Feature articles – July 2026
Our first monthly feature paper, co-authored by researchers from University of Bremen, RWTH Aachen University, RPTU Kaiserslautern-Landau, and DFKI German Research Center for Artificial Intelligence presents the conceptualization, challenges, and prospects of unified 3D network architectures, encompassing space, air, and ground segments, for 6G. The second feature article, coauthored by industry practitioners Zoox Inc, surveys how foundational models address critical challenges in autonomous perception, including limitations in generalization, scalability, and robustness to distributional shifts.
We’ve provided short summaries of these feature articles, written in accessible language that we hope will make your reading experience enjoyable.
Unified 3D Networks: Architecture, Challenges, Recent Results, and Future Opportunities
Mohamed Rihan; Dirk Wübben; Abhipshito Bhattacharya; Marina Petrova; Xiaopeng Yuan; Anke Schmeink; Amina Fellan; Shreya Tayade; Mervat Zarour; Daniel Lindenschmitt; Hans Schotten; Armin Dekorsy
Published in Volume 6, IEEE Open Journal of Vehicular Technology
Read on IEEE Xplore
Summary contributed by Mohamed Rihan (author):
Our paper introduces the concept of unified 3D networks for 6G, where space, air, and ground segments are integrated into one coordinated communication architecture. Instead of treating satellite, aerial, and terrestrial networks as separate systems, we envision them working together to provide seamless, resilient, and ubiquitous connectivity. The paper explains why this integration is important, identifies the key technical challenges, and highlights promising solutions for mobility management, handover, interference mitigation, and efficient resource allocation. In particular, we discuss enabling techniques such as federated learning, advanced beamforming, and energy-efficient offloading that can help turn this vision into a practical network architecture for future applications.
Foundation Models for Autonomous Driving Perception: A Survey Through Core Capabilities
Rajendramayavan Sathyam; Yueqi Li
Published in Volume 6, IEEE Open Journal of Vehicular Technology
Read on IEEE Explore
Summary contributed by Rajendramayavan Sathyam (author):
Autonomous driving systems need to understand a complex world that is constantly changing, from unusual road layouts to difficult weather and incomplete sensor data. This survey focuses on how foundation models for autonomous driving perception can be built to support more general, adaptable, and robust perception systems than traditional task-specific approaches. A key contribution of the paper is a capability-based framework organized around four essential abilities for real-world driving: generalized knowledge, spatial understanding, multi-sensor robustness, and temporal understanding. Generalized knowledge helps vehicles handle diverse objects and rare situations more effectively. Spatial understanding builds a richer 3D view of the environment for safer navigation. Multi-sensor robustness improves reliability when sensor inputs are noisy or degraded. Temporal understanding helps the system track motion, handle occlusion, and understand how scenes evolve over time. Rather than organizing recent work only by method type, the paper uses this capability-driven lens to connect existing approaches with concrete model-development goals, offering practical guidance on how future perception systems can be designed to better satisfy the key requirements of real-world autonomous driving.