore, to 1 dBA.in the Therefore, practical implementation,in the practical implementation, it is recommended it is recommended to choose the to optimal choose thecontroller optimal and controller hardware and hardwareto keep the to processingkeep the processing time to less time than to less10 ms than to maintain 10 ms to maintainthe system’s the system’seffectiveness. effectiveness.

Figure 11. TheThe predicted predicted overall overall reduction reduction performance performance of of two two control control systems systems when when different different delays are added into reference signals.

5. Conclusions Conclusions A local active active noise noise control control system system can can achieve achieve effective effective control control of broadband broadband random road noise when the the reference reference signals signals are are highly highly correlated correlated wi withth the the primary primary disturbance. disturbance. However, However, due due to the to complexthe complex propagation propagation path path between between structural structural and and acoustic acoustic noise, noise, it itis is not not practical practical to to obtain obtain great great coherence in thethe frequencyfrequency bandband ofof interest.interest. ToTo overcome overcome this this limitation, limitation, in in this this paper, paper, the the potential potential of ofapplying applying a multichannel a multichannel simplified simplified hybrid hybrid active active noise noise control control (sHANC) (sHANC) system system to the to rear the left rear seat left of seata large of a SUV large has SUV been has investigated. been investigated. Furthermore, Furthermore, to reduce to reduce the practical the practical implementation implementation costs whilecosts whilemaintaining maintaining great reduction great reduction performance, performance, a modified a modified simplified simplified hybrid control hybrid (msHANC) control (msHANC) system has systembeen proposed has been firstly. proposed The mainfirstly. conclusions The main conclusions are as follows: are as follows: • The attenuation performance of the multichannel sHANC system was off-line off-line analyzed, using • the disturbance signals along with two types of referencereference signals:signals: (a) (a) ten vibration reference signals provided provided by by accelerometers accelerometers on on the the suspensions suspensions and and bodywork bodywork of the of vehicle, the vehicle, which which were werechosen chosen out of out 36 options; of 36 options; (b) ten (b) acoustical ten acoustical reference reference signalssignals provided provided by microphones by microphones on the flooron the of floor the cabin. of the cabin.It has been It has shown been shown that, comp that, comparedared to the to traditional the traditional feedforward feedforward system, system, the sHANCthe sHANC system system using using two two types types of reference of reference signals signals can can produce produce a significant a significant suppression suppression of theof the narrowband narrowband peak peak noise noise between between 75 75 and and 80 80 Hz, Hz, but but the the sHANC sHANC system system lost lost the the control capability in a broadband range of 100–500 Hz by using acoustical reference signals. It was also shown that compared to traditional hybrid co controlntrol systems, the sHANC system can achieve similar noise reduction. • The msHANC system using both vibrational reference signals and acoustical reference signals was proposed firstly. The system cleverly combined the benefits of using two different types of reference signals. The results showed that four reference accelerometers can be substituted by ten microphones without damaging the control performance, achieving around 3.7 dBA overall reduction. The system provided the potential to reduce the required number of accelerometers by using a few low-cost microphones, such as micro-electro-mechanical system (MEMS) microphones, which can further reduce the practical implementation costs.

Sensors 2020, 20, 7190 14 of 15

The msHANC system using both vibrational reference signals and acoustical reference signals • was proposed firstly. The system cleverly combined the benefits of using two different types of reference signals. The results showed that four reference accelerometers can be substituted by ten microphones without damaging the control performance, achieving around 3.7 dBA overall reduction. The system provided the potential to reduce the required number of accelerometers by using a few low-cost microphones, such as micro-electro-mechanical system (MEMS) microphones, which can further reduce the practical implementation costs. The effect of time delays on the reduction performance of the proposed systems was investigated. • It has been shown that the msHANC was more sensitive to the delays compared to the sHANC system using only vibrational reference signals. When the additional delays exceeded 20 ms, the msHANC system lost the capacity to control the disturbance.

