PhD Theses
No. 7 - Jochen Withopf
Jochen Withopf: Signalverarbeitungsverfahren zur Verbesserung der Sprachkommunikation im Fahrzeug
Shaker-Verlag, 2017
Commission
- Prof. Dr.-Ing. Gerhard Schmidt
(first reviewer) - Prof. Dr.-Ing. Rainer Martin
(second reviewer) - Prof. Dr. rer. nat. Steffen Börm
(examiner) - Prof. Dr.-Ing. habil.Franz Faupel
(head of the examination board)
Abstract
Speech communication inside a moving vehicle is often difficult because of the presence of high background noise levels and because the conversational partners do not face each other. In-car communication (ICC) systems help the passengers in such situations by recording the speech with microphones placed close to the talker’s mouth and reproducing it amplified with loudspeakers located close to the listener’s ears. However, by this approach, an improvement in speech intelligibility and speech quality can only be achieved if system stability, despite of the operation in a closed electro-acoustic loop, can be remained at the required system gain. Furthermore, the overall system delay has to be low enough to prevent from the perception of two individual sound sources.
Starting from the boundary conditions of speech communication inside a vehicle, this work develops a generic algorithmic framework which interconnects the signal processing methods for enhancing the microphone signals and distributing them to the available loudspeaker channels. The strict requirement for low signal delay is fulfilled by a special filter bank design which also allows for a reduction in computational complexity. In a basic version, the system stability margin is increased by equalization and signal-dependent feedback suppression. A reduction of non-stationary background noise is obtained by a multi-channel pre-processing scheme for the microphone signals. Based on this, a method for feedback cancelation is derived. Due to the high correlation between the talker signal as the desired signal and the loudspeaker signals as the excitation of the adaptive filters, a suitable method for signal decorrelation is investigated and implemented. A final comparison between different methods for feedback control clearly shows the superior performance of the cancelation approach, but also illustrates the increased requirements in system resources.
All algorithms described in this work are implemented within the real-time signal processing framework KiRAT and tested in an audio laboratory as well as under real driving conditions. Even complex algorithms, such as feedback cancelation, are always considered in the context of the entire ICC-system in order to ensure the development of practical solutions.
No. 6 - Vasudev Kandade Rajan
Vasudev Kandade Rajan: Speech Enhancement in Hands-free Systems for Automobile Environments
Shaker-Verlag, 2017
Commission
- Prof. Dr.-Ing. Gerhard Schmidt
(first reviewer) - Prof. Dr.-Ing. Tim Fingscheidt
(second reviewer) - Prof. Dr.-Ing. Dipl.-Wirt. Ing. Stephan Pachnicke
(examiner) - Prof. Dr. habil. Klaus Rätzke
(head of the examination board)
Abstract
A new microphone position in the automobile where microphones are placed on the seat belt is available. The pros outweigh the cons of the position which makes the it very attractive to be be practically used. In order to be able to use this microphone a set of issues are addressed through signal processing methods. Some of the methods presented are improved versions of the existing ones and some methods are new ideas. They are adapted and assessed in the automobile context. The central work of the thesis is a set of speech enhancement algorithms which are applied to the belt microphones. The speech enhancement chapters presented in the thesis form the basic units of a hands-free system.
Belt microphones when integrated into hands-free systems are used to pick up the speech of the passenger/s in the automobile. The microphone signals also contain the echoes of the remote talker which is played back over the loudspeaker in the automobile. This thesis presents an acoustic echo canceller to remove these echoes. The echo canceller must be able to not only remove the echoes to the extent that a transparent conversation is possible, but also satisfy measures specified by various standards. In order to achieve this an improved way to control the adaptive filters of the echo canceller is presented. The control involves estimation of unmeasurable quantities. Based on theoretically derived optimal quantities an improved step-size control is presented as compared to the existing ones. By utilizing properties of the impulse response such as the inherent delay, slow varying nature, it is shown that the proposed control method out performs the existing method which is based on the same principle. The chapter also presents a way to deal with the moving of the microphones when the body of the passenger moves. The problem of “room change” is handled through the two coupling factors which are in built control mechanisms of the step-size control. The step-size control and the room change is evaluated using standard measures under different realistic automobile scenarios.
