Combination and Selection of Machine Learning Algorithms in GNSS Architecture Design for Concurrent Executions with HIL Testing

Aug 5, 2026

Machine learning–assisted GNSS receivers can improve autonomous navigation, but algorithm choice remains a practical tradeoff between competing goals and limited empirical guidance. This paper proposes Algorithm Selection and Matching with Fuzzy Analytic Hierarchy Process (ASM-FAHP), converting multiple quantitative and qualitative concerns into a multi-criteria decision-making problem. ASM-FAHP enumerates feasible algorithm combinations under hardware resource constraints, ranks candidates via fuzzy hierarchical attributes, and synthesizes fuzzy weights. Performance is verified through concurrent execution on resource-constrained Jetson Nano using HIL with high-fidelity synthetic datasets from Spirent GSS7000 and SimGen.

LanguageEnglish

CategoryResearch & Report

FormatPDF

Pages10

Reading Time25 min

AccessFree to download

Abstract

Index Terms

Introduction

Motivation and GNSS requirements

ML methods for GNSS processing and integrity

Examples of ML applications

Scope and verification approach

2023 IEEE/AIAA 42nd Digital Avionics Systems Conference (DASC), 1-5 October 2023, Barcelona, Spain DOI: 10. 1109/DASC58513 .2023.10311160 Combination and Selection of Machine Learning Algorithms in GNSS Architecture Design for Concurrent Executions with HIL Testing 1st Zhengjia Xu Autonomous and Cyber-Physical Sys. Cranfield University Cranfield, UK [billy.xu@cranfield.ac.uk](billy.xu@cranfield.ac.uk) 2nd Ivan Petrunin Autonomous and Cyber-Physical Sys. Cranfield University Cranfield, UK [i.petrunin@cranfield.ac.uk](i.petrunin@cranfield.ac.uk) 3rd Antonios Tsourdos Autonomous and Cyber-Physical Sys. Cranfield University Cranfield, UK [a.tsourdos@cranfield.ac.uk](a.tsourdos@cranfield.ac.uk) 4th Raphael Grech Spirent Communications Plc Spirent Devon, UK [raphael.grech@spirent.com](raphael.grech@spirent.com) 5th Pekka Peltola Navigation Telespazio UK Luton, UK [pekka.peltola@telespazio.com](pekka.peltola@telespazio.com) 6th Smita Tiwari Navigation Telespazio UK Luton, UK [smita.tiwari@telespazio.com](smita.tiwari@telespazio.com) Abstract—As machine learning (ML) continuing to gain popularity, ML-assisted Global Navigation Satellite System (GNSS) receivers facilitate the performance of Autonomous Systems (AS) navigation solutions. However, selections of ML is often a tradeoff in practice where empirical knowledge is taken to alleviate complexities. Therefore, this paper explores decision-making solutions for maximising determined hardware performance under quantitative and qualitative considerations. This work proposes Algorithm Selection and Matching with Fuzzy Analytic Hierarchy Process (ASM-FAHP) that maps multiple trade-off concerns into a Multi-Criteria Decision-Making (MCDM) problem. The ASM-FAHP firstly searches all the possible alternatives to find possible combinations with hardware resource limitations taken into account. Afterwards, ASM-FAHP prioritizes the most significant candidate by constructing a hierarchical structure with several attributes and scoring with fuzzy numbers. Hereby, the most suitable ML combinations are determined by calculating synthesised fuzzy weights per each alternative. The performance of the ML combination is evaluated by concurrently executing it on resource-constrained hardware, specifically the Jetson Nano board. The ML models are trained and tested using high-fidelity synthetic datasets produced from Spirent GSS7000 simulator and SimGen while connected to hardware-in-the-loop (HIL). It has been discovered that when approaching hardware limits, the selected combination of machine learning algorithms makes full use of memory resources but sacrifices processing speed. Index Terms—GNSS receiver design, machine learning, deep learning, algorithm selection, FAHP I. INTRODUCTION Nowadays, the performance of Global Navigation Satellite System (GNSS) receivers has more requirements in terms of accuracy, availability, continuity, integrity, as well as Size, Weight, Power, and Cost (SwaP-C) . Moreover, the evolution of processing hardware platforms brings more opportunities for This work is performed under the ESA-funded project VTL4AV (NAVISPEL1-066 bis) . implementing complex algorithms. Therefore, Machine Learning (ML) has attracted high potential interest due to relaxing theoretical assumptions and the possibility of improving GNSS performance. Facilitated by self-regression implementations in ML approaches, ML-associated designs enable efficient identification of tightly-coupled dependencies by learning action variables from datasets. By integrating ML method with processing GNSS data, a few applications are anticipated in different ways. For instance, the regressor is commonly used to model error sources induced by ionospheric and troposphere effects, multipath, clock drift, receiver noise, interference, and hardware biases [1] along with error compensations through filtering to reduce degradation effects on accuracy and availability. The self-le

FAQs

What problem does ASM-FAHP address in GNSS architecture design?

It addresses how to select and match machine learning algorithms by maximizing hardware performance while balancing quantitative and qualitative trade-offs under resource constraints.

How does ASM-FAHP rank candidate ML algorithm combinations?

ASM-FAHP searches feasible alternatives, builds a hierarchical structure with multiple attributes, scores candidates using fuzzy numbers, and calculates synthesized fuzzy weights to determine the most suitable combination.

How is the selected ML combination evaluated?

The approach is evaluated by concurrently executing the ML models on resource-constrained Jetson Nano hardware using hardware-in-the-loop testing with synthetic datasets from Spirent GSS7000 and SimGen.

Aug 5, 2026

Upload by Emma Wilson