Feature articles – August 2025
Our journal welcomes not only original high-quality papers covering the theoretical, experimental and operational aspects of electrical and electronics engineering in mobile radio, motor vehicles and land transportation, but also industry-focused publication focusing on research findings and suggesting ideas that may be useful to those conducting similar research.
Below, we highlight two featured peer-reviewed articles:
The first paper, coauthored by researchers and practitioners from TECNALIA, Basque Research and Technology Alliance; University of the Basque Country UPV/EHU; Virginia Tech Transportation Institute; Virtual Vehicle Research GmbH; and CEIT, proposed and verified an efficient lightweight fallback-oriented localization algorithm for automotives in urban scenarios after the failure of the main localization source.
The second paper, coauthored by researchers from Tongji University and United Arab Emirates University, presented and discussed the transformative potential of integrating large language models (LLMs) with unmanned aerial vehicles (UAVs), ushering in a new era of autonomous systems
We’ve provided short summaries of these feature articles, written in accessible language that we hope will make your reading experience enjoyable.
A Fallback Localization Algorithm for Automated Vehicles Based on Object Detection and Tracking
Authors: Mario Rodríguez-Arozamena; Jose Matute; Javier Araluce; Lukas Kuschnig; Christoph Pilz; Markus Schratter; Joshué Pérez Rastelli; Asier Zubizarreta
Published in volume 6, IEEE Open Journal of Vehicular Technology
IEEExplore version
Summary contributed by Mario Rodríguez-Arozamena (Author):
As automated vehicles become more common, ensuring their safety in every possible situation is critical, including the rare moments when their main positioning systems, like GNSS or maps, suddenly stop working. Our research introduces a novel backup localization method designed exactly for these scenarios. Instead of relying on known landmarks or external infrastructure, the vehicle uses something already present on city streets: other parked cars. Even without knowing where these parked cars are located beforehand, the automated vehicle can track their relative positions before a failure and then estimate its own position afterwards by calculating distances to them, similar to how our phone's GPS uses satellites for localization. This technique helps the vehicle maintain enough awareness of its surroundings to perform essential safety actions, such as coming to a controlled stop. We tested this method in simulations, on real-world data, and with two different car models, and the results showed clear improvements over traditional fallback methods that rely solely on internal sensors.
Large Language Models for UAVs: Current State and Pathways to the Future
Authors: Shumaila Javaid; Hamza Fahim; Bin He; Nasir Saeed
Published in volume 5, IEEE Open Journal of Vehicular Technology
IEEExplore version
Summary contributed by Nasir Saeed (Author):
Modern drones excel at tasks such as capturing images or flying preprogrammed routes, but they often struggle to adapt when conditions change or when they need to comprehend complex instructions. In everyday life, we’ve all seen drones get “stuck” because they can’t interpret a spoken command or replan a flight path on the fly. Our work pioneers a simple idea: what if drones could use large language models—“language brains” that power chatbots—to make smarter decisions? Instead of treating drones’ sensing cameras and language-based AI as separate tools, we demonstrate how to merge them into a unified system. Imagine telling a drone, “Survey that collapsed building for survivors,” and having it parse the instruction, plan an efficient search pattern, and even describe what it sees. To make this possible, we tackle real-world hurdles, such as how to run large language models on the limited computers drones carry, how to ensure the AI doesn’t “hallucinate” (i.e., make up information), and how to keep communication secure when drones share data with each other or with a base station.
Our paper walks through the key building blocks, showing how to shrink powerful large language models so they fit on a small flight computer, how to fuse text-based commands with sensor data (camera, LiDAR, GPS) in real-time, and how to verify that the drone’s “understanding” remains reliable under changing conditions (for example, poor lighting or spotty connectivity). By mapping out concrete solutions, such as onboard fine-tuning techniques, multimodal sensor-fusion strategies, and lightweight trust-check layers; we provide a roadmap for transforming today’s drones into truly conversational, adaptive machines.
In a nutshell, this work shows how to fuse advanced large language “brains” with aerial robots in a hands-on, deployable way going beyond theory to tackle real-world hurdles like limited onboard computing, reliable sensor-language fusion, and trustworthy AI output. Instead of simply reviewing past efforts, we offer a clear, step-by-step blueprint that empowers engineers and researchers to build drones that don’t just fly but they can “talk” to us, understand complex instructions, and reason through missions as naturally as having a conversation.
About IEEE Open Journal of Vehicular Technology (OJVT)
The IEEE Open Journal of Vehicular Technology covers the theoretical, experimental and operational aspects of electrical and electronics engineering in mobile radio, motor vehicles and land transportation. (a) Mobile radio shall include all terrestrial mobile services. (b) Motor vehicles shall include the components and systems and motive power for propulsion and auxiliary functions. (c) Land transportation shall include the components and systems used in both automated and non-automated facets of ground transport technology.