Citation Link: https://nbn-resolving.org/urn:nbn:de:hbz:467-12
Innovative Sensorkonzepte und Signalverarbeitungsstrategien zur Bewegungserkennung und Präsenzkontrolle von Personen
Source Type
Doctoral Thesis
Author
Subjects
PIR-Bewegungsmelder
CMOS-Bildwächter
Kalman-Filter
DDC
620 Ingenieurwissenschaften und Maschinenbau
GHBS-Clases
Issue Date
2002
Abstract
Motion observers and presence controllers are essential parts of modern concepts for building automation and energy management. Especially the low-cost passive infrared sensors, based on the pyroelectric effect, are suitable for reducing the overall power consumption by providing information about the presence or absence of persons.
Particularly in indoor applications, conventional motion observers suffer from the low resolution of commercially available pyroelectric detectors. On the other hand, CCD cameras for the visible and near infrared region of light provide extremely high resultions, but for comparatively high cost. Newer developments in CMOS imager technology enable the design of low cost camera sensors. In the scope of this work, preleminary investigations on both, the
CMOS camera and the motion observer with pyroelectric detectors, are presented. Both parts can be realized as separate sensor systems or in combination.
Many years of experience in optical sensor design at the Center of Sensor Systems and closer investigations on signal processing strategies for motion detection culminated in a new superior motion observer. Digital signal processing algorithms reduce the susceptibility to false alarms through a very special evaluation of signal-specific characteristics.
For future improvements, a new evaluation strategy for motion observers was developed, which is based on the foundations of estimation theory. This approach uses a discrete-time linear state space model of the real world process and leads to a linear, optimum estimation algorithm, which is called “Kalman-filter” according to its explorer. The Kalman-filter serves as suppressor for unwanted background signals. Deviations between the predicted behaviour and the behaviour observed at the real system are used for motion detection. This procedure
yields into an increased sensitivity with simultaneously improved false alarm immunity.
Particularly in indoor applications, conventional motion observers suffer from the low resolution of commercially available pyroelectric detectors. On the other hand, CCD cameras for the visible and near infrared region of light provide extremely high resultions, but for comparatively high cost. Newer developments in CMOS imager technology enable the design of low cost camera sensors. In the scope of this work, preleminary investigations on both, the
CMOS camera and the motion observer with pyroelectric detectors, are presented. Both parts can be realized as separate sensor systems or in combination.
Many years of experience in optical sensor design at the Center of Sensor Systems and closer investigations on signal processing strategies for motion detection culminated in a new superior motion observer. Digital signal processing algorithms reduce the susceptibility to false alarms through a very special evaluation of signal-specific characteristics.
For future improvements, a new evaluation strategy for motion observers was developed, which is based on the foundations of estimation theory. This approach uses a discrete-time linear state space model of the real world process and leads to a linear, optimum estimation algorithm, which is called “Kalman-filter” according to its explorer. The Kalman-filter serves as suppressor for unwanted background signals. Deviations between the predicted behaviour and the behaviour observed at the real system are used for motion detection. This procedure
yields into an increased sensitivity with simultaneously improved false alarm immunity.
File(s)![Thumbnail Image]()
Loading...
Name
krieger.pdf
Size
7.49 MB
Format
Adobe PDF
Checksum
(MD5):3af37d6be3577e36ab606a2ff8cd56fa
Owning collection