Heriot-Watt University
ROS: The Research Output Service. Heriot-Watt University EdinburghNot a member yet
4689 research outputs found
Sort by
Vehicle trajectory prediction for safe navigation of autonomous vehicles
Get PDFTrajectory prediction of the other road users in the vicinity of an autonomous vehicle is important for safe navigation in dense traffic. Once an autonomous vehicle
anticipates how the other road actors will react in the near future, path planning is
a lot more simpler and safer. Moreover, the knowledge of future movement of other
road actors allows control of sudden jerks in the planned ego vehicle’s path and thus
makes travel smoother. This trajectory prediction stage can be used at any level,
from restricted driver assistance to full vehicle autonomy. In this thesis two novel trajectory prediction models have been developed. In the
first model, the spatio-temporal features that form the basis of behaviour prediction were captured using a Convolutional Long Short Term Memory (Conv-LSTM)
neural network architecture consisting of three modules: 1) Interaction Learning to
capture the motion of and interaction with surrounding cars, 2) Temporal Learning
to identify the dependency on past movements and 3) Motion Learning to convert
the extracted features from these two modules into future positions. In addition,
a novel feedback scheme was introduced in which the current predicted positions
of each car are leveraged to update future motion, encapsulating the effect of the
surrounding cars. In the second model a conventional Long Short Term Memory
(LSTM) cell based encoder-decoder architecture was developed which uses not only
the historical observations but also the associated map features. Moreover, unlike
existing architectures, the proposed method incorporates and updates the surrounding vehicle information in both the encoder and decoder, making use of dynamically
predicted new data for accurate prediction in longer time horizons. This seamlessly
performs four tasks: first, it encodes a feature given the past observations, second,
it estimates future maneuvers given the encoded state, third, it predicts the future
motion given the estimated maneuvers and the initially encoded states, and fourth,
it estimates future trajectory given the encoded state and the predicted maneuvers
and motions. Both the developed models were evaluated extensively on two publicly available datasets which include both multi-lane highway and signalled intersections,
to benchmark the prediction accuracy with the state-of-the-art models. Later, the
conventional encoder-decoder model was also evaluated with a newly collected “Radiate” dataset which includes two intersections, the Kingussie T-junction and the
Edinburgh four-way junction, both without traffic signals. The accuracy of the predicted trajectories on the benchmark datasets are comparable with state-of-the-art
methods. Moreover, evaluation on the latter dataset (“Radiate”) made it possible
to understand better the effect of inter-vehicle interactions on future motion without
any influence from mandatory traffic signals.Engineering and Physical Sciences Research Council (EPSRC) funding
Understanding business model innovation in start-ups – a dynamic managerial capabilities perspective
Get PDFAlthough start-ups and incumbent firms both engage in business model innovation, the
research literature on business model innovation has largely focused on incumbent firms.
Start-ups play an important role in the growth of an economy, and with their failure rate being
high, the need for an adequate business model and continuously innovating it, is crucial for
financial performance and competitive advantage. Underpinning this study is the
understanding that a business model represents how a firm creates, delivers, and captures
value. Moreover, business model innovation involves reconfiguring components or the
architecture of a business model for the benefit of the firm, which requires capabilities.
The importance of the dynamic capabilities of a firm, and the need to quickly identify and
respond to opportunities, and consequently innovate business models has previously been
noted. However, gaps still exist in how business model innovation is understood with respect
to dynamic managerial capabilities in start-ups. This study contributes to the body of
knowledge by exploring the influence of capabilities on business model innovation in start-ups.
Based on the theory of dynamic managerial capabilities, the following research question was
examined: What capabilities allow managers (e.g. founders, decision makers) in start-ups to
innovate their business models?
This study was situated within the critical realism paradigm and used multiple explorative
cross-sectional case studies on start-ups. The primary source of data was the subjective
experience of managers, who were recruited through purposive sampling. Semi-structured
interviews were used to elicit the data, which was recorded and transcribed verbatim. The
transcripts were analysed using thematic analysis, and documentary evidence was used for
triangulation. The findings highlight two main dimensions of capabilities that enable start-up
managers to innovate their business models. These are collaboration capabilities (comprising
of networking, commitment, and internal cooperation) and capitalization capabilities
(comprising of experience, searching, and maximizing resources). This study draws attention to
the need for managers to foster these capabilities, and the implications for professional
managerial practises and the research literature are delineated
How digitalisation in manufacturing drives business model innovation in the optical lens industry
Get PDFDigitalisation as a megatrend disrupts traditional manufacturing value chains and
associated business models. Deciders face substantial challenges driven by digitalisation
in manufacturing whilst shaping business models and enabling value creation. Yet,
academic literature reveals gaps in understanding, how digitalisation in manufacturing
can leverage business model innovation.