The speech enhancement of the microphone also deals with the estimation of the background noise in the automobile environment. The noise estimation chapter of the thesis proposes a new noise estimation scheme applicable to the belt microphones. The scheme considers noise scenarios involving nonstationary signals like the sudden change in the noise properties when the window of the automobile is lowered. By tracking long term average, the long term level, and taking into account the short term dynamics of the background noise, a multiplicative constant based scheme is proposed. The basic idea involves the classification of the current state of the noise as either speech, slowly changing noise, or fast changing noise. Based on this classification an estimate is made combined in the end with the instantaneous background noise. By doing so it has been shown that it is possible to track the background noise aggressively at the same time avoid tracking speech. This scheme has been compared with two other well known schemes. The evaluation shows that the proposed scheme is the better choice in the evaluated scenarios.
The traditional Wiener filter approach to noise suppression has been re-looked in the final speech enhancement chapter of the thesis. The existing modifications of the Wiener filter are presented as a basis for proposing newer modifications. The first proposed modification moves from the retention of the background noise as a suppressed version to a shaping of the suppressed noise. Two ways to reshape the background noise is proposed, first involving the low frequency noise present in the belt microphones due to its proximity to the windows, second involving the equalization of the noise in the remaining frequencies. The idea is to permanently apply the low frequency modification and apply the broadband equalization to only non-speech frequencies. The second proposed modification tries to reduce the speech distortion caused when the background noise needs to overestimated to avoid the so called “musical noise”. The modifications are subjectively tested and the improvements of the methods are shown through spectrogram plots. A hands-free system where the above proposed speech enhancement algorithms has been implemented in a real-time system. This system has been tested in two cars. The software and hardware implementation details are described in the real-time implementation chapter of the thesis. The evaluation of the hands-free system into which the speech enhancement units are integrated is shown.
No. 5 - Kolja Pikora
Kolja Pikora: Automatische Mehrzielverfolgung als Grundlage für die Kontaktfusion und Parameterschätzung in einem Aktivsonarsystem
Shaker-Verlag, 2017
Commission
- Prof. Dr.-Ing. Gerhard Schmidt
(first reviewer) - PD. Dr. Wolfgang Koch
(second reviewer) - Prof. Dr.-Ing. Michael Höft
(examiner) - Prof. Dr. rer. nat. Lorenz Kienle
(head of the examination board)
Abstract
An active sonar system is used for the detection, classification and tracking of underwater objects. It consists of a transmitter which sends out acoustic energy into the water and a receiver (towing ship) which makes use of a towed hydrophone array (towed array) for detecting acoustical reflections in the environment. Within a common used signal processing chain the reflections are processed by signal enhancement and beamforming algorithms to create sonar contacts each representing one reflection. Based on these generated contacts, tracks of possible target trajectories are estimated in a subsequent automatic multitarget tracking which is within this work realized by a cardinalized PHD-filter. In parallel to the contact generation parameter estimation techniques could be used for approximating system parameters whose knowledge is essential for a correct contact generation. If several waveforms are used for transmission each of these waveforms is processed in an individual signal processing chain. The resulting contacts can be fused in an explicit contact fusion procedure or within the multitarget tracking to achieve an increasing localization accuracy of the targets.
A central problem in target tracking is demonstrated by the fact, that the uncertainty in the generation of sonar contacts affects the position of target tracks proportionally. The accuracy and probability of detection of sonar contacts can be mitigated by a loss of coherence in the transmitted signal waveforms and by a maneuver of the towing ship (and thus of the towed array). The purpose of this work is the study of a feedback of tracking information back to the contact generation for an improvement of the target localization. The contact fusion and the parameter estimation within the contact generation are possibilities to use the information provided by the feedback.
Within the context of this work, a so called semicoherent contact fusion is developed which counteracts the loss of coherence in the signal forms. This fusion technique contains an association of contacts generated from hyberbolic frequency modulated waveforms by using target state information from the target tracking as additional input. It is shown, that the semicoherent contact fusion leads to increasing tracking performance while suffering from multitarget situations. Within a parameter estimation for approximating the hydrophone positions the semicoherent contact fusion is used for calculating multitarget likelihoods. Within this work, the estimation of hydrophone positions is realized in a selftracking routine which also uses target state information from the target tracking. The focus of the estimation of hydrophone positions is on the moment of a towing ship maneuver. It turns out, that a self-tracking during a maneuver is possible and neither knowledge of non acoustical sensor data of the hydrophone array nor prior knowledge about sources of opportunity in the environment are necessary. The discussion of the new developed algorithms is based on Monte-Carlo simulations generating track metrics which describe the quality of the tracks at the output of the multitarget tracker. The results of the semicoherent contact fusion are compared with the iterative update procedure within the cardinalized PHD-filter. The results of the self-tracking procedure is compared with tracking results based on different knowledge about the maneuver of the array. As shown in the results, the feedback of information from the multitarget tracker back to the contact fusion and parameter estimation is a capable approach for increasing the performance of an active sonar system.
