Profile & expertise

Created in 1960 as a part of the Latvian Academy of Science, EDI (Institute of Electronics and Computer Science) is a state research institute in Riga dedicated to fundamental and applied research in five main areas:

  1. Extremely precise event timing incl. space domain
  2. Remote sensing and space data processing
  3. Robotics and machine perception
  4. Signal processing and embedded intelligence
  5. Smart sensors and IoT

By combining expertise in these areas, EDI has developed a niche specialization of Smart Embedded Cooperative Systems (SECS). The main fields impacted by the developments of EDI include space (extremely precise timing, remote sensing, image processing), mobility (components for electric and automated vehicles, sensor fusion, embedded intelligence), production (wireless sensor networks, computer vision, robotics), health (wearables, signal processing, UWB radars), digital life (human-computer interaction, smart cities, IoT).

EDI aims to bridge the gap between basic research at the universities and the application of technology in the industry, thus EDI is deliberately moving towards becoming an RTO, by tackling user- and problem-oriented research.

Currently, EDI has a workforce of more than 80 scientific stuff, including 12 PhDs, 4 PhD candidates, and 15 PhD students.

Role in the Project

In VIZTA EDI contributes to the development of a short-range 3D multi-camera system for Industry 4.0 (Task E of WP3).  For this task, EDI brings its expertise in the robotics and machine perception and its experience in the field of production automation. Our previous research in AI-based perception systems for production includes the use of computer vision with regular, stereo-vision, and depth-sensing cameras for object localization, detection, and recognition. EDI has also implemented classic and neural network-based vision algorithms on embedded systems (FPGA based System on a chip) and automated a robotic arm for picking-up different objects in changing task conditions. Using this experience, EDI participates in defining an industry 4.0 use-case for a 3D sensor S1, which is being developed in the VIZTA project. Also, EDI is responsible for developing a demonstrator application that will include the new Z-sensing technology, adaptive control of a robotic arm, signal processing, and AI-based perception models, and implementation of the solutions on the edge using ISD Herodion platform.

EDI logo

MAIN CONTACT

Roberts KADIKIS

roberts.kadikis@edi.lv

+37 (0)1 67 55 82 06

http://www.edi.lv/en/home/

Riga, Latvia

 

Ričards Cacurs joined EDI in 2013 within “National Research program”, participating in development of accelerometer sensor network for surface reconstruction. Has been involved in ESF project “Smart City Technologies for Human Lives Improvements” researching methods for 3D information reconstruction from 2D images in smart city applications. Also has been involved in EDI and ltd “EUROLCDS” collaboration project “Transformation of the DICOM standard image data for the 3D spatial image display”, where computer software was designed for lung blood vessel 3D model reconstruction from computer tomography scans. Main interests: Digital Signal Processing, Computer Vision, Image Processing, Wearable Sensor Networks, Cyber Physical Systems, Parallel Computing (nVidia CUDA platform), software development (Android, JAVA, Python, C/C++, CUDA).

Richards Cacurs

Dr. sc. ing. Roberts Kadiķis is a senior researcher, head of Robotics and Machine Perception laboratory, and a chairman of the Scientific Council at EDI, where he works since 2011. The main research areas are computer vision and machine learning, while specific interests include deep neural networks (convolutional neural networks CNN, recurrent neural networks RNN, generative adversarial networks GAN), the learning of meaningful embeddings, explainable AI, generation of synthetic training data, and computationally efficient vision algorithms. The applied experience of using deep models and other image processing tools includes the development of object detection, tracking, and image segmentation methods for tasks in transportation, biomedicine, and industry.  The efficiency of the object detection algorithms was the main emphasis of the PhD thesis that was defended in the field of electronics at Riga Technical University in 2018.

Roberts Kadikis

Rihards Novickis started his technical career in industry as an embedded Linux-based system programmer and joined EDI where he works with FPGA-based System-on-Chip technology. Rihard’s experience includes design of on-chip communications architectures, tool design for efficient ANN implementation in FPGA logic, development of asymmetric multiprocessing subsystem for ARMv7-A architecture, Linux-based system and device driver development and development of image processing accelerators. Currently, Rihards is developing his PhD thesis on stereo-vision algorithms for heterogeneous embedded systems. Rihards believes that electronics is one of the most fascinating fields to be currently involved in and considers it a privilege to work with System-on-Chip technology.

Rihards Novickis