DR-NTU (Digital Repository of NTU)
Not a member yet
116018 research outputs found
Sort by
Localization-assisted fast and robust beam optimization for mmWave communications
No full textThe millimeter wave (mmWave) communication becomes a key enabler for the future Internet of Things (IoT) due to its capability for supporting high rate and low-latency traffic. However, beamforming in the mmWave band faces issues of low efficiency since the narrow beam of mmWave devices would increase the search delay and overhead. Inspired by this, we utilize localization over sub-6 GHz band to assist the mmWave base station in performing fast and robust adaptive beamforming (RABF). Different from existing works, we focus on the indoor scenario and consider the effects of several practical issues, including localization errors and hardware defects. Specifically, a novel two-step access scheme is proposed. During the first step, we design a novel localization method customized for indoor scenarios, jointly considering the time of flight and angle of arrival. The localization error is further analyzed to determine the mmWave scanning angle and an optimal beamwidth expression is derived in closed-form to maximize system throughput with the considerations of the search delay. Moreover, considering the mismatch of the steering vector caused by the hardware defects, we propose an RABF method in closed-form. Simulation results demonstrate that the proposed scheme can effectively reduce the search delay and realize robust beamforming to enhance the mmWave communication performance.Submitted/Accepted versionThis work was supported in part by the Key Research and Development Program of Hangzhou City under Grant 2023SZD0073, and in part by the Key Research and Development Program of Zhejiang Province under Grant 2024C01140
Design of a rail-to-rail output analog buffer
No full textThe design of a low-voltage rail-to-rail op amp used for high-speed signal buffering or generating ac ground is presented in this paper. The op amp consists of a bulk-driven input transconductance stage with Gm-compensator circuit, a transimpedance gain stage with class-AB bias control, and a rail-to-rail output stage with Miller compensation. As many submicron devices from the GlobalFoundries 55nm BCD-Lite process were used, the circuit was systematically designed and optimized using the increasingly popular Gm/Id methodology. Pre-layout simulation results in the typical corner show that the op amp has an operating supply voltage range from 1.2V to 1.8V. For small-signal performance, it achieves an open-loop gain of 60.9dB, unity gain bandwidth of 66.7MHz and a phase margin of 58.90 while driving a resistive load of 40Ω and capacitive load of 60pF. Besides, the op amp achieves a good THD of less than -55.6dB at 1kHz, PSRR+ and PSRR- of 54.4dB and 56.0dB respectively, positive and negative slew rates of 27.6V/us and 85.1V/us respectively, input-referred noise of 855nV/√Hz at 1kHz and 39nV/√Hz at 1MHz, and CMRR of 113dB with a standard deviation of 14dB. With a quiescent power of 8.09mW, the FOM of the op amp is calculated to be 495MHzpF/mW. Partial completion of layout design allows the estimate of silicon area to be around 0.0232mm2. Finally, in addition to typical corner simulation, the robustness of the circuit was tested with PVT corner simulations and Monte Carlo simulations.Bachelor's degre
Advanced motion planning using learning and non-learning-based methods for mobile robots
No full textAutonomous Mobile Robots (AMRs) require robust navigation strategies to operate effectively in dynamic
and cluttered environments, such as warehouses, while avoiding static obstacles. This study evaluates four
local navigation planners—Dynamic Window Approach (DWA), Timed-Elastic Band (TEB), Free-space
Motion Tube (FMT), and Reinforcement Learning (RL)—focusing on their performance in static obstacle
avoidance for the non-holonomic Clearpath Jackal AMR in narrow environments. DWA, TEB, and FMT
was implemented using the ROS1 Navigation Stack for global path guidance, optimizing their parameters
for narrow corridors. For RL, a collision-avoidance policy was developed using Proximal Policy Optimization
(PPO) in NVIDIA Isaac Sim, leveraging its GPU-accelerated pipeline for efficient training and its PhysX
engine for accurate physics simulation. To improve the exploration capability of the RL agent during training,
we incorporated curriculum learning, progressively increasing task difficulty from obstacle-free navigation
to handling 90% obstacle spawning rates, and introduced soft constraint rewards, penalizing gradually as
proximity to obstacles decrease to encourage accurate and slightly less conservative navigation.
