Implementation of an Adaptive Feedback Canceller for Hearing Aid - Assignment Example

IMPLEMENTATION OF AN ADAPTIVE FEEDBACK CANCELLER FOR HEARING AID Name: Course: Instructor: Institution: Date of Submission: ABSTRACT Objective: The research paper explores the implementation of adaptive feedback canceller for hearing aids. The adaptive filter in adaptive feedback cancellation is regularly employed for modelling the acoustic feedback pathway amid the loudspeaker of the hearing device and the microphone. Methods: Adaptive filters are a significant parameter as they provide an exchange between fast conjunction and low, stable state misalignment. The research paper defines the cancellation feedback process on hearing devices. It is the modernized hearing tool. AFC can be used for howling hearing aids suppression without experiencing any environmental influences. Results: Feedback cancellation diverges from other known feedback reduction methods, as a forward gain reduction method. The increase is perceived in the device while though it is used in similar formats. The result observed is when the feedback is reduced through an internal estimation of the actual device signal, which is then subtracted to attain the microphone signal approximation. The AFC limits the signal degradation because of the feedback signal. The signal reference is correlated with the feedback signal, not the input signal. Attaining such signals is hard as related to hearing aids. The output signal can be used for AFC cancellation. The compressor frequency is utilized as a decorrelator. Conclusion: The results indicate that the proposed system and proper system choice parameters for deteriorating the output signal quality. It also focuses on refining the approximation of the adaptive feedback cancellation as contained in the hearing devices while utilizing the forecast error technique (PEM) Keywords: Adaptive Feedback Cancellation, Feedback Reduction, Hearing Aid & receiver, Feedback Signal, PEM, INTRODUCTION Adaptive Feedback Cancellation is the best approach available for cancelling the audio feedback process perceived in acoustic systems such as the hearing aids in this study. Acoustic feedback systems encounter leakage paths that are eliminated with the implementation of feedback cancellation. Adaptive feedback canceller is an input signal that is used in acoustic systems such as hearing aids. It is used for cancelling the recurrent distortion in the artifacts as entrainment is generated. Adaptive feedback cancellation has its applications employed in echo cancellation where the error between the actual and desired output is provided as the adaptive feedback to the processor. It is also used in the process of adjusting the coefficients of the cancellation with the intention of minimizing error (Spriet, Simon, Marc, & Jan 2008). Feedback in hearing aids arises once the microphone of the hearing aid captures the receiver also identified as the loudspeaker signal. The response/signal is amplified, which affects the device by leading to shared loops within the system. Thus, once feedback is attained, it leads to loud tonal signals that are disturbing. Similarly, feedback can mainly be achieved on the stipulation that the hearing volume of the device is increased. Thus, the algorithms are tactics employed for adaptive feedback cancellation with the purpose of estimating the path of transmission between the microphone and the loudspeaker. Thus, the approximation attained is utilized for implementing a counteracting electric feedback pathway that overpowers the feedback signal of the tone. Research Problem and Objective Recently, patients encountering hearing problems or using the hearing devices has steadily increased. Furthermore, the devices used to assist these patients have been advanced over the years while most of them encounter problems with the quality of performance they present. The research shows that improving the application of adaptive feedback cancellation improves the performance of the devices leading to the development of quality hearing aid devices. Some hearing aids can be used to support hearing of those with impaired hearing. It includes the utilization of open fitting without an earmold. Thus, the hearing aids are one of the devices that will improve the problem of hearing among patients. The research objective is to present an analysis of the literature and research made on the hearing aid to assist the growing number of patients with the problem. The study shows some of the devices with the utilization of technique as provided in the paper. The devices range from open-fitting aids that are vulnerable to audio feedback mainly apparent as the annoying whistling in the acoustic hearing devices among others as the 3 On Semiconductor. The acoustic feedback cancellation requires a fast and vigorous algorithm for feedback cancellation. There exist some solutions to minimize acoustic feedback. (AFC) Adaptive feedback cancellation exists as the most recognized and utilized resolution. To approximate the urge response occurring within the hearing aid loudspeaker and the hearing aid microphone the adaptive filter is employed in the acoustic response pathway of an adaptive