The Institute for Electronics and Computer Science (EDI) invites students from RTU, LU, and other universities (including from abroad) to develop their master’s and bachelor’s thesis using the institute’s resources and scientific guidance.
We can match you with a supervisor: simply write to info@edi.lv or call 67554500.
The development of master’s and bachelor’s works are managed by experienced scientific staff, and facilities are provided for the carrying out of theoretical studies and experimental developments. Writing your master’s or bachelor’s thesis can be combined with work in projects at the institute!
Detailed information on selecting a thesis topic, internships, and work opportunities at EDI.
Topics
Extremely accurate measurement of event timing
- Super-high precision (picosecond) time and time-amplitude measurement systems, as well as time synchronization systems
Remote research and processing of space data
- Treatment of hyperspectral images with applications for classification of objects, such as forest classification
- Using the information theory to classify hyperspectral image objects
Robotics and Machine Perception
- Computer vision and deep learning (artificial neural networks) for detecting, localizing, and segmenting objects in images and video (1D, 2D, 3D). The theory here will have to face image processing, machine learning, convolution neural networks, recurrent neural networks. In practice, it will include defining and training neural networks, preparing and generating training data sets
- Clustering of unlabeled data with artificial neural networks or Deep Learning
- Extending training data with 3D computing and simulation software and generative neural networks
- Methods of interpretative artificial intelligence systems, explanations for decision-making of artificial intelligence systems (“black box”), software and methods for explaining artificial intelligence (functional, local, structural, perturbation, modular), interpretation and explanation
- Data compression with artificial neural networks / Deep Learning
- Realtime RGB/IR data fusion
- Embedded implementation of drone indoor localization algorithms
- Automatic staining of medical images with deep learning tools – GAN, Stable Diffusion, etc.
- Classification of medical images using deep neural networks
- Enlargement of medical image datasets with image augmentation and generation methods – from Blender to Stable Diffusion
- Federated learning-based algorithms for efficient system and component model building
- Implementation of neural networks on multiple energy-constrained devices (edge devices)
- Particle systems for physical simulation of gases and liquids
- Machine learning with differentiable physics
- Deep machine learning for program code generation
- Combinatorial optimization with graph neural networks
Robotics
- Computer vision and industrial robot applications in the management of various types of industry, etc., and the development of a human tracking system using an industrial robot and a camera
- Applications of stimulated learning in the control of industrial and mobile robots
- Bio-inspired navigation in mobile robots
- Manipulation of soft objects by robots
- Use of physical simulation of gases and liquids in robotics
Signal processing and embedded intelligence
- Ultra-broadband (UWB) signal generation, recording, processing methods, and sub-technical solutions for applications of the “Through-wall imaging”, terrestrial mapping, contact-free diagnostics, small-object motion detection, aerial selection for ultrawide-band signal discharge using the impactor method, etc. Handling of signals from various sources (radar, medical, sports, social, etc.) using the Karhunen-Loeve method
- Ultra-wideband radar tomograph: a study of the internal structure of a 100-year old oak
- Generating pulses ranging from parts of the volt to a thousand volts and fronts from hundred picoseconds to more than one nanosecond
- The human body as a data-transmission environment: experimental measurements on the spread of an electrical field and magnetic field through the human body
- Linux routine operation acceleration using FPGA
- Application-specific integrated processor design for image processing related tasks
- Applications of zero-knowledge protocols
Smart sensors and the Internet of Things
- Obtaining and using bioimpedance measurements to obtain information about heart rate, breathing, and other parameters of people
- Development of wireless sensing networks and IoT to improve quality of life (including for people with special needs), smart house, smart class, smart laboratory, environmental improvements, and/or industrial equipment performance (Industry 4.0)
- Identification of human activities using wearable devices equipped with an accelerometer and other sensors and/or infrastructure sensors such as stationary WiFi or IEEE 802.15.4 devices
- Protocols (BLE, IEEE 802.15.4, LoRa etc.) for low-power wireless networks with a large number of nodes and dense deployment
- Design and development of self-driving cars: development of the MPC (Model Predictive Control) controller, implementation of a programming language for describing self-driving car maneuvers, effective creation of a neighborhood 3D point-cloud map from LiDAR data, GPS and video camera data
- Detection of drone orientation in space using an optical sensor
- Wearable sensors in healthcare. Investigate the use of wearable sensors for monitoring vital parameters (heart rate, blood pressure, oxygen saturation) and compare them with traditional medical devices
- Wearable sensors in sports. Examine the use of wearable sensors for assessing athletic performance (speed, distance, workload, fatigue). Investigate how these sensors can be used to prevent injuries or optimize training plans
- Wearable sensors in human-computer interfaces. Investigate how wearable sensors can be used to detect user gestures, facial expressions, and other non-verbal control commands
Cyber security
- Analysis and practical feasibility of costs of exploiting SS7 vulnerabilities
- Social engineering attacks against e-Signature users and recommendations to mitigate them
- The resilience of home/office alarms to physical attacks.
The Institute for Electronics and Computer Science is a national scientific body that handles research in the fields of electronics, signal processing, telecommunications, and computer science, and conducts the development of science-intensive electronic equipment and software using modern technologies and solutions. In line with the cooperation agreements concluded with the RTU, LU, and VeA, the Institute is a research base for these universities and has equivalent rights and responsibilities in the management of academic works. The activities of the Institute are based on the execution of projects financed by various sources of Latvia, the European Commission, the European Structural Funds, etc., in which original solutions are found for the tasks assigned.