Studentská vědecká konference

Recently, electric cars are gaining more and more popularity, due to their efficiency, environmental friendliness, and ease of maintenance. As the popularity of electric vehicles grows, so does the need to expand the electric vehicle charging infrastructure. However, sometimes the problem is that parking spaces for charging are occupied by cars that do not need it. As a result, EV owners are unable to charge and revenue from charging stations are lost. The goal of this project is to develop the system for determining the presence of a car in a parking space. The main task was to choose the most suitable sensors for this project. Several types of sensors were chosen: 9DOF, 3-axis magnetometer, wide and narrow angle ultrasonic. All sensors are connected to system controlled by single-board computer. IP-camera is used as an additional monitoring device. Now, the entire system is connected and working properly. The connection is implemented using Python3 and CircuitPython. Also, part of the system is a local network, which is used for remote control of SBC. To visualize real-time information from sensors, we use a programmable LEDs. Currently, the project is at the testing stage, the data from experiments are analyzed using visualization in Python.  

Hydrogeloví roboti jsou malé objekty s rozměrem v řádu stovek mikrometrů, které lze ovládat pomocí laserového paprsku. Vzhledem k jejich velikosti, experimenty s nimi probíhají pod mikroskopem. Tyto experimenty jsou časově náročné, proto by bylo vhodné místo experimentů in vitro jejich chování simulovat in silico. Vhodným výstupem této simulace by byl i uměle generovaný obraz z mikroskopu. Cílem této práce je ověřit možnost generovat fotorealistiké obrazy tohoto experimentu pomocí neuronových sítí. Prvním krokem je ověření, zda jsou neuronové sítě schopné generovat pouze jednotlivé roboty. Druhým krokem je hledání možností, jak vygenerovat obraz experimentu, který má rozdílné rozměry než data použitá pro učení.

The key bioprocess parameters, such as produced biopolymer and microbial biomass concentrations, are currently measured off-line by a laboratory analysis, which is unsuitable for on-line bioprocess monitoring and control. A possible solution is the application of a so-called software sensor, i.e. a computational algorithm (e.g. based on Artificial Neural Network (ANN)), which can estimate the key parameters from a set of other commonly on-line measured parameters. This work aims at analysing data obtained from a fed-batch cultivation process and reviewing various types of software sensors based on ANN that could be used for the on-line estimation of biopolymer and microbial biomass concentrations in that process. The analysed process is a fed-batch cultivation of the bacterium Pseudomonas putida KT2442 used for producing medium-chain-length polyhydroxyalkanoates (mcl-PHAs). The cultivations were carried out in a laboratory bioreactor at the Bioprocess Control Laboratory of the University of Chemistry and Technology in Prague. The obtained data involve on-line measurements of off-gas composition analysis and culture medium capacitance; and off-line mcl-PHAs biopolymer and microbial biomass concentrations, considered as inputs and outputs to the software sensors, respectively.

Meteorologická stanice VŠCHT Praha umožňuje měření vnějších a vnitřních podmínek prostředí, teploty a vlhkosti půdy a koncentrace prachových částic v atmosféře. Data z této stanice jsou odesílána do internetového cloudu. Cílem projektu bylo provést důkladný průzkum možností komunikace webové aplikace s meteorologických cloudem Weatherlink pomocí aplikačního rozhraní WeatherLink V2 API a otestovat způsoby předávání a vizualizace meteorologických dat. Výsledkem projektu je funkční kód schopný sběru, zpracování, klasifikace, vizualizace a ukládání hodnot fyzikálních veličin naměřených senzory teploty, tlaku, vlhkosti, rychlosti proudění, srážek, koncentrace prachových částic a dalších. Dále byly otestovány grafické knihovny napsané v jazyce Java Script pro grafické vizualizace naměřených dat v čase.  

Goal of this project is to analyze the possibility of using 9 Degree of Freedom Motion Processing Unit (9 DoF IMU) to detect presence of a vehicle.   The proposed solution for this problem is to use a Neural Network (NN) to process the signals from the 9 DoF IMU. The NN shall classify whether a car is absent, present, arriving or leaving. Additionally, this project focuses on experiment data visualization and processing using Python with numerous libraries (e. g. NumPy, SciPy, scikit-learn, Matplotlib, SciKit). Although data visualization is not strictly needed to create a dataset, it improves our understanding of the data. The signals were filtered to eliminate high frequency noise. Furthermore, scikit-learn's PCA was used to transform the three dimensional magnetometer signals to a single dimension.  Main application of this research is in Electric Vehicle (EV) charging stations. There is currently no solution that would not require additional installation to detect a vehicle that is not charging. These vehicles block the charging station until they are reported. Our smart charging stations could detect their occupation status even when the vehicle did not start charging with using only two additional sensors in the system.​​​​​​​

Má práce se zabývá vývojem mobilní aplikace, která za použití kamery telefonu skenuje obličej člověka a vyhodnocuje 468 obličejových bodů, na základě kterých pak dochází k vyhodnocení cviku. Cvikem je zde myšlen úsměv, zvednutí obočí, mrknutí a další rehabilitační cviky. Tato aplikace má sloužit pro rehabilitační účely pacientům, kteří mají například ochrnutou část obličeje nebo trpí jiným onemocněním nebo ochabením obličejových svalům. První prototyp mobilní aplikace byl napsán v jazyce Kotlin s použitím rozšířené reality Google AR Core. Z důvodu rozšíření aplikace i na jiné platformy se aplikace nyní programuje v jazyce C# s využitím .NET MAUI. Z důvodu právě multiplatformní podpory byl Google AR Core nahrazen MediaPipe.  

The development of drug delivery systems based on polymeric nanoparticles can address one of the main challenges in the fight against cancer, namely the improvement of drug bioavailability in the body. Therefore, the main goal is to create such a system that supports the targeting of the drug and allows it to be released in a controlled manner. At the same time, we have to think about the high toxicity of cytostatic drugs and so the possibility of protecting healthy cells. The copolymer of glycolic acid and lactic acid (poly(lactide-co-glycolide) - PLGA) belongs to the most widely used biodegradable polymers, due to their good biological compatibility and the possibility of controllable degradation by hydrolysis. Another advantage is the possible modification of the PLGA surface to the desired properties (e.g. increased solubility). The encapsulation of colchicine in a polymer system has potential because of the cytostatic effects of colchicine, but at the same time for reducing its high toxicity.