Browsing by Author "Kuzovkin, Ilya, juhendaja"

Now showing 1 - 3 of 3

Real-Time Detection of Robot Failures by Monitoring Operator’s Brain Activity with EEG-based Brain-Computer Interface
(Tartu Ülikool, 2024) Podliesnova, Veronika; Kuzovkin, Ilya, juhendaja; Kruusamäe, Karl, juhendaja; Tartu Ülikool. Loodus- ja täppisteaduste valdkond; Tartu Ülikool. Tehnoloogiainstituut
The rapidly growing field of Brain-Computer Interfaces (BCIs) provides an innovative method for human-robot interaction, enabling machines to be controlled directly via brain signals. This study investigates the use of BCI technology to monitor the operator’s brain activity in real-time and detect robot malfunctions through the identification of Error-Related Potentials (ErrPs). In this thesis, a real-time system was developed to monitor the brain activity of robot operators continuously using an OpenBCI consumer-grade EEG device. Machine learning algorithms were implemented to analyze brain activity, specifically targeting ErrPs, which indicate that the operator is noticing something unexpected. By recognizing these brain signals, our system can identify potential robot malfunctions based on the operator’s cognitive response and trigger an immediate halt to the operation. This system integrates several key components: signal processing techniques such as resampling, filtering, and normalization to prepare the EEG data for analysis; machine learning classifiers to identify ErrPs associated with robot malfunctions; a robotic simulation that generates realistic scenarios to elicit ErrPs in participants’ brain activity for safe system testing; and real-time brain signal acquisition allowing immediate detection of ErrPs and response to faults. All these components are packed into one coherent system that detects robot malfunctions and triggers intervention. The final system underwent live testing with 10 participants, demonstrating its capability to detect ErrPs associated with robot malfunctions effectively. These tests showed that the system achieves an average sensitivity of 0.53 while maintaining a specificity of 0.98, suggesting that the system rarely reacts without a valid reason while being able to correctly detect more than half of the events of interest. For some users, sensitivity reached as high as 0.7 or more. These findings demonstrate the potential of consumer-grade EEG devices for practical applications of BCI-based robot fault detection.
Replicating DeepMind StarCraft II reinforcement learning benchmark with actor-critic methods
(2018) Ring, Roman; Kuzovkin, Ilya, juhendaja; Matiisen, Tambet, juhendaja; Tartu Ülikool. Matemaatika ja statistika instituut; Tartu Ülikool. Loodus- ja täppisteaduste valdkond
Reinforcement Learning (RL) is a subfield of Artificial Intelligence (AI) that deals with agents navigating in an environment with the goal of maximizing total reward. Games are good environments to test RL algorithms as they have simple rules and clear reward signals. Theoretical part of this thesis explores some of the popular classical and modern RL approaches, which include the use of Artificial Neural Network (ANN) as a function approximator inside AI agent. In practical part of the thesis we implement Advantage Actor-Critic RL algorithm and replicate ANN based agent described in [Vinyals et al., 2017]. We reproduce the state-of-the-art results in a modern video game StarCraft II, a game that is considered the next milestone in AI after the fall of chess and Go.
Web-based Toolbox for Interactive 3D Visualization of Neural Recordings
(Tartu Ülikool, 2021) Stomakhin, Fedor; Kuzovkin, Ilya, juhendaja; Zafra, Raul Vicente, juhendaja; Tartu Ülikool. Loodus- ja täppisteaduste valdkond; Tartu Ülikool. Arvutiteaduse instituut
The visualization of brain activity is an approach that aids neuroscience researchers and medical professionals to explore the data they work with. In particular, 3D visualization of brain activity is a technique used when the spatial positions of data points in the brain are important. Numerous tools have been developed for the analysis and editing of various forms of brain activity. In this thesis, a web-based toolbox for interactive 3D visualization of neural recordings was implemented. The use cases of the toolbox were demonstrated by adapting it to visualize intracortical LFP recordings from 100 human subjects.