feedback cancellation. Research Methodology/ Plan The study involves using literature and research to present the advancements and devices that are developed for patients with a hearing aid. That is; it includes presenting the significant developments that have occurred regarding the devices. It also shows how the performance and quality of the devices to increase hearing performance have developed. The research presents the challenges and advancements that the hearing devices have encountered over the years. The first methodology is research, which will then be followed by a practical study/experimental process of the devices to understand the adaptive feedback cancellation implementation in hearing devices. The plan of the research is to review numerous articles and previous research on the hearing devices adaptive feedback/response cancellation. The theoretical model ensures that before fully engaging in a period, one can understand the possible challenges, gains, impulses, frequencies and styles to expect. More importantly, it informs one of what to expect when practically dealing with the hearing devices. My plan is to understand the implementation and operation of the devices in detail as presented here, with the intention of providing viable recommendations on how to improve the performance of the devices in future. LITERATURE ANALYSIS Adaptive Feedback Cancellation Typically, normalized minimum mean square procedure or the slightest mean squares system are used to evaluate the reaction deriving from the acoustic path of response. However, the projected impulse response is usually biased out of the relationship amid the incoming sign and the feedback signal (Williamson, Bustamante, & Worrall, 1989). The hearing devices acoustic response is identified as the acoustical combination between the earpiece (amplifier) and the microphone of the earshot device. The figure one below displays the issue linked to the response of acoustic path when dealing with devices that use only a particular microphone (Spriet, Proudler, Moonen, & Wouters, 2005). Forward path G (q) =goq0 + - - - - + gLG-1q-LG+1 where LG signifies the filter length titled G (q), which embodies the systematic path of the signal linked to the device also called the frequency-specific gain, density and noise decline. The results are accurate on the supposition that G (q) has a predelay also defined as dG of one sample. That is; it is presented as go = 0. F (q, n) denotes the path of response of the amplifier and the receiver. The microphone and amplifier hints are recorded as y [n] and u1 [n] correspondingly. The response pathway signal is given as v [n] = F (q, n) u1 [n] (Onsemi) while the signal anticipate is presented by x [n]. The acoustic response amplifies the sound sent by the receiver and sent into the microphone leading to the closed-loop system (Williamson, Bustamante, & Worrall, 1989). G (q) 1-G (q) F (q, n) The signal recorded as x [n], which is the adaptive feedback, is sent to the loudspeaker (receiver) signal u1 [n] for the loop system. However, the loop unsteadiness transpires to a gain expressed as |G (eiw) F (eiw)|, which surpasses one with a pointed regularity ω = [0,π], where the phase of the loop is equal to a manifold of 2π [1]. U1 [n] Forward Path G (q) F (q, n) audio response pathway Y[n] v[n] X [n] (Onsemi) Figure 1. Acoustic feedback Acoustic feedback as presented in the figure above occurs when the sound is intensified through the hearing device into the earl canal. The sound then leaks out. The microphone (loudspeaker), which amplifies the signal, again picks the sound. The amplified signal route travels back to the microphone while following the feedback path. The response cancellation scheme needs to adapt to the feedback instantaneously before it can be noticed. The cancellation process increases the accuracy and efficiency of the device that upsurges the sound quality (Williamson, Bustamante, & Worrall, 1989). The stability of the devices is determined and affected by the distance it takes for the receiver to send the signal to the microphone as described in the essay. Other factors include the physical terms of the devices, the acoustic energy propagation of the devices, distance of the devices signal and other constraints among other factors (Spriet, Proudler, Moonen, & Wouters, 2005). Additionally, the position the microphone is placed also influences the success of the device to pick and send the sound as depicted to support speech understanding and optimizing the directional microphone effects. Consequently, the need for stability within the devices for increased performance is implemented with the adaptive feedback cancellation process. To minimize the undesirable consequences presented through the response of acoustic where some processes are perceived. The approaches are largely categorized through the process of feedforward conquest and feedback termination methods. The regular signal processing path is linked to the feedforward suppression in the hearing devices also denoted as G (q) where the steadiness of the device is attained through F (q,n). The use of a notch filter being the most common technique is implemented in the signal processing progression. Feedback