
Kaspars Ozols
Dr. sc. ing.
Direktora vietnieks attīstības jautājumos, vadošais pētnieks
Dr. Kaspars Ozols ir Elektronikas un datorzinātņu institūta (EDI) direktora vietnieks attīstības jautājumos, kā arī vadošais pētnieks. EDI strādā kopš 2011. gada. Kopš 2023. gada ir INSIDE industriālās asociācijas zinātniskās padomes loceklis. EDI ietvaros Dr. Ozols vada arī dažādas komisijas, piem., intelektuālā īpašuma vērtības noteikšanas; izsoles; ētikas; dzimumu līdztiesības komisijas, u.c. Absolvējis Rīgas Tehnisko universitāti (Rīga, Latvija), kur 2017. gadā ieguvis zinātņu doktora grādu elektronikas studiju programmā.
Ir bijis iesaistīts dažādos starptautiskos projektos, kas saistīti ar inovatīvu datu iegūšanas, pārraides un signālu apstrādes metožu izstrādi, kā arī unikālu ierīču izstrādi dažādiem lietojumiem transporta, veselības, ražošanas u.c. jomās, tostarp piedalījies 15+ Eiropas līmeņa pētniecības projektu īstenošanā Horizon Europe, Horizon 2020, Chips-JU, KDT-JU, ECSEL-JU, ARTEMIS-JU, ERA-NET, EIT programmās. Latvijas pārstāvis COST CA16222 "Wider Impacts and Scenario Evaluation of Autonomous and Connected Transport" (WISE-ACT) aktivitātē. Vairāk kā 7 gadus strādājis arī kā projekta pieteikumu izvērtēšanas eksperts, kā arī priekšsēdētājs (vice-chair), kas uzrauga ekspertu darbu, priekš Eiropas Komisijas un dažādu valstu zinātniskajām padomēm.
Kompetence IKT sistēmu simulēšanā, modelēšanā, projektēšanā, izstrādē, ieviešanā, testēšanā un validācijā. 50+ zinātnisko rakstu un 1 patenta autors. Pētniecībā īpaši interesējas par pašbraucošiem/autonomiem automobiļiem, kiber-fizikālām sistēmām, signālu apstrādi, bezvadu komunikāciju, bezvadu sensoru tīkliem, lietu internetu, iegultām sistēmām, zema enerģijas patēriņa risinājumiem, mākslīgo intelektu, robotiku, industrijas digitalizācijas risinājumiem u.c. Aktīvi iesaistās starptautisku projekta pieteikumu izstrādē gan kā partneris, gan kā koordinators.
Ir uzrakstījis, iesniedzis un ieguvis finansējumu vairāk kā 20 dažādiem projektiem (lielākoties iepriekš minētajās starptautiskajās programmās), pienesot EDI vairāk kā 10mEUR finansējumu zinātnei. "International workshop on Embedded Digital Intelligence (IWoEDI’2023)" konferences organizācijas komitejas priekšsēdētājs. Dr. Ozols ir arī tehnoloģiju pārneses eksperts un ir aktīvi iesaistījies un palīdzējis komercializēt 5+ zinātņu ietilpīgas tehnoloģijas. Viņš ir arī 10+ inženiera, bakalaura un maģistra darbu, kā arī 3 promocijas darbu vadītājs. Mentors 10+ jaunuzņēmumiem. 1. vieta ResearchSlam konkursā. Latvijas Zinātņu akadēmijas (LZA) balvas saņēmējs par vienu no nozīmīgākajiem sasniegumiem Latvijas zinātnē 2018. gadā un LZA prezidenta atzinības raksta saņēmējs par vienu no nozīmīgākajiem sasniegumiem Latvijas zinātnē 2025. gadā.
Projekti
3 rezultāti
- Vairāk nekā Mūrs transformējošām inovācijām jaudas elektronikā (Moore4Power) 01.05.2026. – 30.04.2029.
- Būtisko elektronisko komponentu un sistēmu (ECS) mozaika mūsu automatizētai digitālajai nākotnei rūpniecībā un mobilitātē (MOSAIC) 01.07.2025. – 30.06.2028.
