Nicholas Sadjoli

Author Statistics

Papers

Citation

H-Index

i-10 index

Research Field

Pcd Reconstruction, Object Classification and Pose Estimation for Underwater Vehicles Using Orthogonal Multibeam Forward Looking Sonar Fusion

Nicholas Sadjoli Basman Elhadidi Yiyu Cai Gerald Gim Seet Gim Lee

Orthogonal Multibeam Sonar Fusion (OMSF) is a recent fusion based method capable of producing accurate underwater 3D Point Cloud Data (PCD) from Multibeam Forward Looking Sonars (MFLS), enabling accurate seabed mapping and scanning. This article provides methodical testing of OMSF reconstruction for MFLS accounting for operating frequency effects and object shapes. The article then proposes novel perception-based applications for OMSF, consisting of classification technique integrating OMSF PCDs with a PCD based Convolutional Neural Network (CNN), and pose estimation method combining Orthogonal Feature Matching (OFM) bounding box regression with a pose regression CNN. Reconstruction test results show that OMSF is more accurate and efficient using higher frequency MFLS and achieves 43\% higher accuracy on solid surfaces compared to hollow frames. Application tests based on an underwater garage dock show classification using OMSF PCDs can achieve 25% and 37% higher success rate and confidence while being more efficient, compared to using raw 3D sonar data. OFM based regression produces 4.28% higher mean Intersection over Union (IoU), and 10\% increase in >25% IoU metric compared to methods based on more traditional MFLS filtering. Similarly, pose estimation achieves 6.25% higher success rate with OFM bounding boxes comparedÂ to ones obtained using traditional MFLS filtering.

10.2139/ssrn.4521497

Cite

Citations (1)

Acoustic PCD-Based AUV Perception for Enhanced Automated Garage Docking

OCEANS 2021: San Diego – Porto (2021)

Nicholas Sadjoli Chia Hao Gen Dinh Quang Huy Yiyu Cai Gerald Seet Gim Lee

Automated docking for AUVs is an important application for prolonged AUV usage. However, current AUV perception systems experience several limitations that are accentuated in more turbid waters. A potential factor that causes this issue is the limited spatial information of the objects extractable from the 2D images utilized by both acoustic and optical-based modalities commonly found on such perception systems today.Inspired by the current progress done for underwater Point Cloud Data (PCDs), this paper thus proposes an acoustic PCD-based system that can synthesize PCD data with minimal acoustic image inputs, and utilize the additional spatial data from PCDs to potentially enhance the AUV perception for complex applications, such as automated docking. The proposed system consists of two main components: acoustic-based PCD reconstruction module, and a PCD-based classifier/pose-estimator (CPE) Convolutional Neural Network (CNN) module. Several simulation and in-field based experiments have been conducted to validate the feasibility of the system's modules. Current results discussed in this paper show a potential feasibility for the proposed system for use in complex applications such as automated garage docking, noting further works to be conducted to improve the design and viability of the proposed system.

Docking (animal)

10.23919/oceans44145.2021.9705817

Cite

Citations (1)

Feasibility of dictionary-based sparse coding for data compression in machine condition-based monitoring

Wang Ling Goh Nicholas Sadjoli

On-time machine maintenance and upkeep are critical to ensure on-time production. In recent years, Condition-Based Monitoring (CBM) has been regarded as one of the most state-of- the-art machine maintenance techniques that can significantly lower unscheduled maintenance cost and provide greater efficiency. This however means that more data needs to be streamed and stored for CBM to work properly, which might prove costly in the long run in terms of data storage. This paper therefore focuses on the feasibility of using Sparse Coding as a method for signal compression for CBM purposes. Specifically, the feasibility of the method will be measured for vibration signals coming from the spindle component of machines.

Condition Monitoring

Maintenance engineering

Neural coding

10.1109/icis.2018.8671663

Cite

Citations (1)

A ROS Approach on Multi-mode Control Systems for Remote Operating Vehicle

OCEANS 2021: San Diego – Porto (2021)

Hao Gen Chia Nicholas Sadjoli Dinh Quang Huy Yiyu Cai Gearld Seet

The paper presents a multi-mode control system implemented to a ROV. The multi-mode system enables pilots to train on a virtual simulator or to deployed for a physical mission. This provides a cost effective and risk-reducing option arising from using two different or dissimilar systems. In particular, the paper highlights the design methodology and architecture of the system to be built on ROS. The structure of ROS enable data to be transferred easily between models through nodes and topics. This allows flexibility in adding, or removing, features. Furthermore, ROS is compatible on serval platforms which extends applicability and permits codes to be reused on different platforms. By implementing on ROS, it saves on resources and programming effort, compared to developing on a specified package. Design and coding time were found to be faster than implementing a solution solely based on a custom algorithm with native codes

Mode (computer interface)

10.23919/oceans44145.2021.9706126

Cite

Citations (5)

Effectiveness of 3D PCD Reconstruction using Orthogonal Multiple Sonar Fusion on Frame and Surface type Objects

OCEANS 2022, Hampton Roads (2022)

Nicholas Sadjoli Yiyu Cai Gerald Seet Gim Lee Basman Elhadidi

Recent development of orthogonally-oriented multi-beam sonars fusion (OMSF) has enabled effective reconstruction of underwater 3D point cloud data (PCD) using commonly available acoustic imaging sonars. However, its effectiveness on various object shapes has not been carefully investigated. Addressing such gap is important as underwater structures have varying and complex configurations.This paper investigates the effectiveness of OMSF on three common types of shapes encountered underwater: flat 'hollow' frames, flat 'solid' surfaces, and curved 'solid' surfaces. In-depth analysis of OMSF indicates that its interpolation process tends to fit estimated PCD points onto large surface shapes, leading its PCD results to be more accurate on objects with solid surfaces and not as effective on hollow objects with frames. This analysis is supported by simulation and in-field experiments, with in-field results showing OMSF obtaining up to 60% higher reconstruction accuracy on flat solid surfaces compared to frames with similar outer dimensions.Overall, the analysis and results in this paper shows a limitation of OMSF method for frame-like structures.

Sensor Fusion

Surface reconstruction

10.1109/oceans47191.2022.9976953

Cite

Citations (1)

Object perception in underwater environments: a survey on sensors and sensing methodologies

Ocean Engineering (2022)

Dinh Quang Huy Nicholas Sadjoli Abu Bakr Azam Basman Elhadidi Yiyu Cai

Modalities

10.1016/j.oceaneng.2022.113202

Cite

Citations (55)