In addition to project activities, we also participate in collaboration with industry. Results are in the form of software or hardware applications. These are often related to the diploma or dissertation thesis of the team members. Below is an overview of the most significant results in last several years and implementations listed in chronological order:
Mobmon – Parameter Measurement Tool for 2G-4G Mobile Networks
In cooperation with the Department of Telecommunication Technology at CTU (Czech Technical University in Prague), we have designed a hardware device to measure the parameters of the 2G-4G mobile network. The task of our team was to design and implement a software module for Mobmon platform. The module has the following functionality: Universal Application Interface for Communication with Mobile Network Modem, Local and Remote Command Options, AT Command Interpreter, Mobile Network Data Recorder with Options Local storage, automatic and manual switching of 2G, 3G and 4G networks, measurement of connected mobile network parameters, scanning of available mobile networks, data connection / disconnection from the mobile network.
Responsible Person: Jan Rozhon
Voice control of a UAV devices
As part of the implementation of the Innovation Voucher at the MIT, and in cooperation with Skysystems Europe, s.r.o., we developed proof of concept voice-control UAV accessories – voice control of the camera carrier gimbal (Kardan’s hinge). It was necessary to design and implement a software application for mobile devices that allowed the insertion of voice commands. The Bluetooth module then transfers the signal to the modified hardware platform of the gimbal itself.
Responsible person: Filip Řezáč
Shock sensor prototype and communication gate for LoRa IoT technology
A member of our team and student of the first year of doctoral studies – Erik Grešák created a prototype of a wireless shock sensor on LoRa technology, one of the most used technology in the IoT (Internet of Things) concept. Together with the sensor, he designed and implemented both the software and the hardware part of the LoRa Gateway, that is collecting information from individual sensors and communicating with the so-called MQTT Broker.
Responsible person: Erik Grešák
Huawei Probe – Mobile network coverage measurement and 2G-4G network quality measurement services
In cooperation with Huawei, we have created a mobile application that measures the qualitative parameters of the mobile network. Mobile network coverage and service quality measurement are designed for 2G, 3G and 4G networks. The application displays the results on a map with the GPS positions of the mobile device. Tests are available to run directly from the application or remotely from the web interface. Users can export results to standard output formats for further processing such as csv, JSON, XML.
Responsible persons: Martin Mikulec, David Vašíček
BeeSIP project and individual use-cases
We have developed and maintained a BeeSIP platform over a number of years in close cooperation with CESNET. BeeSIP currently serves as a build system for OpenWRT operating system, as well as a standalone platform itself. Originally, we developed it as a multiplatform and multi-application solution focused mainly on the extension of IP telephony solutions for SOHO. We distribute the most used use-case packages in the form of deployment-ready packages.
Responsible persons: Lukáš Macura, Jan Rozhon, Filip Řezáč, Jakub Šafařík, Jiří Šlachta
Honeypot probes and automatic classification of IP phone attacks
One of the BeeSIP use-case packages (see above) is a honeypot probe that monitors malicious IP telephony traffic. The probe sends data to a Beekeeper tool, which automatically classify attack type with a neural network. The modified kernel of the tool is also used under the name LISAC as an information system with a web interface for autonomous detection and analysis of SIP attacks within the TACR project. The whole Beekeeper concept describes Jakub Safarik in his dissertation thesis.
Responsible person: Jakub Šafařík
Experimental Mobile Network in CESNET Infrastructure
The CESNET Development Fund project has provided an opportunity to learn about the possibilities of implementing its own BTS / NodeB mobile cell using open-source tools. The aim of this project was to launch a mobile cell for 2G and 4G mobiles and to measure qualitative parameters to determine if this solution could be used as a possible replacement of commercial cells. The method of solution and project results summarise the final report.
Responsible person: Martin Mikulec
Classification of emotions and speaker identification from speech signal
Within the framework of several ongoing projects, our team has created an accurate methodology for classifying emotions and identifying a speaker from a sound recording. Neural networks and proprietary algorithms are used to compare and refine results. These methods have already been deployed in several use-cases, and our team works on the further development of the tool. A comprehensive description of the current state of the solution can be found in the dissertation thesis of Pavol Partila and Jaromir Tovarek.
Responsible persons: Pavol Partila, Jaromír Továrek
Distribution of voice messages in emergency situations
The problem of the distribution of warning messages in crisis situations was covered by the dissertation thesis of Filip Řezáč when the topic was also elaborated in the project of security research of the Ministry of the Interior. Research and development deal with a communication system to alert people to dangers. This system has the task of distributing voice messages with pre-recorded content to inform the target audience of events around them using Voice over IP technology.
Responsible person: Filip Řezáč
Model for predictive call quality determination in VoIP solutions
The long-term research goal within our team is also the area of speech quality determination in telephone calls, mostly using VoIP technology. Jan Rozhon, in his dissertation thesis, devoted himself to the design and creation of a new predictive model for determining the quality of speech using neural networks. The result of the dissertation thesis is a generally applicable model implementing a non-intrusive methodology for estimating call quality that improves and refines the methods used so far and finds use in the field of IP and SIP protocol performance testing.
Responsible Person: Jan Rozhon