No. 4 - Anne Theiß
Anne Theiß: Instrumentelle Evaluierung von Innenraum-Kommunikationssystemen
Pdf-based submission (available freely via the MACAU system), 2017
Commission
- Prof. Dr.-Ing. Gerhard Schmidt
(first reviewer) - Prof. Dr.-Ing. Sebastian Möller
(second reviewer) - Prof. Dr.-Ing. habil. Ludger Klinkenbusch
(examiner) - Prof. Dr.-Ing. Jeffrey McCord
(head of the examination board)
Abstract
The communication between the occupants of a speeding vehicle can be impaired. In particular, the conversation between the passengers of the front row and those of the rear seats is due to the occuring background noise and the arrangement of the passengers difficult. Usually the passengers change their position and their voice in order to compensate for this impairment. Another method to improve the communication situation is a so-called in-car communication system (ICC system). Such systems take the speech signal of a talking person using the already installed microphones, process and play back the amplified desired signal inside the passenger compartment through the loudspeakers nearby the listening passengers. Thus, the power of the desired signal inside the compartment increases as well as the intelligibility.
Following the successful implementation and development of such systems, the question arises to the quality of the ICC system. To answer this question, in this work, first, the term ”quality” referring to an ICC system is defined. Subsequently, an evaluation strategy, which consists of a variety of individual instrumental methods, is presented. For a plausible classification of the methods and a transparent investigation of the overall quality, these methods are divided into three different evaluation groups. These evaluation groups examine the characteristics of the vehicle, the behavior of the ICC system itself, as well as the quality of the communication. Within each of these groups first instrumental methods are described and validated. Subsequently, the individual results of the instrumental methods are weighted by previously calculated attributes and combined into an overall result. The attributes of the vehicle and the ICC system are determined based on the determined characteristics of the vehicle. For the development of the instrumental methods a quality reference which reflects the subjective perception of passengers is required. For this purpose, some auditory evaluation methods are introduced in addition to the instrumental methods and the results analyzed. Therefore, not only listening tests, but also psychoacoustic perception experiments have been carried out and analyzed. The evaluation strategy implemented into a real-time framework is finally tested on a specific ICC system and an overall view of the quality of all its results is proposed.
No. 3 - Manuel Haide
Manuel Haide: Durchflussmessung mit Ultraschall-Phased-Array-Sensoren
External publication 2016
Commission
- Prof. Dr.-Ing. Gerhard Schmidt
(first reviewer) - Prof. Dr.-Ing. Wolfgang Schroer
(second reviewer) - Prof. Dr.-Ing. habil. Thomas Meurer
(examiner) - Prof. Dr.-Ing. Werner Rosenkranz
(head of the examination board)
Abstract
Future development of reliable flow measurement in urban water management, flood control, process and power engineering are driven by developing wear-free and flexible measurement systems with high accuracy. Current ultrasonic measurement systems are limited in their scope and thus signal processing and sensor design need to be improved. The basis of the discussed measuring principle are sound-reflecting suspended solids in the fluid, which move slip-free along the stream.
This thesis presents fused signal processing methods to determine velocities of suspended solids in fluids. In combination with a phased-array ultrasonic sensor, a successive measurement of the velocity prole across the entire channel or pipe cross section can be performed. Moreover, signals are transmitted with a frequency hopping coding in order to detect a position-dependent information of the transit time and Doppler frequency. With the objective of noise-resistant and accurate flow measurement, object tracking theories or cluster methods are used to fuse the received echo information.
The object tracking is realized by a Kalman lter and various local and global association methods such as the Neares-Neighbour- and (Joint) Probabilistic Data Associations method. In contrast to this time-based tracking of streamlines, the cluster method determines the Doppler frequency by an application-oriented MUSIC method. Areas which are beyond of the phased-array sensor coverage, are complemented by a computational optimized function to generate a velocity distribution across the entire channel / pipe cross section. The technical feasibility of the analyzed evaluation methods as well as the sensor system has been adduced by a hardware platform.
By using the data fusion methods measurement errors have been reduced for high in flow velocities. Ambiguities in the Doppler frequency interpretation are cancelled by the data fusion, and thus enlarge the measurement range. In combination with the ultrasonic phased-array technology inhomogeneous velocity proles are determined across the entire channel / pipe cross section.
More Articles ...
Page 3 of 4