Here we present a number of examples which were successfully tackled using WhizniumSBE and WhizniumDBE.
Multi-Spectral Detector
To showcase WhizniumSBE’s and WhizniumDBE’s capabilities, at MPSI we’ve got a modular vision demonstrator which we can re-configure from low-cost (e.g. ZedBoard and USB webcams) to high end (e.g. Minnowboard Turbot, Xilinx Kintex and industrial GigE cameras), and everything in between.
Derived and adapted from the ICARUS detector (see below), our hardware demonstrator features:
- two cameras
- a thermal imager
- a modulated laser pointer mounted on a tilt/pan unit
- high-power LED’s and accelerometer
It allows to implement functionalities such as stereo imaging, sensor fusion and visual feature tracking using the laser pointer.
MultiSpectralDetectorControl WhizniumSBE project (msdc):
MultiSpectralDetectorDevice WhizniumDBE project (msdd):
FabSight
A typical IIoT application is to continuously monitor a machine’s operation parameters on the factory floor in order to determine its state and health. Higher-level applications such as predictive maintenance furthermore require the aggregation of data for many machines over long periods of time in the cloud.
Our demo application FabSight shows how this can be done using the Whiznium tools. Highlights of the project include:
- an FPGA board configured as oscilloscope with peak detection and FFT, processing current and voltage (I/V) on the supply line of a mobile beer cooling box
- a data model on the edge computer which allows to store raw data and insights derived from them
- simple machine learning for the interpretation of I/V traces and spectra
- multiple connectivity options for the edge-based tool to receive additional known process parameters via software and to display live data through a web-based HMI
- the cloud-based counterpart to the edge device database which can be used for periodic synchronization via its API
View screencast tour on YouTube now (hint: use HD!)
Download .pdf See FabSight.Analytics in action See FabSight.BeerCooler in action
FabSight.Device WhizniumDBE project (sfcd):
Download .zip WhizniumDBE in the cloud
The ICARUS Detector
WhizniumSBE and WhizniumDBE gained their street credibility as powerful developer tools for embedded systems with the ICARUS detector as a challenging use case.
Within a European Commission funded robotics project for Search and Rescue by the name of ICARUS, the task at hand was to design a complex detector system featuring:
- two cameras and stereo vision
- a thermal imager
- auxiliary functionalities of actuation (tilt/pan), high-power LED illumination and a modulated laser pointer
- custom hardware (FPGA among others) along with gumstix/ARM embedded computer
ICARUS WhizniumSBE project (idec):
ICARUS WhizniumDBE project (idhw):
BeamRelay
The roots of WhizniumSBE are in distributed simulation projects running in the cloud (C++ as common basis with today’s embedded focus). BeamRelay was the first “data science” project which was developed using WhizniumSBE throughout the entire project lifecycle.
In a private effort and out of curiosity, the task was to determine the feasibility of using commercial airplanes as relay stations for free-space optical communication e.g. for high-speed on-board internet access. The effort resulted in a conference presentation at ICNS 2014 and a scientific publication. Features of the underlying software tool include:
- reading and analyzing airline timetable information into a database
- calculating mutual visibilities airports/airplanes and airplanes/airplanes using a parallelized geometrical algorithm
- establishing possible communication relay paths
- accumulating statistics
- graphical result output