cancellation algorithm is a more capable clarification for the acoustic response. Figure 2 below presents, it describes the feedback cancellation algorithm (Onsemi). The response canceller F (q,n) yields an approximate of z [n] of the signal response v [n] and deducts the estimated z [n] from the microphone signal. Thus, the single ignal attained is conserved as G (q), which is also the input (Onsemi). U1 [n] U[n] Response Forward G (q) F (q,n) cancellation path F(q,n) audio response path path z [n] v [n] e [n] y [n] x [n] Figure 2. Adaptive response canceller. TECHNICAL ADVANCEMENTS Adaptive Feedback Cancellation On Semiconductor Feedback cancellation processes, features and algorithms will be discussed in this section regarding the latest ON Semiconductor digital hearing aid amplifiers. Unlike other existing feedback reduction approaches, the feedback cancellation exists as a tactic that guarantees the stability of the devices forward gain under standard situations of operation. That is; feedback cancellation process has an algorithm that operates by decreasing the inner feedback approximation formulate leading to the proper hearing aid feedback signal. Consequently, the approximate is then deducted from the signal provided by the microphone leading to the approximate and accurate preferred signal. The device does not engage in any form of squeal, as the actual feedback process is cancelled (Spriet, Simon, Marc, & Jan 2008). Thus, the accurate feedback cancellation process occurs as explained. Adaptive feedback cancellation three titled the ON Semiconductor exists as a proper response/phase cancellation. The procedure functions by deducting the inner assessment of the response signal of the device from the microphone signal. The inner response signal is achieved when the output signal of the device through an inner pathway of response signal. The model of the pathway in adaptive feedback canceller 3 is intended to equal the response time of the peripheral feedback pathway. The positive impulse response is the signal to be detected at the output level of the microphone if the hearing device formed an impulsive signal. As an illustration, the time measured response of a representative BTE hearing device occurs as perceived below (Onsemi). The graph below displays a primary period of exactly 0.5 – 0.75 ms with miniature vitality. The time opening relays signals from analog to digital or vice versa, conversion predelays as well as auditory predelays brought by the features of the device or position such as sound propagation, tubing, and transducers through the air. The total consequence of the algorithm from the feedback canceller is to eradicate a definite share of the impulse response (Williamson, Bustamante, & Worrall, 1989). Due to limited computational resources inside the hearing aid, a quota of the next response route can be negated. To reach better efficiency of the adaptive feedback cancellation, the response path should consist the cancelled portion with prevalent energy. For instance, constraining the cancellation space to 1ms, then the pathway of response as perceived in the figure 3-a, the inner model of response should be similar to the graph in figure 3-b. On the stipulation that it is deducted from the response pathway, the attained total is as shown in figure 3-c. The equivalent regularity response is exposed in figure 4. Therefore, the response canceller in this illustration should offer About 12 dB of additional gains (ASG) attained devoid of an influence on the devices forward gain (Leber & Schaub, 2006). Unluckily, the response path recorded in the hearing device lacks continuity. Notable inconsistency in the response path of various hearing supports such as acoustic tubing, transducer assortment, and earmold styles among others. Besides, the response path of a specified hearing tool can alter with time because of user movement and ecological variations. To uphold ideal cancellation outputs, adaptive feedback cancellation 3 executes constant reforms of the inner response pathway model. These apprises will befall routinely and exploit present signals from the device, therefore eliminating the necessity of intervention leading to minimal disruption for the hearing-aid wearer (Onsemi). Figure3: BTE time response measurements (Onsemi) Figure 3. BTE impulse response measurements: a) Overall exterior feedback path, b) Feedback canceller inner pathway c) Feedback canceller net outcomes Figure 4 present the process of change in the response path incidence earlier and later the application of feedback canceller Influence of the Hearing Aid Style and Optimizing Performance Some system-level factors sway the algorithm response canceller performance of the hearing utility. Feedback Canceller presentation is predisposed by the acoustic response, hearing-aid style, audio signal bandwidth, interactions with other hearing aid features and input signal characteristics. The following are strategies to achieve optimum performance from the feedback canceller. The delay is due to the required time for the sound to broadcast where dissimilarities are perceived from the ITC to a BTE. The auditory transmission time increases to the delays perceived at