Clemens Gühmann

Low-complexity compression of short messages

We describe a low-complexity scheme for lossless compression of short text messages. The method uses arithmetic coding and a specific statistical context model for prediction of single symbols. Our particular contribution is a simple yet effective approach for storing highly complex statistics in a succinct yet effective data model that can easily be trained by text data. The proposed model already gives good compression rates with a RAM memory size of 128 kByte, thus making lossless data compression with statistical context modeling readily applicable to small devices like wireless sensors or mobile phones

Fractional Wavelet Filter for Camera Sensor Node with external Flash and extremely little RAM

This paper introduces the fractional wavelet filter as a technique to compute fractional values of each wavelet subband, thus allowing a low-cost camera sensor node with less than 2 kByte RAM to perform a multi-level 9/7 picture wavelet transform. The picture dimension can be 256x256 using fixed-point arithmetic and 128x128 using floating-point arithmetic. The technique is applied to a typical 16 Bit sensor node architecture with external flash memory, which allows to line-wisely read and write picture data.

Wavelet Image Two-Line Coder for Wireless Sensor Node with Extremely Little RAM

This paper gives a novel wavelet image two-line (Wi2l) coder that is designed to fulfill the memory constraints of a typical wireless sensor node. The algorithm operates line-wisely on picture data stored on the sensors flash memory card while it requires approximatively 1.5 kByte RAM to compress a monochrome picture with the size of 256x256 Bytes. The achieved data compression rates are the same as with the set partitioning in hierarchical trees (Spiht) algorithm. The coder works recursively on two lines of a wavelet subband while intermediate data of these lines is stored to backward encode the wavelet trees. Thus it does not need any list but three small buffers with a fixed dimension. The compression performance is evaluated by a PC-implementation in C, while time measurements are conducted on a typical wireless sensor node using a modified version of the PC-code.

Compression of Short Text on Embedded Systems

The paper details a scheme for lossless compression of short data series larger than 50 Bytes. The method uses arithmetic coding and context modeling with a low-complexity data model. A data model that takes 32 kBytes of RAM already cuts the data size in half. The compression scheme just takes a few pages of source code, is scalable in memory size, and may be useful in sensor or cellular networks to spare bandwidth. As we demonstrate the method allows for battery savings when applied to mobile phones.

External Flash Filesystem for Sensor Nodes with sparse Resources

This paper describes a free filesystem for external flash memory to be employed with low-complexity sensor nodes. The system uses a standard secure digital (SD) card that can be easily connected to the serial port interface (SPI) or any general input/output port of the sensor’s processor. The filesystem is evaluated with SD- cards used in SPI mode and achieves an average random write throughput of about 40 kByte/sec. For random write access throughputs larger than 400 kByte/sec are achieved. The filesystem allows for storage of large amounts of sensor or program data and can assist more memory expensive algorithms. It requires 7 kByte of program memory and about 570 Bytes of RAM.

Sliding mode position control for automotive dry clutch system: Nonlinear observer/sliding mode control design for an electro-hydraulic actuation system

Automatic position control of automotive dry clutches is a crucial factor for comfort and efficiency aspects. In this paper the nonlinear approach of sliding-mode control (SMC) with underlying nonlinear Luenberger-observer for trajectory tracking is designed analytically and tested in simulation.

Reliability based lifetime prediction and optimization for permanent magnet synchronous machines

This paper presents a methodology for on-line evaluation of electrical motors current state, used to predict its life-time and to increase the residual lifetime. Based on the estimated motor state, the residual lifetime can be extended drastically by reducing the load. Also maintenance costs can be reduced.

Fault Diagnosis on Bearings of Electric Motors by Estimating the Current Spectrum

Abstract The current spectrum of a motor with a faulty bearing contains the rotation frequency and its higher harmonics as well as narrow sidebands produced by frequency and amplitude modulations. A physical model is developed to predict the current spectrum of an universal motor with a faulty bearing. Numerical simulations are compared with measu-rements and the results are used to find significant features for the classification. To classify faulty and faultless motors an algorithm for the extraction of these significant features is introduced. For the feature extraction a model based estimation approach (Prony’s method) is presented.

Fault Diagnosis of Electric Low-Power Motors by Analyzing the Current Signal

Abstract To detect and localize faults occurring in the rotor of universal motors analysis of the current signal in the time and frequency domain is presented. The information about the condition of the rotor is contained mainly in the periodical components of the current signal. Because these harmonics are speed dependent and the speed is not constant, digitizing synchronously with speed is carried out. The method of linear prediction is used as a spectral modeling technique in which the current spectra are modeled by all-pole spectra. The coefficients of the linear predictors are the basis for automatic fault detection and classification.

Automated parameter identification for a dry clutch

Detailed models of automated passenger car transmissions are used for developing new shift/ control algorithms [1], [2]. The modeling process is based on good system knowledge and appropriate parametrization. Without a suitable parametrization the model will not fit the real system behavior. This paper describes a two-step method to identify the mechanical parameters of a dry clutch in order to parametrize a simulation model. This is done at a transmission test bench by using Structured Recurrent Neural Networks and an automated identification process. In the first step the inertias of the clutch are identified. Furthermore the sliding friction of the test bench drive train is also determined. In the second step, the torque capacity of the dry clutch is ascertained.