Thus, the purpose of this study is to explore, how digitalisation in manufacturing drives
business model innovation for the optical lens manufacturing industry. The research is
based on a case study to understand experiences and investigate patterns, how value
creation opportunities from emerging digitalisation adoptions in manufacturing can be
exploited for business model innovation. Following a pilot study, semi-structured
interviews are applied to gain in-depth understanding about experiences with the
phenomenon. The interviews are conducted with people in key roles, who put business
model innovation and digitalisation in manufacturing into practice. The collected data is
assessed adopting a thematical analysis approach to derive themes and patterns.
Adding to recent academic theory and managerial implications regarding business model
innovation, the findings of this thesis contribute to an understanding, how digitalisation
adoptions can be exploited for business model innovation, how business model
innovation can be deployed, and how value creation can be driven by digitalisation in
manufacturing. The findings provide insights into the challenge that business model
innovations often happen on a greenfield, whereas manufacturing processes behind the
business model are already established. Best practices for business model innovation
seem to emerge out of personal experiences. External and internal motivators trigger
business model innovation, however, customers stimulate innovations by rather
communicating their problems instead of particular ideas. Furthermore, the value added
by business model innovation is not always assessed looking at financials, but includes
industry-specific performance criteria as well
Visually grounded representation learning using language games for embodied AI
Get PDFThe ability to communicate in Natural Language is considered one of the ingredients
that facilitated the development of humans’ remarkable intelligence. Analogously, developing artificial agents that can seamlessly integrate with humans requires them to
understand, and use Natural Language, just like we do. Humans use Natural Language to coordinate and communicate relevant information to solve their tasks—they
play so-called “language games”. In this thesis work, we explore computational models
of how meanings can materialise in situated and embodied language games. Meanings
are instantiated when language is used to refer to, and to do things in the world. In
these activities, such as “guessing an object in an image” or “following instructions to
complete a task”, perceptual experience can be used to derive grounded meaning representations. Considering that di↵erent language games favour the development of specific
concepts, we argue it is detrimental to evaluate agents on their ability to solve a single
task. To mitigate this problem, we define GroLLA, a multi-task evaluation framework
for visual guessing games that extends a goal-oriented evaluation with auxiliary tasks
aimed at assessing the quality of the representations as well. By using this framework,
we demonstrate the inability of recent computational models to learn truly multimodal
representations that can generalise to unseen object categories. To overcome this issue,
we propose a representation learning component that derives concept representations
from perceptual experience, obtaining substantial gains over the baselines—especially
when unseen object categories are involved. To demonstrate that guessing games are
a generic procedure for grounded language learning, we present SPIEL, a novel self-play procedure to transfer learned representations to novel multimodal tasks. We show
that models trained in this way can obtain better performance as well as learn better
concept representations than competitors. Thanks to this procedure, artificial agents
can learn from interaction using any image-based datasets. Additionally, learning the
meaning of concepts involves understanding how entities interact with other entities in
the world. For this purpose, we use action-based and event-driven language games to
study how an agent can learn visually grounded conceptual representations from dynamic scenes. We design EmBERT, a generic architecture for an embodied agent able
to learn representations useful to complete language-guided action execution tasks in
a 3D environment. Finally, learning visually grounded representations can be achieved
when watching others completing a task. Inspired by this idea, we study how to learn
representations from videos that can be used for tackling multimodal tasks such as commentary generation. For this purpose, we define Goal, a highly multimodal benchmark
based on football commentaries that requires models to learn very fine-grained and rich
representations to be successful. We conclude with some future directions for further
progress in computational learning of grounded meaning representations
Deaf business owners’ experiences of and strategies in navigating an audist normative structured labour market in Denmark
Get PDFThis PhD investigates deaf-led businesses, an emerging phenomenon in Denmark between
2000-2017. The data consists of interviews with nine deaf business owners, supported by
observations made on visits to the businesses and interviews with three employees. The
businesses fall into two groups: those oriented towards hearing, private customers, and those
oriented towards deaf customers for whom the services receive public funding. This study
employs Bourdieu’s theoretical framework regarding how people navigate their social contexts
based on their habitus and forms of capital.
The study demonstrates the ways in which deaf business owners’ life experiences in Denmark
are influenced by structures of inequality. Research into disabled people’s work experiences
has shown that the labour market rests on ableist values. Ableist values encompass audist
values, whereby hearing and speaking are privileged; deaf people are, therefore, disadvantaged
by not fitting the template of ‘ideal worker’. Deaf people set up businesses so as to become
their own boss and pursue professional interests; this study also reveals that direct or indirect
discrimination may motivate them to seek alternatives to traditional employment in the
‘hearing’ labour market.