the converter and those at the impulse response attained from the feedback pathway. Compensation through internal model is required to increase canceller performance. Figure 5 shows an illustration of impulse responses of ITE, ITC, and BTE. It is vital to identify the sway of the impulse responses that ensues from the hearing devices styles with the superior loudspeaker/microphone separation (Onsemi). Figure 5. Shows impulse response for the three different hearing aids. ITC, ITE, BTE Adaptive feedback cancellation three uses diverse hearing devices styles from ON semiconductor through a modifiable audio pre-delay factor (Onsemi). The factor permits the enhancement of the feedback canceller time location operation for various hearing device styles. Figure 6 shows an illustration of the impact of the acoustic pre-delay on the adaptive feedback cancellation three performance, additional stability advances were calculated for several hearing aids styles. The styles include; ITE, ITC, BTE, BTE-open and receiver in the ear (RITE). (Onsemi, pg 5) Figure 6. Stability gains measurements for devices styles selection regularity. The graph above presents the acoustic predelay results of an experiment (Onsemi) presented for different hearing aid styles. The maximum added stable gains measured in each case, as shown in Table 1, adaptive feedback cancellation 3 provides up to 30 dB of added stable advances, subject to the hearing-aid style and audio sample rate. Table 1. Measured additional stable gain from the diverse hearing aid styles These dimensions signify a solitary hearing aid in the style categories. As publicized in Table 2, auditory pre-delay is diverse for the two frequency samples presented. Indeed, the aural pre-delays for 16kHz are roughly partial those of 32kHz. Hearing Devices Styles The appropriate acclaimed acoustic predelay The distance of the mic receiver (cm) 16 kHz 32kHz BTE 13 25 14 ITC 8 15 2.5 CIC 7 14 2 ITE 8 15 3.5 RITE 10 20 8 Table 2. Endorsed audio pre−delay values for various hearing devices styles Acoustic Response Hearing aids with similar styles have significant performance difference. Time response is a term that refers to where some feedback paths of some devices operate with a short time duration in sending and receiving the feedback. That is; the feedback canceller can eliminate a big portion of the net energy. On the other hand, feedback paths with long time duration engage in the process of eradicating the vigor to a smaller portion of the net. The process results in a lower added stable gain. Thus, the unsteadiness of the devices is not experienced, and performance is improved, and efficient (Leber & Schaub, 2006). Figure 7 below exists as an illustration of the two ITE styles responses. However, from the top graph, the response has a more duration than the reaction revealed in the graph below (figure 8). Figure 8 shows the feedback path frequency responses for similar systems, from the chart it shows clearly that, the first style of ITE has a lengthier time leading to a weak impulse response frequency plot. Various aspects can contribute to such a peaky feedback-path response. The factors may range to sharp resonances encountered in the transducer responses (mostly in loudspeakers/receivers), and the vibration coupling problems (Onsemi). Figure 7 & 8. As presented in the figure above, these are algorithms of two devices with similar physical sizes but different performances. Both of the devices employ the ITE style of response. Thus, the figure presents their impulse responses. The first graph figure 7 as presented takes a longer time to receive feedback while the second one takes a shorter time. The information in the graph presents that the second graph (figure 8) as it takes less time to receive feedback. Thus, the second one due to adaptive feedback cancellation leads to stable gains and encounters less unsteadiness. On the other hand, the first one is less stable and presents major loops. Therefore, it lacks the feedback canceller (Onsemi). The graphs also provide the information that figure 7 has a sharper response peak linked to the impulse feedback and the increased stability gain in the device. Figure 8 is the opposite presenting less stability, loops that lead to unsteadiness and less performance. Additionally, the performance of the devices maybe linked to the acoustic tube resonance. However, these factors steer to feedbacks with high signals and eventually degrade the performance of the feedback canceller. Therefore, it is only important if such resonances are eliminated in the system design, for optimum feedback canceller performance (Williamson, Bustamante, & Worrall, 1989). Additionally, the signal deriving from the microphone y [n] derives from the signal stated as x [n] and the link to the signal response f [n]. That is; the method can be presented in a mathematical calculation format as follows: y [n] = f [n] + x [n] f [n] is the density of the response pathway signal of the acoustic titled h [n] and it exists as a limited impulse reaction filter of length Lh. Therefore, it is clear that f [n] can be expressed as in the procedure below: f [n] = h T [n] u [n]. With h [n] = [h0 [n] h1 [n] : : : hLh-1 [n] T and u [n] = [u [n] u [n-1] . . . . . u [n - Lh+1] T. (Onsemi) The filter h[n] can be symbolized as a polynomial function of transfer in q, i.e. H (q,n) = h T [n] q with q = [1q-1 . . . . .q-Lh+1] T (Onsemi). Hence, f [n] can be represented as F [n] = H (q,n) u [n] in the figure below (Onsemi) Figure 9 u[n] G (q,n) H (q,n) H (q,n) F[n] e [n] y[n] x[n Figure 9 above presents a schematic process of a standard adaptive feedback cancellation system. The signal for error e[n] corresponds to an approximation of the inward x[n] signal and is computed as E [n] = y [n] – H (q,n) u [n] Where H (q,n) is provided as an estimate of H (q,n). H (q,n) is the loudspeaker(receiver) signal where u [n] is then calculated by amplifying e [n] using the forward path gain G (q,n). The process can be presented mathematically through the following formula: U [n] = G (q,n) e [n], Where it is routinely assumed that G (q,n) contains a delay that is also displayed as dG ≥ 1 [3,4]. Desired Signals of Adaptive Feedback Cancellers Models and Tracking Performance Improvement Using the Shadow Filter Approach The adaptive feedback cancellation techniques exploit previous information regarding the response path to develop the estimate precision of the average continuous adaptive feedback. These practices have the capability of engaging signal distortion decrease, which is trafficked beside the model fitness differences in the response pathway. However, the process is also considered biased as the standard continuous adaptive feedback is condensed by including a fixed or unstable archetypal of the anticipated signal x [n] in the recognition (Leber & Schaub, 2006). The preferred model signal utilized to predict the anticipated signal component in the receiver (loudspeaker) and the microphone signals (Spriet, Proudler, Moonen, & Wouters, 2005). The methodology has been espoused from undeviating closed-loop structure acknowledgement with the prediction error method (Forssell & Ljung, 1999). The feedback canceller necessities a relaxed merging speed, small phase proportions, to obtain precise response path estimation at angular regularities with a lesser round gain. With these rates, the signal desired power to feedback ratio signal is slight, which gives a big error of the feedback canceller. To improve or enhance the tracking of outcomes of the prediction error method, a second quicker adaptive filter is utilized. It is positioned parallel to the adaptive response canceller (Mader, Puder, & Schmidt, 2000). Once a variation in feedback path transpires, the filter shadow converges faster than the feedback canceller does. Thus, producing a healthier response path estimate. Conclusion & Results Adaptive feedback cancellation three from ON Semiconductor is an accurate response cancellation procedure, eliminating auditory feedback devoid of conceding the hearing device forward gain. It accomplishes additional steady gain (ASG) of up to 30dB whereas giving protection to sound feedbacks as well as least entrainment relics. The procedure is intended to cohabit with added adaptive hearing tool features and permits the devices designers to realize the elasticity to enhance outcomes for specific products. In this research, it has been proposed an ideal adaptive feedback cancellation system that uses the two adaptive filters from the combination of an affine with various step sizes to give fast merging and the stable misalignment of the collective filter. The results indicate that the proposed system and proper system choice parameters for deteriorating the output signal quality. It also focuses on refining the approximation of the cancellation with prediction error method (PEM) in hearing devices adaptive feedback. References Bershad, N. J., Bermudez, M. J., & Tourneret, Y. J. (2008). An Affine Combination Adaptive Filters - Transient Mean-Square Analysis. IEEE Transactions on Signal Processing, 56(5), 1853-1864. Forssell, U., & Ljung, L. (1999). Closed-loop identification revisited . Elsevier Automatica, 35(7), 1215-1241. Leber, R., & Schaub, W. (2006). Circuit and method for the adaptive suppression of an acoustic feedback, US Patent, US 6611600, 2003. . ARTICLE IN PRESS 572 A. Spriet et al. / Journal of the Franklin Institute, 545–573. Mader, A., Puder, H., & Schmidt, U. G. (2000). Step-size controls for acoustic echo cancellation filters . Journal of Signal Processing - Special Issue on current topics in adaptive filtering for hands-free acoustic communication and beyond, 80(9), 1697-1719. Onsemi. (n.d.). Adaptive Feedback Cancellation 3 from ON Semiconductor. AND9020/D, 1-10. Retrieved from http://www.onsemi.com/pub/Collateral/AND9020-D.PDF Spriet, A., Proudler, I., Moonen, M., & Wouters, J. (2005). Adaptive feedback cancellation in hearing aids with linear prediction of the desired signal. IEEE Transactions on Signal Processing, 53(10), 3749-3763. Spriet, A., Simon, D., Marc, M., & Jan, W. (2008). Feedback control in hearing aids. In Springer Handbook of Speech Processing (pp. 979-1000). Berlin, Germany: Springer Berlin Heidelberg. Williamson, M. J., Bustamante, D. K., & Worrall, T. L. (1989). Measurement and adaptive suppression of acoustic feedback in hearing aids. Glasgow, Scotland, 3, 2017–2020. Read More