ACCESS-AV: A daptive C ommunication- C omputation Cod e sign for S u s tainable A utonomous V ehicle Localization in Smart Factories CCS: Computer systems organization Embedded and cyber-physical systems CCS: Computer systems organization Real-time systems CCS: Hardware Digital signal processing CCS

Rajat Bhattacharjya, Arnab Sarkar, Ish Kool, Sabur Baidya, Nikil Dutt · 2025

Read ACCESS-AV: A daptive C ommunication- C omputation Cod e sign for S u s tainable A utonomous V ehicle Localization in Smart Factories CCS: Computer systems organization Embedded and cyber-physical systems CCS: Computer systems organization Real-time systems CCS: Hardware Digital signal processing CCS by Rajat Bhattacharjya, Arnab Sarkar, Ish Kool, Sabur Baidya, Nikil Dutt (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2507-20399.txt.

Introduction

Rajat Bhattacharjya and co-authors present, in a 2025 paper, a low-energy way for delivery vehicles to find their place inside smart factories. These driverless carts haul goods through halls linked by 5G networks. Pinpointing a cart's spot takes heavy computing, which drains power. Their system, ACCESS-AV, reuses the 5G signals already broadcast in the hall, so no extra roadside units or sensors are needed.

The carts read the angle of incoming signals to fix their position. A smart strategy shifts the balance between radio use and computing as signal strength and speed change. In tests, the system cut energy use by about forty-three percent versus fixed versions of the same method. It still placed carts within a hand's span of their true spot, and it lowers both setup and running costs for factory fleets.

Catalog record

CREATORRajat Bhattacharjya, Arnab Sarkar, Ish Kool, Sabur Baidya, Nikil Dutt
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE54 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2507-20399