Evaluations in Isaac Sim and real-world warehouse settings revealed RL’s superior adaptability, achieving
a consistent 100% success rate in both real world and simulation, with the fastest navigation time, and zero
collisions, while smoothly navigating narrow passages. FMT also performed well in simulations with 100%
success rate and a slightly slower navigation time, but struggled in real-world tests due to noisy costmaps,
while TEB and DWA consistently failed, getting trapped in local minima in both simulations and real world
scenarios. These findings highlight RL’s potential for complex environments and provide practical insights
for selecting navigation methods in industrial applications, particularly in warehouse robotics.Bachelor's degre
Effect of chloride contamination levels on corrosion kinetics of steel bars in OPC and GGBS concretes
No full textCorrosion is one of the most common types of damage that steel reinforcement in concrete structures
experience when they are used in coastal structures. High levels of chlorine in the environment
accelerates corrosion initiation process and corrosion rates of steel bars. This study aims to study the
extent chloride has on the corrosion rate of rebars. This is done through completely submerging concrete
slabs in varying concentrations of chloride. This study also compares the effect of ground granulated
blast furnace slag (GGBS) in concrete on the corrosion rate. GGBS is used as a replacement material
for OPC as it is more sustainable and reduces CO2 emission.
Various measurement techniques were used for data collection. A potentiostat was used in a threeelectrode system where the concrete slabs had a reference electrode of copper sulfate to determine the
relative potential. Data was recorded using the EC-Lab software on a computer connected to the
potentiostat. Open circuit potential, linear polarisation resistance, ohmic drop and potentiostatic
electrochemical impedance spectroscopy were techniques applied across the span of nine weeks.
Thereafter, corrosion potential (Ecorr), polarisation resistance (Lp), corrosion current density (icorr) and
impedance (Z) are recorded and calculated. Trends are studied from the data obtained.
Experimental findings show that chloride concentration have different effects on different GGBS
replacement ratios. 40% GGBS ratio concretes experience higher corrosion rates as chloride
concentration increases. GGBS ratio has an interesting effect on the corrosion rate of rebars in concrete.
0% GGBS ratio concretes have the lowest corrosion resistance as hypothesised. However, 20% GGBS
ratio concrete has a higher corrosion resistance than 40% GGBS ratio concrete. More studies were
explored later to provide an explanation for this phenomenon.
Future work could be done to include more datapoints in the experiment. For 0% GGBS concrete, the
effect of chloride concentration can be studied properly by taking measurements at shorter intervals
before severe corrosion has taken place. More GGBS ratios can also be included to find the optimal
GGBS ratio in concrete that provides the most amount of corrosion resistance against pitting attacks by
chlorides.Bachelor's degre
CabMate
No full textCabMate is a ride-sharing platform designed to offer a sustainable, affordable, and scalable
alternative to traditional transportation options in urban areas. The platform connects
commuters with similar travel routes and timings, allowing them to share rides, thereby
reducing commuting costs and minimizing environmental impact. The primary focus of
CabMate is to alleviate the high cost of travel and provide an eco-friendly solution by enabling
users to find suitable ride partners through an easy-to-use web-based interface.
The platform leverages advanced technologies such as Firebase for backend services, ensuring
real-time data handling, secure user authentication, and efficient data storage. By
crowdsourcing the data and connecting commuters with similar schedules, CabMate provides
a solution that is both cost-effective and practical, especially for frequent commuters and
students.
The project involved comprehensive market research to validate the need for such a service,
alongside user interface design to ensure a seamless and user-friendly experience. A functional
prototype of the platform was developed and is in preparation for beta testing. The testing will
focus on gathering feedback from users to refine the platform’s features and ensure its
effectiveness in real-world scenarios. With a focus on scalability, CabMate is poised to expand
its services beyond initial use cases, including transportation between universities and airports,
and eventually extend to broader urban mobility solutions. The goal of the project is not only
to provide a practical solution for ride-sharing but also to empower users to make
environmentally conscious commuting choices.Bachelor's degre
Flow visualization and analysis of a L-shaped sprinkler
No full textThe purpose of this study is to investigate the influence of flow patterns on the behaviour of an L-shaped sprinkler system. This study seeks to address the long-standing debate regarding the accuracy of theoretical predictions made by the angular momentum equation when applied to systems like sprinklers. Furthermore, this study also addresses a sprinkler geometry, L-shaped, which does not have much prior studies into it. The experiment involved designing an L-shaped sprinkler setup that aspirates water from one end and expels it from the bent end, ensuring the flow within the system remains within the laminar region. Flow patterns were visualized by introducing dye into the flow, which formed visible streak lines. Under steady-state conditions, these streak lines could be interpreted as streamlines, representing the actual flow of water through the system. The findings of this study revealed that the flow at both the inlets and outlets was never completely uniform or parallel. Factors such as flow separation played a key role in disrupting the uniformity of the flow. These separations also caused pressure differences that induced recirculatory flows in certain regions of the sprinkler, further disturbing the flow profiles. This non-uniformity in the flow is identified as the primary cause of the inaccuracies in the predictions made by the angular momentum equation. The equation's assumptions during its derivation fail to account for these complexities, suggesting that a more general and comprehensive model is required to consider all the influencing factors and provide a more accurate prediction of sprinkler behaviour.Bachelor's degre
