kth.sePublications KTH
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf
Obstacle Detection and Tracking using LiDAR for Hydrofoiling Unmanned Surface Vehicle
KTH, School of Electrical Engineering and Computer Science (EECS).
2025 (English)Independent thesis Advanced level (degree of Master (Two Years)), 20 credits / 30 HE creditsStudent thesis
Abstract [en]

The use of Unmanned Surface Vehicles (USVs) in dynamic maritime environments demands robust real-time perception systems for safe obstacle avoidance. This thesis develops a LiDAR-based obstacle detection and tracking framework for hydrofoiling USVs. The proposed pipeline fuses 3D LiDAR point clouds from an Ouster OS1-128 with position and orientation estimates from an SBG Ellipse-D GNSS-aided INS. An important component of this work is a robust point cloud preprocessing step that filters water surface detections using attitude-informed RANSAC water surface estimation and dynamic radius outlier removal. The resulting clean data supports a Multi-Object Tracking pipeline that combines Euclidean clustering, Global Nearest Neighbor association, and Kalman Filter tracking. In parallel, a Rolling Local Occupancy Grid map is maintained to represent static obstacles, with a segmentation step separating static and dynamic clusters for each of the methods. All components operate in real-time and are integrated with ROS2. The pipeline was validated both in a Unity simulation environment, developed within this thesis, and with real-world datasets collected from the Evolo hydrofoiling USV. Results demonstrate strong tracking performance in sparse conditions (MOTA: 91.6%, MOTP: 1.99 m). However, in cluttered nearshore environments, local occupancy mapping proved significantly more robust than object-based tracking. Combined with the proposed preprocessing step, this map-based approach is recommended as the perception layer for future onboard obstacle avoidance on Evolo. The evaluation also confirms LiDAR as a suitable sensor for this application, proving capable of detecting even small obstacles with sufficient range (65 m) and resolution.

Abstract [pt]

A utilização de Veículos de Superfície Não Tripulados (USVs) em ambientes marítimos dinâmicos exige sistemas de perceção robustos e em tempo real para garantir navegação segura e evitar colisões. Esta tese desenvolve um algoritmo de deteção de obstáculos baseado em LiDAR, concebido para USVs hidrofólios, que integra nuvens de pontos 3D de um Ouster OS1-128 com estimativas de posição e orientação de um SBG Ellipse-D (INS auxiliado por GNSS). Um dos componentes mais relevantes deste trabalho é o pré-processamento da nuvem de pontos, que filtra detecções da superfície da água através de RANSAC informado pela orientação do veículo e remoção dinâmica de outliers por raio. Os dados filtrados alimentam um algoritmo de seguimento de múltiplos objetos que combina clustering Euclidiano, associação por Vizinho Global Mais Próximo (GNN) e Filtros de Kalman. Em paralelo, é mantido um mapa de ocupação local para obstáculos estáticos, com segmentação entre elementos estáticos e dinâmicos. Todos os módulos correm em tempo real e estão integrados em ROS2. O algoritmo foi validado num ambiente de simulação desenvolvido em Unity e com dados reais recolhidos do USV hidrofólio Evolo. Os resultados demonstram bom desempenho em ambientes esparsos (MOTA: 91,6%, MOTP: 1,99 m), mas em zonas costeiras o mapeamento local revelou-se mais robusto do que o seguimento de objetos. Combinado com o pré-processamento, este método é recomendado como perceção base para futuros sistemas de desvio de obstáculos. A avaliação confirma ainda que o LiDAR é adequado para esta aplicação, conseguindo detetar até pequenos obstáculos com alcance (65 m) e resolução suficientes.

Abstract [sv]

Användningen av obemannade ytfarkoster (USV:er) i dynamiska marina miljöer kräver robusta perceptionssystem i realtid för säker hinderundvikning. Denna avhandling presenterar ett LiDAR-baserat ramverk för hinderdetektion och spårning, särskilt utformad för bärplansbaserade USV:er. Det föreslagna systemet kombinerar 3D-punktmoln från en Ouster OS1-128 LiDAR med positions- och orienteringsdata frånen SBG Ellipse-D INS/GNSS. En viktig del av arbetet är ett förbehandlingssteg som filtrerar bort reflektioner från vattenytan genom att använda attitydinformerad RANSAC-planpassning och dynamisk radiebaserad borttagning av avvikande punkter. De rensade data används i ett multiobjektspårningssystem som kombinerar euklidisk klustring, Global Nearest Neighbor-association och Kalman-filterspårning. Samtidigt byggs en rullande lokal ockupanskarta för att representera statiska hinder, med segmentering mellan statiska och dynamiska kluster. Alla moduler körs i realtid och är integrerade med ROS2. Systemet validerades i både en Unity-simuleringsmiljö, utvecklad inom ramen för detta arbete, samt med verkliga data insamlade från USV:n Evolo. Resultaten visar god spårningsprestanda i glesa miljöer (MOTA: 91,6 %, MOTP: 1,99 m). I tätare kustmiljöer visade sig dock lokal ockupanskartläggning vara betydligt mer robust än objektbaserad spårning. I kombination med förbehandlingen rekommenderas därför den kartbaserade metoden som perceptionslager för framtida hinderundvikning ombord på Evolo. Utvärderingen bekräftar också att LiDAR är en lämplig sensor för denna tillämpning, med förmåga att upptäcka även små hinder på upp till 65 meters avstånd med tillräcklig upplösning.

Place, publisher, year, edition, pages
2025. , p. 126
Series
TRITA-EECS-EX ; 2025:947
Keywords [en]
LiDAR, Obstacle Detection, Point Cloud, Local Mapping, Multi-Object Tracking
Keywords [pt]
LiDAR, Deteção de Obstáculos, Nuvem de Pontos, Mapeamento Local
National Category
Computer and Information Sciences
Identifiers
URN: urn:nbn:se:kth:diva-377021OAI: oai:DiVA.org:kth-377021DiVA, id: diva2:2040379
Subject / course
Systems Engineering
Educational program
Master of Science - Systems, Control and Robotics
Supervisors
Examiners
Available from: 2026-02-27 Created: 2026-02-20 Last updated: 2026-02-27Bibliographically approved

Open Access in DiVA

fulltext(78461 kB)20 downloads
File information
File name FULLTEXT01.pdfFile size 78461 kBChecksum SHA-512
adb45a7aee8a4ba91ba059df071d4db250aa867b7ff1541fbd34f191c8058ab8f66b2f8d8c6211d54f6f94d30b664c8d3201a4873e65ba32b514298386ade86a
Type fulltextMimetype application/pdf

By organisation
School of Electrical Engineering and Computer Science (EECS)
Computer and Information Sciences

Search outside of DiVA

GoogleGoogle Scholar
The number of downloads is the sum of all downloads of full texts. It may include eg previous versions that are now no longer available

urn-nbn

Altmetric score

urn-nbn
Total: 733 hits
CiteExportLink to record
Permanent link

Direct link
Cite
Citation style
  • apa
  • ieee
  • modern-language-association-8th-edition
  • vancouver
  • Other style
More styles
Language
  • de-DE
  • en-GB
  • en-US
  • fi-FI
  • nn-NO
  • nn-NB
  • sv-SE
  • Other locale
More languages
Output format
  • html
  • text
  • asciidoc
  • rtf