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

ACCESS-AV: A daptive C ommunication- C omputation Cod e sign for S u s tainable A utonomous V ehicle Localization in Smart Factories CCS: Computer systems organization Embedded and cyber-physical systems CCS: Computer systems organization Real-time systems CCS: Hardware Digital signal processing CCS: Networks Wireless access networks CCS: Networks Cyber-physical networks CCS: Applied computing Rajat Bhattacharjya, Arnab Sarkar, Ish Kool, Sabur Baidya, Nikil Dutt Abstract . Autonomous Delivery Vehicles (ADVs) are increasingly used for transporting goods in 5G network-enabled smart factories, with the compute-intensive localization module presenting a significant opportunity for optimization. We propose ACCESS-AV , an energy-efficient Vehicle-to-Infrastructure (V2I) localization framework that leverages existing 5G infrastructure in smart factory environments. By opportunistically accessing the periodically broadcast 5G Synchronization Signal Blocks (SSBs) for localization, ACCESS-AV obviates the need for dedicated Roadside Units (RSUs) or additional onboard sensors to achieve energy efficiency as well as cost reduction. We implement an Angle-of-Arrival (AoA)-based estimation method using the Multiple Signal Classification (MUSIC) algorithm, optimized for resource-constrained ADV platforms through an adaptive communication-computation strategy that dynamically balances energy consumption with localization accuracy based on environmental conditions such as Signal-to-Noise Ratio (SNR) and vehicle velocity. Experimental results demonstrate that ACCESS-AV achieves an average energy reduction of 43.09% compared to non-adaptive systems employing AoA algorithms such as vanilla MUSIC, ESPRIT, and Root-MUSIC. It maintains sub-30 cm localization accuracy while also delivering substantial reductions in infrastructure and operational costs, establishing its viability for sustainable smart factory environments. † † footnotetext: Authors’ version posted for personal use and not for redistribution. 1. Introduction Figure 1. A Smart Factory Infrastructure Powered by Private 5G: Autonomous Delivery Vehicles (ADVs) transport goods across plants while leveraging 5G Synchronization Signal Blocks (SSBs) from roadside Remote Radio Heads (RRHs) for localization. With rapid advances in computing systems and robust network infrastructure, factories are increasingly transitioning toward autonomy. Modern smart factories, often equipped with private 5G infrastructure ( pvt5g ) , span vast areas covering several acres of land ( hplc ) . Safety regulations dictate significant separation between plants, with distances ranging from a few hundred meters to a half kilometer ( separation ) . These extensive layouts pose logistical challenges for the efficient and safe transportation of goods across facilities. To address these challenges, autonomous delivery vehicles (ADVs) equipped with Level 4 (L4) autonomy have emerged as a promising solution to transport goods between plants ( uisee ) , as shown in Figure 1 . These vehicles reduce human dependency and improve operational efficiency and safety ( factory ) . However, the deployment of ADVs requires overcoming critical challenges related to both onboard energy consumption (extending battery life for long-duration operations ( adv ) ) and overall system sustainability (reducing deployment costs and the environmental impact of additional infrastructure). Toward these ends, we explore energy-efficient localization strategies that leverage existing 5G networking infrastructure. Our goal is to simultaneously extend ADV operational durations and improve smart factory sustainability by eliminating the need for dedicated transmission infrastructure— also known as Roadside Units (RSUs) ( rsu ) — thereby significantly reducing both embodied and operational carbon footprints ( jones ; cc ) . We focus on localization, an integral perception module in an autonomous vehicle’s computational stack ( chauffeur ; avstack ) , which presents an opportunity to address the intertwined energy and sustainability bottlenecks in ADVs. Specifically, we focus on Vehicle-to-Infrastructure (V2I) localization using Angle of Arrival (AoA)-based estimation via the Multiple Signal Classification (MUSIC) algorithm ( music_lite ) . MUSIC provides high-resolution AoA estimates and has proven effective in diverse applications, from biomedical sensors ( use2 ) and seismic monitoring ( seismic ) , to joint radar-communication systems ( musicuse ) and challenging multipath-rich V2I environments ( italy ) . Studies show that AoA-based localization can offer greater accuracy compared to GPS-based positioning ( italy ; italy2 ) and is well-suited for robust ADV localization ( music_lite ) . Our approach offers a holistic and sustainable solution by primarily focusing on (1) eliminating the need for dedicated Roadside Units (RSUs) ( rsu ; aviral ) . We achieve this by repurposing the factory’s existing private 5G infrastructure for V2I communication, a foundational shift that drastically reduces deployment costs, complexity, and carbon footprint ( jones ) . Within this sustainable architecture, we employ an adaptive communication-computation codesign approach to dynamically balance localization accuracy and onboard energy consumption. This adaptivity is crucial, leveraging real-time factors like environmental quality (e.g., SNR) and vehicle velocity to ensure localization is both trustworthy and computationally efficient. Furthermore, our strategy enhances sustainability on the vehicle itself by (2) minimizing onboard hardware complexity and cost through reliance solely on a single wireless receiver, obviating the need for additional sensors (like cameras or LiDAR) for this task. Figure 1 is a cartoon of a smart factory deploying a private 5G network that uses Remote Radio Heads (RRHs) ( rrh ) strategically positioned along roads interconnecting adjacent factory plants. These RRHs, part of the existing communication infrastructure, effectively serve as proxy RSUs in our system. Our framework, ACCESS-AV , opportunistically exploits the 5G Synchronization Signal Blocks (SSBs) ( ssb ; alexandra ) periodically broadcast by these RRHs to enable efficient V2I localization. ACCESS-AV leverages its adaptive communication-computation approach to dynamically calculate the AoA on the ADV’s onboard system, achieving precise localization while minimizing energy consumption. Operating within the factory’s predictable, closed-loop environment, ACCESS-AV ensures energy-efficient navigation along predefined ADV paths. By specifically addressing the computational demands and energy-efficiency challenges of the MUSIC algorithm ( music_lite ) through this adaptive codesign, ACCESS-AV dynamically balances localization accuracy against onboard energy expenditure in real-time. By intelligently leveraging existing network resources (5G RRHs) and non-dedicated signaling (SSBs), the framework ensures accurate, scalable, cost-effective, and energy-efficient localization, supporting the broader vision of sustainable smart factories. To the best of our knowledge, ACCESS-AV is the first work to jointly address ADV onboard energy constraints and infrastructure sustainability considerations through a cohesive, adaptive localization approach leveraging opportunistic V2I communication. ACCESS-AV ’s main contributions are that we: • Eliminate the need for dedicated RSUs and additional signaling by opportunistically leveraging periodically broadcast SSBs from existing 5G-enabled RRHs in the factory infrastructure. This approach reduces deployment costs, system complexity, and overall carbon footprint ( jones ) . • Simplify onboard hardware requirements by utilizing only an Orthogonal Frequency-Division Multiplexing (OFDM) wireless receiver for localization, eliminating the need for additional sensors and their associated maintenance costs while promoting sustainable development. • Present a communication-computation codesign framework that dynamically balances localization accuracy and onboard energy consumption via an algorithm that adapts to real-time conditions. • Demonstrate average energy reduction of 43.09% compared to non-adaptive systems employing vanilla MUSIC, ESPRIT ( esprit ) , and Root-MUSIC ( root ) , while maintaining high localization accuracy with worst-case error below 0.3 meters. • Further quantify the sustainability benefits of our approach via empirical cost estimation and analysis, which demonstrates over a 130\times reduction in onboard sensor costs. The rest of this paper is structured as follows: Section 2 provides relevant background; Section 2 details the system model; Section 4 presents the adaptive localization framework; Sections 5 , 6 , and 7 discuss the communication-computation codesign framework, experimental setup, and evaluation, respectively; Section 8 reviews related work;

Read the complete text