- Eiropas elektronisko komponenšu un sistēmu (ECS) vērtības ķēžu suverenitātes nodrošināšana, izstrādājot nākamās paaudzes ECS autobūves nozarei (ShapeFuture) 01.05.2024. – 30.04.2027.
Publikācijas
- 2026
Differentially Private Federated Learning for Remaining Useful Life Prediction(atveras jaunā logā)
Arturs Nikulins, Kārlis Freivalds, Ivars Namatēvs, Kaspars Sudars, Audris Arzovs, Wilhelm Söderkvist Vermelin, Madhav Mishra, Kaspars Ozols.
Applied Sciences, Sēj. 16 (6), lpp. 2784 (raksta numurs), MDPI
Kopsavilkums
Accurate remaining useful life (RUL) prediction is essential for the safe and cost-effective operation of safety-critical systems such as electronic components and engines. While data-driven machine learning approaches have demonstrated strong performance for RUL estimation, their effectiveness is limited by the lack of full run-to-failure data and by strict privacy and intellectual property constraints in industrial settings. Federated learning (FL) enables collaborative model training across multiple data owners without direct data sharing, but it does not, by itself, provide formal privacy guarantees and remains vulnerable to information leakage. This paper presents a privacy-preserving DP-enhanced FL setup for RUL prediction that combines federated learning with differential privacy (DP). We describe an end-to-end implementation based on the Opacus DP library, highlight practical challenges arising from the integration of DP into recurrent neural network architectures, and propose solutions to address them. Using two representative RUL datasets (CMAPSS and SiC MOSFET), we analyze the effect of DP noise on prediction performance and on the functional dependence between the predicted RUL and the already lived life feature. The results demonstrate that differential privacy can be integrated into federated RUL prediction with limited degradation in predictive performance, providing practical insights for deploying privacy-aware collaborative models in industrial environments.
Indeksēts Scopus / WoS · Scopus procentile: 79 · SNIP: 0.956
- 2026
ConMonity: An IoT-Enabled LoRa/LTE-M Platform for Multimodal, Real-Time Monitoring of Concrete Curing in Construction Environments(atveras jaunā logā)
Namatēvs, I., Gaigals, G., & Ozols, K.
Sensors, Sēj. 26 (1), lpp. 14, MDPI
Kopsavilkums
Monitoring the curing process of concrete remains a challenging and critical aspect of modern construction, often hindered by labour-intensive, invasive, and inflexible methods. The primary aim of this study is to develop an integrated IoT-enabled platform for automated, real-time monitoring of concrete curing, using a combination of LoRa-based sensor networks and an LTE-M backhaul. The resulting ConMonity system employs embedded multi-sensor nodes—capable of measuring strain, temperature, and humidity‒connected via an energy-efficient, TDMA-based LoRa wireless protocol to an LTE-M gateway with cloud-based management and analytics. By employing a robust architecture with battery-powered embedded nodes and adaptive firmware, ConMonity enables multi-modal, multi-site assessments and demonstrates stable, autonomous operation over multi-modal, multi-site assessment and demonstrates stable, autonomous operation over multi-month field deployments. Measured data are transmitted in a compact binary MQTT format, optimising cellular bandwidth and allowing secure, remote access via a dedicated mobile application. Operation in laboratory construction environments indicates that ConMonity outperforms conventional and earlier wireless monitoring systems in scalability and automation, delivering actionable real-time data and proactive alerts. The platform establishes a foundation for intelligent, scalable, and cost-effective monitoring of concrete curing, with future work focused on extending sensor modalities and enhancing resilience under diverse site conditions.
