Umeå University's logo

In this paper, we present a haptic-based wireless electronic system that helps users with profound visual impairment avoid obstacles and map their surroundings through an unknown landscape. Our aim has been to design a lightweight and low-cost system that can easily and discretely be worn by the user. The system comprises two parts: 1) the “sensor module”which is laser-based and can be worn on the belt, and 2) two “haptic modules” which can be worn on each wrist with its vibrators directly placed on the skin. These two parts communicate wirelessly which makes it comfortable for the user to wear the system. The sensor module creates a representation matrix of the surrounding objects, and as an object gets closer, a higher pulse-width-modulation (PWM) duty cycle (i.e., a higher vibratorypower) is delivered by the corresponding vibrator(s), indicating its direction and proximity. This obstacle avoidance system was tested on seven individuals between 50 and 78 years of age (mean=63.5, SD=9.3 years) with profound vision impairment. The participants were instructed to walk through two path configurations: one with and another without the proposed assistive system. According to our results, significantly (p<0.05) shorter time(Median = 64s (IQR 32-81) versus Median = 95s (IQR 51-134) was needed and slightly fewer collisions occurred once they used this assistive system. The participants stated that they found the haptic signals intuitive and easy to understand. Hardware schematics and codes are publicly available for future development.

This paper proposes a computationally light adaptivefiltering approach, normalized least mean squares (NLMS), to model the nonlinearity caused by the current transformer (CT) iron core saturation. A simplified CT model was used to generate adataset considering four different nonlinear iron core magnetic characteristics. The preliminary results show satisfactory results in the cases where the CT iron core nonlinearity is within certain limits.

Power-quality standards provide limited guidance on frequency quality for short time scales, such as less than one hour. Capturing frequency variations and events requires high time resolutions, e.g., 0.1 seconds or less, resulting in significant data storage requirements. However, power-quality monitors typically report averaged values at intervals of 10 minutes, 15 minutes, or one hour, depending on the disturbance type. To address these challenges, we propose calculating statistic indices from high-resolution data within a 15-minute or one-hour window length, thus avoiding storing high-resolution data. We apply the basic statistic indices to frequency data measured in Finland for a single day in June 2023, demonstrating their effectiveness in capturing frequency variations and two significant events during that day.

It is believed that the phase of the cochlear wave propagation might play a crucial role in binaural perceptionand sound localization by generating cochlear disparities. Experimental data demonstrate that, while a tone is being played,an excitation pattern is formed extending from the base to the apex of the cochlea. The phase of the excitation patterndecays along the cochlear length until the slope of the phase curve (i.e. group delay) reaches its maximum at a locationwith the characteristic frequency (CF) that matches the frequency of the input tone. Thereafter, the phase stays almostconstant (group delay equals zero) until the apex. Computational models have been devised to simulate the cochlearresponses and thereby illuminate the underlying electromechanics of the human inner ear. These computational models canbe divided, according to their topology, into two groups: Parallel filterbanks that model the cochlea as several independentdecoupled filters versus cascade filterbanks (including transmission lines) which assume that the filters are coupled inseries. Due to their modeling principles, cascade filterbanks intrinsically include the longitudinal traveling wavepropagation whereas the parallel filterbank models lack this intrinsic feature since there is no longitudinal relation betweenthe filter stages in these models. The objective of this study is to verify if cascade filterbanks are actually more successfulin simulating the phase responses than parallel filterbanks. The excitation patterns generated by seven cochlear models(four parallel filterbanks, two cascade filterbanks and a transmission-line model) in response to 4 and 9 kHz tones wereestimated using an impulse, and the results were compared with corresponding experimental data recorded at comparablelengths in living mammalian cochleae. The accuracy of the model predictions was reported in mean absolute error (MAE)relative to the experimental data. The results show that cascade filterbank models are remarkably more successful withCARFAC and VERHULST models reproducing the experimental data most closely. However, DRNL model (a parallelfilterbank) also produces outcomes that are comparable in accuracy with those generated by cascade filterbank models.Further investigations showed that this specific parallel filterbank model externally incorporated correct phase delays inthe impulse responses of its filter stages. The results indicate that, if parallel filterbanks incorporate the phase delays intheir impulse responses according to the proposed method, they could successfully simulate the timing of the longitudinalwave propagation along the cochlea within the same accuracy range as cascade filterbank models do.