Author Contributions: Software, Z.J.; investigation, X.Z.; data curation, Q.Z.; writing—original draft preparation, Z.J.; writing—review and editing, X.Z. and Z.H.; supervision, Y.Q.All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by the National Natural Science Foundation of China, grant number 51705454 and 51876188. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Chu, Z.; Wang, H.; Chen, C.; Yan, H.; Kang, R. Source path contribution analysis for vehicle indoor pass-by noise. Sound Vib. 2017, 51, 2–6. 2. Liu, N.; Sun, Y.; Wang, Y.; Guo, H.; Gao, B.; Feng, T.; Sun, P. Active control for vehicle interior noise using the improved iterative variable step-size and variable tap-length LMS algorithms. Noise Control Eng. J. 2019, 67, 407–414. [CrossRef] 3. Elliott, S.J.; Nelson, P.; McDonald, A.M.; Quinn, D.C.; Saunders, T. The active control of engine noise inside cars. In Proceedings of the INTER-NOISE, Avignon, France, 30 August–1 September 1988; pp. 987–990. 4. Hasegawa, S.; Tabata, T.; Kinoshita, A.; Hyodo, H. The Development of an Active Noise Control System for Automobiles. Soc. Auto Eng. 1992, 922086. [CrossRef] 5. Schirmacher, R.; Kunkel, R.; Burghardt, M. Active Noise Control for the 4.0 TFSI with Cylinder on Demand Technology in Audi’s S-Series. Soc. Auto Eng. 2012, 1533. [CrossRef] 6. Cheer, J. Active Control of the Acoustic Environment in an Automobile Cabin. Ph.D. Thesis, University of Southampton, Southampton, UK, 2012. 7. Duan, J.; Li, M.; Lim, T.; Lee, M.-R.; Cheng, M.-T.; Vanhaaften, W.; Abe, T. Combined Feedforward Feedback Active Control of Road Noise Inside a Vehicle Cabin. J. Vib. Acoust. 2014, 136, 041020. [CrossRef] 8. Padhi, T.; Chandra, M.; Kar, A.; Swamy, M.N. Design and analysis of an improved hybrid active noise control system. Appl. Acoust. 2017, 127, 260–269. [CrossRef] 9. Wu, L.; Qiu, X.; Guo, Y. A simplified adaptive feedback active noise control system. Appl. Acoust. 2014, 815, 40–46. [CrossRef] 10. Wu, L.; Burnett, I.; Guo, Y. Decoupling feedforward and feedback structures in hybrid active noise control systems for uncorrelated narrowband disturbances. J. Sound Vib. 2015, 350, 1–10. [CrossRef] 11. Sutton, T.; Elliott, S.; McDonald, A.; Saunders, T. Active control of road noise inside vehicles. Noise Control Eng. J. 1994, 42, 137–147. [CrossRef] 12. Oh, S.-H.; Kim, H.; Park, Y. Active control of road booming noise in automotive interiors. J. Acoust. Soc. Am. 2002, 111, 180–188. [CrossRef][PubMed] 13. Mohammad, J.; Elliott, S.; Mackay, A. The performance of active control of random noise in cars (L). J. Acoust. Soc. Am. 2008, 123, 1838–1841. [CrossRef][PubMed] 14. Cheer, J.; Elliott, S. Multichannel control systems for the attenuation of interior road noise in vehicles. Mech. Syst. Signal Proc. 2015, 60–61, 753–769. [CrossRef] 15. Jung, W.; Elliott, S.; Cheer, J. Local active control of road noise inside a vehicle. Mech. Syst. Signal Proc. 2018, 121, 144–157. [CrossRef] 16. Elliot, S.J. Signal Processing for Active Control; Academic Press: London, UK, 2001. Sensors 2020, 20, 7190 15 of 15

17. Mazur, K.; Wrona, S.; Pawełczyk, M. Active noise control for a washing machine. Appl. Acoust. 2019, 146, 89–95. [CrossRef] 18. Jung, W. Mid-Frequency Local Active Control of Road Noise. Ph.D. Thesis, University of Southampton, Southampton, UK, 2018. 19. Cheer, J.; Elliot, S.J. Active noise control of a diesel generator in a luxury yacht. Appl. Acoust. 2016, 105, 209–214. [CrossRef] 20. Bai, M.R.; Lin, Y.; Lai, J. Reduction of electronic delay in active noise control systems—A multirate signal processing approach. J. Acoust. Soc. Am. 2002, 111, 916–924. [CrossRef][PubMed]

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).