SH
S.A. Hijlkema
info
Please Note
<p>This page displays the records of the person named above and is not linked to a unique person identifier. This record may need to be merged to a profile.</p>
2 records found
1
Mobile devices are getting increasingly powerful, becoming compatible
for an ever increasing set of functionality. Applications based around
neural networks however still have to offload parts of their computations
to the cloud since current Artificial Neural Networks (ANNs) are
still too computationally expensive for any practical standalone use in
energy constrained mobile devices. Developments in the next generation
of ANN: Spiking Neural Network (SNN), are expected to bring
neural networks directly to the edge. Even though SNNs are becoming
a reality, they can not (yet) effectively operate on raw sensory input
data. For this, a preprocessing algorithm can be used to extract low-level
features in an efficient way to boost the neural network efficiency.
A parallel can be found in biology with the cochlea that, for audio,
provides preprocessing for the brain. Recent research has shown that
an SNN is capable of reaching high classification accuracy when combined
with an biologically plausible audio preprocessing stage. To be
of interest for edge-computing it however also needs to be area and
energy efficient. This thesis will provide the first steps in researching
the optimal configuration of a specific audio preprocessing algorithm
by mapping its current software simulation to embedded hardware.
For this purpose the software simulation is analyzed and an efficient
hardware implementation is designed. For evaluation a prototype,
and its hardware constrained simulation, is developed and optimized. ...
for an ever increasing set of functionality. Applications based around
neural networks however still have to offload parts of their computations
to the cloud since current Artificial Neural Networks (ANNs) are
still too computationally expensive for any practical standalone use in
energy constrained mobile devices. Developments in the next generation
of ANN: Spiking Neural Network (SNN), are expected to bring
neural networks directly to the edge. Even though SNNs are becoming
a reality, they can not (yet) effectively operate on raw sensory input
data. For this, a preprocessing algorithm can be used to extract low-level
features in an efficient way to boost the neural network efficiency.
A parallel can be found in biology with the cochlea that, for audio,
provides preprocessing for the brain. Recent research has shown that
an SNN is capable of reaching high classification accuracy when combined
with an biologically plausible audio preprocessing stage. To be
of interest for edge-computing it however also needs to be area and
energy efficient. This thesis will provide the first steps in researching
the optimal configuration of a specific audio preprocessing algorithm
by mapping its current software simulation to embedded hardware.
For this purpose the software simulation is analyzed and an efficient
hardware implementation is designed. For evaluation a prototype,
and its hardware constrained simulation, is developed and optimized. ...
Mobile devices are getting increasingly powerful, becoming compatible
for an ever increasing set of functionality. Applications based around
neural networks however still have to offload parts of their computations
to the cloud since current Artificial Neural Networks (ANNs) are
still too computationally expensive for any practical standalone use in
energy constrained mobile devices. Developments in the next generation
of ANN: Spiking Neural Network (SNN), are expected to bring
neural networks directly to the edge. Even though SNNs are becoming
a reality, they can not (yet) effectively operate on raw sensory input
data. For this, a preprocessing algorithm can be used to extract low-level
features in an efficient way to boost the neural network efficiency.
A parallel can be found in biology with the cochlea that, for audio,
provides preprocessing for the brain. Recent research has shown that
an SNN is capable of reaching high classification accuracy when combined
with an biologically plausible audio preprocessing stage. To be
of interest for edge-computing it however also needs to be area and
energy efficient. This thesis will provide the first steps in researching
the optimal configuration of a specific audio preprocessing algorithm
by mapping its current software simulation to embedded hardware.
For this purpose the software simulation is analyzed and an efficient
hardware implementation is designed. For evaluation a prototype,
and its hardware constrained simulation, is developed and optimized.
for an ever increasing set of functionality. Applications based around
neural networks however still have to offload parts of their computations
to the cloud since current Artificial Neural Networks (ANNs) are
still too computationally expensive for any practical standalone use in
energy constrained mobile devices. Developments in the next generation
of ANN: Spiking Neural Network (SNN), are expected to bring
neural networks directly to the edge. Even though SNNs are becoming
a reality, they can not (yet) effectively operate on raw sensory input
data. For this, a preprocessing algorithm can be used to extract low-level
features in an efficient way to boost the neural network efficiency.
A parallel can be found in biology with the cochlea that, for audio,
provides preprocessing for the brain. Recent research has shown that
an SNN is capable of reaching high classification accuracy when combined
with an biologically plausible audio preprocessing stage. To be
of interest for edge-computing it however also needs to be area and
energy efficient. This thesis will provide the first steps in researching
the optimal configuration of a specific audio preprocessing algorithm
by mapping its current software simulation to embedded hardware.
For this purpose the software simulation is analyzed and an efficient
hardware implementation is designed. For evaluation a prototype,
and its hardware constrained simulation, is developed and optimized.
Motional Feedback in a Bass Loudspeaker
Digital Implementation
This thesis describes the digital implementation of a motional feedback system for a bass loudspeaker. Motional feedback is used to suppress the linear and non-linear distortions produced by the loudspeaker, especially at the low frequencies. An accelerometer is mounted on the cone of the loudspeaker to provide the feedback signal. The controller which consists of a PI controller and an equalizer are implemented on an FPGA. The equalizer, which is the inverse of the linear model of the loudspeaker, is used to compensate for the linear distortion. The PI controller with negative feedback is used to suppress the non-linear distortion. Not all measurement results are available at the moment of submission of this thesis. However, simulations were carried out on the model of the loudspeakers which show that the linear distortion is fully suppressed. The reduction of the non-linear distortion due to the controller can not be seen in the simulations.
...
This thesis describes the digital implementation of a motional feedback system for a bass loudspeaker. Motional feedback is used to suppress the linear and non-linear distortions produced by the loudspeaker, especially at the low frequencies. An accelerometer is mounted on the cone of the loudspeaker to provide the feedback signal. The controller which consists of a PI controller and an equalizer are implemented on an FPGA. The equalizer, which is the inverse of the linear model of the loudspeaker, is used to compensate for the linear distortion. The PI controller with negative feedback is used to suppress the non-linear distortion. Not all measurement results are available at the moment of submission of this thesis. However, simulations were carried out on the model of the loudspeakers which show that the linear distortion is fully suppressed. The reduction of the non-linear distortion due to the controller can not be seen in the simulations.