Acoustic echo cancelation (AEC) is a system identification problem that has been addressed by various techniques and most commonly by normalized least mean square (NLMS) adaptive algorithms. However, performing a successful AEC in large commercial vehicles has proved complicated due to the size and challenging variations in the acoustic characteristics of their cabins. Here, we present a wideband fully linear time domain NLMS algorithm for AEC that is enhanced by a statistical double-talk detector (DTD) and a voice activity detector (VAD). The proposed solution was tested in four main Volvo truck models, with various cabin geometries, using standard Swedish hearing-in-noise (HINT) sentences in the presence and absence of engine noise. The results show that the proposed solution achieves a high echo return loss enhancement (ERLE) of at least 25 dB with a fast convergence time, fulfilling ITU G.168 requirements. The presented solution was particularly developed to provide a practical compromise between accuracy and computational cost to allow its real-time implementation on commercial digital signal processors (DSPs). A real-time implementation of the solution was coded in C on an ARM Cortex M-7 DSP. The algorithmic latency was measured at less than 26 ms for processing each 50-ms buffer indicating the computational feasibility of the proposed solution for real-time implementation on common DSPs and embedded systems with limited computational and memory resources. MATLAB source codes and related audio files are made available online for reference and further development.

An off-axis digital holographic interferometry technique integrated with a Mach–Zehnder interferometer based setup is demonstrated for measuring the temperature and temperature profile of a transparent medium. This technique offers several advantages: it does not require precise optomechanical adjustments or accurate definition of the frequency carrier mask, making it simple and cost-effective. Additionally, high-quality optics are not necessary. The methodology relies on measuring the phase difference between two digitally reconstructed complex wave fields and utilizing the temperature coefficient of the refractive index. In this way, we presented an equation of the temperature as a function of phase changes and the temperature coefficient of refractive index. This approach simplifies the calculation process and avoids the burden of complicated mathematical inversions, such as the inverse Abel transformation. It also eliminates the need for additional work with the Lorentz–Lorentz equation and Gladstone–Dale relation and can be extend for 3D measurements.

In an open world with a long-tail distribution of input samples, Deep Neural Networks (DNNs) may make unpredictable mistakes for Out-of-Distribution (OOD) inputs at test time, despite high levels of accuracy obtained during model training. OOD detection can be an effective runtime assurance mechanism for safe deployment of machine learning algorithms in safety–critical applications such as medical imaging and autonomous driving. A large number of OOD detection algorithms have been proposed in recent years, with a wide range of performance metrics in terms of accuracy and execution time. For real-time safety–critical applications, e.g., autonomous driving, timing performance is of great importance in addition to accuracy. We perform a comprehensive and systematic benchmark study of multiple OOD detection algorithms in terms of both accuracy and execution time on different hardware platforms, including a powerful workstation and a resource-constrained embedded device, equipped with both CPU and GPU. We also profile and analyze the internal details of each algorithm to identify the performance bottlenecks and potential for GPU acceleration. This paper aims to provide a useful reference for the practical deployment of OOD detection algorithms for real-time safety–critical applications.

Since the introduction of hands-free telephony applications and speech dialog systems in automotive industry in 1990s, microphones have been mounted in car cabins to capture, and route the driver's speech signals to the corresponding telecommunication networks. A car cabin is a noisy and reverberant environment where engine activity, structural vibrations, road bumps, and cross-talk interferences can add substantial amounts of acoustic noise to the captured speech signal. To enhance the speech signal, a variety of real-time signal enhancement methods such as acoustic echo cancellation, noise reduction, de-reverberation, and beamforming are typically applied. Moreover, the recent introduction of AI-driven online voice assistants in automotive industry has resulted in new requirements on speech signal enhancement methods to facilitate accurate speech recognition. In this chapter, we focus on spatial filtering techniques that are designed to spatially enhance signals that arrive from certain directions while attenuating signals that originate from other locations. The fundamentals of conventional beamforming and echo cancelation are explained and are accompanied by some real-world examples. Moreover, more recent techniques (namely blind source segregation, and neural-network based adaptive beamforming) are presented in the context of automotive applications. This chapter provides the readers with both fundamental and hands-on insights into the fast-growing field of automotive speech signal processing.

A biophysically inspired signal processing model of the human cochlea is deployed to simulate the effects of specific noise-induced inner hair cell (IHC) and outer hair cell (OHC) lesions on hearing thresholds, cochlear compression, and the spectral and temporal features of the auditory nerve (AN) coding. The model predictions were evaluated by comparison with corresponding data from animal studies as well as human clinical observations. The hearing thresholds were simulated for specific OHC and IHC damages and the cochlear nonlinearity was assessed at 0.5 and 4 kHz. The tuning curves were estimated at 1 kHz and the contributions of the OHC and IHC pathologies to the tuning curve were distinguished by the model. Furthermore, the phase locking of AN spikes were simulated in quiet and in presence of noise. The model predicts that the phase locking drastically deteriorates in noise indicating the disturbing effect of background noise on the temporal coding in case of hearing impairment. Moreover, the paper presents an example wherein the model is inversely configured for diagnostic purposes using a machine learning optimization technique (Nelder–Mead method). Accordingly, the model finds a specific pattern of OHC lesions that gives the audiometric hearing loss measured in a group of noise-induced hearing impaired humans.