<p>Hearing loss afflicts as many as 10\% of our population.Fortunately, technologies designed to alleviate the effects ofhearing loss are improving rapidly, including cochlear implantsand the increasing computing power of digital hearing aids. Thisthesis focuses on theoretically sound methods for improvinghearing aid technology. The main contributions are documented inthree research articles, which treat two separate topics:modelling of human speech recognition (Papers A and B) andoptimization of diagnostic methods for hearing loss (Paper C).Papers A and B present a hidden Markov model-based framework forsimulating speech recognition in noisy conditions using auditorymodels and signal detection theory. In Paper A, a model of normaland impaired hearing is employed, in which a subject's pure-tonehearing thresholds are used to adapt the model to the individual.In Paper B, the framework is modified to simulate hearing with acochlear implant (CI). Two models of hearing with CI arepresented: a simple, functional model and a biologically inspiredmodel. The models are adapted to the individual CI user bysimulating a spectral discrimination test. The framework canestimate speech recognition ability for a given hearing impairmentor cochlear implant user. This estimate could potentially be usedto optimize hearing aid settings.Paper C presents a novel method for sequentially choosing thesound level and frequency for pure-tone audiometry. A Gaussianmixture model (GMM) is used to represent the probabilitydistribution of hearing thresholds at 8 frequencies. The GMM isfitted to over 100,000 hearing thresholds from a clinicaldatabase. After each response, the GMM is updated using Bayesianinference. The sound level and frequency are chosen so as tomaximize a predefined objective function, such as the entropy ofthe probability distribution. It is found through simulation thatan average of 48 tone presentations are needed to achieve the sameaccuracy as the standard method, which requires an average of 135presentations.</p>

Hearing loss afflicts as many as 10\% of our population.Fortunately, technologies designed to alleviate the effects ofhearing loss are improving rapidly, including cochlear implantsand the increasing computing power of digital hearing aids. Thisthesis focuses on theoretically sound methods for improvinghearing aid technology. The main contributions are documented inthree research articles, which treat two separate topics:modelling of human speech recognition (Papers A and B) andoptimization of diagnostic methods for hearing loss (Paper C).Papers A and B present a hidden Markov model-based framework forsimulating speech recognition in noisy conditions using auditorymodels and signal detection theory. In Paper A, a model of normaland impaired hearing is employed, in which a subject's pure-tonehearing thresholds are used to adapt the model to the individual.In Paper B, the framework is modified to simulate hearing with acochlear implant (CI). Two models of hearing with CI arepresented: a simple, functional model and a biologically inspiredmodel. The models are adapted to the individual CI user bysimulating a spectral discrimination test. The framework canestimate speech recognition ability for a given hearing impairmentor cochlear implant user. This estimate could potentially be usedto optimize hearing aid settings.Paper C presents a novel method for sequentially choosing thesound level and frequency for pure-tone audiometry. A Gaussianmixture model (GMM) is used to represent the probabilitydistribution of hearing thresholds at 8 frequencies. The GMM isfitted to over 100,000 hearing thresholds from a clinicaldatabase. After each response, the GMM is updated using Bayesianinference. The sound level and frequency are chosen so as tomaximize a predefined objective function, such as the entropy ofthe probability distribution. It is found through simulation thatan average of 48 tone presentations are needed to achieve the sameaccuracy as the standard method, which requires an average of 135presentations.

Lithium-ion (Li-Ion) batteries are widely used in various applications and are viable for automotive applications. The effective management of Li-Ion batteries in battery electric vehicles (BEV) plays a crucial role in performance and range. One can achieve good performance and range by using efficient battery models in battery management systems (BMS). Hence, these battery models play an essential part in the development process of battery electric vehicles. Physics-based battery models are used for design purposes, control, or to predict battery behaviour, and these require much information about materials and reaction and mass transport properties. Model parameterization, i.e., obtaining model parameters from different experimental sets (by fitting the model to experimental data sets), can be challenging depending on model complexity and the type and quality of experimental data. Based on the idea of parameter sensitivity, certain current/voltage data sets could be chosen that theoretically has a more considerable sensitivity for a given model parameter that is of interest to extract. In this thesis work, different methods for extracting model parameters for a Nickel-Manganese-Cobalt (NMC) battery composite electrode are experimentally tested and compared. Specifically, model parameterization using \emph{optimal experiments} based on performed parameter sensitivity analysis has been benchmarked against a 1C discharge test and low rate pulse tests. The different parameter sets obtained have then been validated on a drive cycle and 2C pulse tests. Comparing the methods show some promising results for the optimal experiment design (OED) method, but consideration regarding the state of charge (SOC) dependencies, the number of parameters has to be further evaluated. / Litiumjonbatterier (Li-jon) används i olika applikationer och är ett bra alternativ förfordonsapplikationer. Den effektiva hanteringen av litiumjonbatterier i elbilar har en viktigroll för fordonens prestanda och räckvidd. Man kan nå bra prestanda och räckviddgenom att använda bra batterimodeller i batteriets övervakningssystem (BMS). Därförspelar dessa batterimodeller en viktig roll i utvecklingen av elbilar. Fysikbaseradebatterimodeller används för design, reglering eller för att prediktera beteendet hos batteriet,vilket kräver mycket information om material samt dess reaktion och andra beskaffenheter.Modellparametrisering, dvs. att införskaffa modellparametrar från olika experiment (genom attanpassa modell till experimentella data) kan vara utmanande beroende på modellkomplexitetoch typen samt kvalitén på experimentell data. Baserat på idén om parametersensitivitet kan data om ström och spänning väljas så att de teoretiskt har mer sensitivitet för engiven modellparameter som är av intresse att extrahera. I detta examensarbete testas ochjämförs olika metoder för att extrahera modellparametrar för en Nickelmangankobolt (NMC)batterielektrod. Mer specifikt, modellparametrisering genom optimala experiment baseradepå genomförd parametersesitivitetsanalys jämförts med 1C urladdningstest och låg nivåpulstest. Jämförande av metoderna visar goda resultat för OED metoden men flera parametrarmåste fortsatt utvärderas gällande laddningstatusberoenden (SOC).