Indeksēts Scopus / WoS · Scopus procentile: 88
- 2025
Privacy Auditing of Lithium-Ion Battery Ageing Model by Recovering Time-Series Data Using Gradient Inversion Attack in Federated Learning(atveras jaunā logā)
Kaspars Sudars, Ivars Namatēvs, Arturs Nikulins, Kaspars Ozols
Applied Sciences (Switzerland), Sēj. 15 (10), MDPI
Kopsavilkums
The exchange of gradients is a widely used method in modelling systems for machine learning (e.g., distributed training, federated learning) in privacy-sensitive domains. Unfortunately, there are still privacy risks in federated learning, as servers can reconstruct clients‘ training data through gradient inversion attacks. To protect against such attacks, we need to know when and how privacy is being undermined, largely due to the black box nature of deep neural networks. Although gradient inversion has been used to classify images and text data, its use in time-series is still largely unexplored. In this paper, we empirically evaluate the practicality of recovering lithium-ion battery time-series data from the gradients of a transformer-based model, both without and with differential privacy, in a time-series federated learning framework. It is especially significant in the case of electric vehicles (EVs). As shown in this paper, additional protection by differential privacy leads to the saturation of gradient inversion attacks, i.e., the reconstructed signal maintains a certain error level, depending on the applied privacy budget level. With this empirical evaluation, we provide insights into effective gradient perturbation directions, unfairness with respect to privacy, and privacy-preferred model initialisation. © 2025 Elsevier B.V., All rights reserved.
- 2025
Ground Segmentation Method Based on Polar Grid for 3D LiDAR Point Cloud(atveras jaunā logā)
Katrīna Tupule, Maksis Celitāns, Kaspars Ozols
2025 22nd International Conference on Networking, Sensing, and Control (ICNSC), lpp. 377-383, IEEE
Kopsavilkums
In autonomous driving platforms, one of the key tasks is the perception of the environment. Light Detection and Ranging (LiDAR) sensors are widely used to obtain a point cloud representation of the surrounding environment. Raw point clouds contain measurement points from different objects in the environment, and distinguishing and semantically segmenting them can be challenging, but beneficial for better understanding the environment. One of the challenges of preprocessing raw point clouds is their segmentation into groups based on the type of object they originated from. In this paper, we propose a real-time, CPU-based ground segmentation method that structures the point cloud into a polar grid and estimates the ground points based on the local features of the grid cells. Our ground segmentation method is evaluated on the SemanticKITTI dataset and achieves high accuracy with real-time performance. Index Terms—ground segmentation, autonomous driving, LiDAR, point clouds, perception, environment model.
Indeksēts Scopus / WoS
- 2025
Privacy Auditing of Lithium-Ion Battery Ageing Model by Recovering Time-Series Data Using Gradient Inversion Attack in Federated Learning(atveras jaunā logā)
Sudars, K., Namatevs, I., Nikulins, A., & Ozols, K.
Applied Sciences, Sēj. 15 (10), lpp. 5704, MDPI
Kopsavilkums
The exchange of gradients is a widely used method in modelling systems for machine learning (e.g., distributed training, federated learning) in privacy-sensitive domains. Unfortunately, there are still privacy risks in federated learning, as servers can reconstruct clients‘ training data through gradient inversion attacks. To protect against such attacks, we need to know when and how privacy is being undermined, largely due to the black box nature of deep neural networks. Although gradient inversion has been used to classify images and text data, its use in time-series is still largely unexplored. In this paper, we empirically evaluate the practicality of recovering lithium-ion battery time-series data from the gradients of a transformer-based model, both without and with differential privacy, in a time-series federated learning framework. It is especially significant in the case of electric vehicles (EVs). As shown in this paper, additional protection by differential privacy leads to the saturation of gradient inversion attacks, i.e., the reconstructed signal maintains a certain error level, depending on the applied privacy budget level. With this empirical evaluation, we provide insights into effective gradient perturbation directions, unfairness with respect to privacy, and privacy-preferred model initialisation.
Indeksēts Scopus / WoS · Scopus procentile: 79
Patenti
2024 LR patents Nr. 15877
Gatis GAIGALS, Kaspars OZOLS METODE SAKARU IEKĀRTU SAVIENOJUMA UZLABOŠANAI UN PATĒRĒTĀS ENERĢIJAS SAMAZINĀŠANAI VIDĒS AR MAINĪGIEM FIZIKĀLIEM PARAMETRIEM
Apraksts un saite
Skatīt patentu Espacenet datubāzē(atveras jaunā logā)LR patents Nr. 15877, Gatis GAIGALS, Kaspars OZOLS METODE SAKARU IEKĀRTU SAVIENOJUMA UZLABOŠANAI UN PATĒRĒTĀS ENERĢIJAS SAMAZINĀŠANAI VIDĒS AR MAINĪGIEM FIZIKĀLIEM PARAMETRIEM