Software receiver front-end for GNSS positioning
No full textThis report aims to evaluate and modify a GNSS positioning program while tailoring
it to work effectively with the Adalm Pluto and Taoglas GNSS Antenna for improved
signal acquisition and positioning accuracy. The report first introduces what GNSS is
and its history. After this, the three GNSS positioning principles are explored and
researched, and for Doppler Shift, different algorithms are investigated to assess their
performance and improvement in the positioning accuracy of their receiver. The
experimental setup I tested in both Outdoor and Indoor conditions is analyzed, and
the code variables are then explained to justify actions in modifying those variables,
and their functions are summarized. Lastly, the project highlights the importance of
appropriate hardware selection for GNSS applications. It provides recommendations
for future work, including using active antennas or external amplifiers, improved
signal processing techniques, and alternative SDR platforms to enhance acquisition
reliability and accuracy.Bachelor's degre
Additive manufacturing of lightweight and high strength metal lattice structures
No full textAdditive Manufacturing (AM) is a cornerstone of modern day manufacturing, enabling the production of complex geometries with a wide variety of materials. This is particularly applicable to lattice structures, where conventional manufacturing processes are not able to produce lattice structures of certain topologies due to their manufacturing constraints. The ability of AM to produce high quality, complex lattice structures with high precision makes AM the preferred choice for the manufacturing of many parts with intricate designs and specifications. Although there has been extensive research on the properties of AM metal lattice structures, exploration into how the wall thickness affects the compressive strength of the lattice structures is limited. SS316L, a corrosion-resistant stainless steel, was chosen for its superior mechanical properties and suitability for applications requiring both strength and lightness, such as aerospace and biomedical devices.Bachelor's degre
The role of Tau, amyloid-β and neuroinflammation in the association between cognition and white matter hyperintensities in a southeast Asian cohort
No full textBackground: Elevated Glial Fibrillary Acidic Protein (GFAP) is associated with increased Phosphorylated Tau 181
(pTau181) induced neurodegeneration in Alzheimer’s Disease.
Objective: However, the role of GFAP and pTau181 in vascular/mixed dementias requires elucidation within the
Southeast Asian context, where their burden is considerable.
Design: Population based cross-sectional study.
Setting: Biomarkers and Cognition Study, Singapore (BIOCIS).
Participants: Baseline data from n = 583 (40.3 % male), non-demented but at risk, Southeast Asian community
participants, were included in this analysis. All participants displayed cognitive symptoms on the Subjective
Memory Complaints Questionnaire, although they may or may not have objective cognitive deficits and did not
meet the criteria for dementia as per the DSM – 5.
Methods: Neuropsychological assessments for executive function evaluation, volumetric White Matter Hyperintensities
(WMH) measurement and plasma biomarker expression, were determined in non-demented but at risk,
Southeast Asian research participants. Partial correlation analysis demonstrated variable associations. Simple
moderation analysis revealed the ability for plasma biomarkers to influence the relationship between executive
function and WMH.
Results: WMH burden positively correlated to Neurofilament-Light (NfL) and pTau181. Executive function and
processing speed negatively correlated to WMH burden. GFAP positively correlated to pTau181 and negatively
correlated to executive function. NfL, GFAP, pTau181, and Amyloid beta 42/Amyloid beta 40 (Aβ42/Aβ40) ratio
independently moderated, the relationship between executive function/processing speed and WMH burden.
Conclusion: Inflammatory mechanisms represented by GFAP were linked to tau pathology and WMH and also
moderated the association between WMH on cognitive performance.Ministry of Education (MOE)National Medical Research Council (NMRC)Ministry of Health (MOH)Nanyang Technological UniversitySubmitted/Accepted versionThis study received funding support from the Lee Kong Chian School of Medicine, Nanyang Technological University, Singapore, Strategic Academic Initiative Award: SP1CLNT900-NTU-A630-PJ- 03INP001400A630; the Ministry of Education, Singapore, Start-Up Grant and Academic Research Fund Tier 1 Award: RT02/21 and Tier 3 Award: MOE2017-T3-1-002 and the National Medical Research Council, Singapore, Clinician Scientist Award: MOH CSAINV18nov-0007
Development of boom-less deployment mechanisms for Solar Sail Spacecraft
No full textThis study aims to develop and test boom-less Solar Sail Spacecraft deployment systems with improved compactness, affordability, and reliability. Three ideas for deployment systems were explored: HDRM-based, centrifugal force, and foldable arm assembly, with iterative improvements in reach, space efficiency, and passive deployment. This report explains the design justification, testing processes, and comparative assessment of each system, along with an examination of origami-inspired sail-folding techniques. The results show the promise of boom-less Solar Sail Spacecraft propulsion systems, in which Design 3 was found to be the most feasible. Suggestions for further research include addressing reversibility on deployment and adaptation to microgravity.Bachelor's degre