Secondly, the study explains how deaf people strategically navigate their hearing surroundings
as business owners. Owners of businesses oriented towards the hearing market use adaptive
strategies to ‘pass’ as business owners, undertaking significant invisible labour to expand their
‘hearing’ cultural and social capital. Those with businesses aimed at deaf customers show the
opposite approach, using strategic isolationism (e.g. avoiding social contact with hearing
people; preferring to employ deaf people) to oppose the audist values of the surrounding labour
market. Being seen as deaf creates the expectation of ‘deaf cultural and social capital’, e.g. sign
language skills, understanding deaf customers’ needs, and the ability to navigate deaf contexts.
Thirdly, the study shows that the emergence of deaf business ownership has created new
opportunities for deaf people in the labour market and a professional context where deaf skills
can be capitalised on and where deaf people’s social networks, behaviours and values are
advantages rather than disadvantages. However, deaf-led businesses provide a limited number
of new jobs, and deaf people still face challenges in the broader labour market.
This study contributes theoretically to the body of research concerning minority-led businesses,
and also to the disciplines of Deaf Studies and Disability Studies in general
Transformative service research, service dominant logic and financial wellbeing : exploring a service ecosystem approach to student financial capability support at a Scottish University
Get PDFThis thesis proposes adopting a service eco-system approach to developing financial
capability that can result in an improvement of the financial well-being (FWB) amongst
university students. To achieve this aim, the thesis draws on insights and implications
from transformative services research (TSR) and service-dominant logic (SDL) to
develop a financial capability intervention. The thesis focuses on student FWB because
students are more vulnerable to the consequences of negative financial behaviour since
they are at a stage in life where their skills, attitudes and financial practices are still
emerging. There is extensive literature confirming that poor financial management can
affect students’ health, academic performance and future employment opportunities.
The thesis answers the research question, “how can an actor-to-actor ecosystem provide
transformative service that results in improved financial well-being for students?” This
was done by developing the Well-being Support Actor to Actor (A2A) Ecosystem
Framework that identified a service ecosystem made up of different actors who are co creating value and interacting through service co-design and resource integration that
brings about wellbeing outcomes for end-users.
The designed conceptual framework was tested using a Mixed Methods Research
approach. Data was collected through an online survey (n=149) of university students
that preceded the piloting of a collaborative intervention design, in which a co-design
workshop was an integral element. The qualitative data was collected through
interviews (n=26) with participants from within the university ecosystem, observations
made during the co-design workshop and the roll -out of the intervention. The
qualitative evidence shows that universities can play a crucial role in improving
financial well-being through student-led financial capability programs. There is also
evidence to the fact that institutional arrangements and context of co-creation can limit
and affect the results of user generated solutions. The study also contributes to
knowledge about the psychological influences (attitudes towards debt, attitude towards
money and locus of control) on the financial behaviours, FWB and financial status of
students. On the practical side, data shows how transformative service organisations -
like credit unions and universities - and end-users like students can develop
interventions that improve the financial capability of young people. There is ample
evidence that students can lead the process and use their lived experiences to help their peers relate with their own struggles as they navigate through the financial management
maze. Policymakers and decision-makers at universities should deliberately move to
close the observed gaps between university managers’ claims about the need to provide
financial capability support to students and their lack of response when there was an
emerging solution coming from the end-users themselves. Measures can be put in place
to provide the requisite support for student-led financial education and counselling to
take place within the university setting
General Relativity analogues in nonlinear optical systems
Get PDFThis thesis presents the results of the experimental and numerical investigations of
nonlinear optical systems in relation to their analogy with some General Relativity
phenomena.
After a short introduction on the context of this work, the first Chapter starts
with an introduction on the Newton-Schrodinger Equation and its optical analogue,
then the investigations of two main effects (the analogue Newtonian gravitational
interaction of two optical beams, and violent relaxation process in analogue galaxy
formation) are illustrated. For these, the theoretical background is presented, then
experimental setups and results are compared with numerical simulations.
The second Chapter starts with a brief literature review on Penrose superradiance
and its optical analogue. This is followed by the experimental setup and results of the
investigation of the nonlinear optical properties of the medium and the measurement
of the time stability of a vortex beam in the medium, followed by the investigation
of the nonlinear interaction of a pump-probe setup by means of a four-wave mixing
mechanism, giving rise to superradiant scattering. The corresponding experimental
setups and results are illustrated in details, as well as the comparison with numerical
simulations.
A short conclusion summarises the whole work
Antecedents of consumer pre-banking behaviour in the Ghanaian banking sector
Get PDFConsumer behaviour is constantly evolving. The service marketing landscape therefore continues to
evolve along with the dynamic ways in which consumers choose to interact with their service
providers. Accordingly, the study of consumer decision-making is one of the dominant trends in
consumer behaviour research. While consumer pre-purchase behaviour is a thoroughly studied field in
other service sectors including the hospitality and telecommunication (Zhang et al., 2019; Tommasetti
et al., 2018), the pre-purchase phase of consumer behaviour for financial services especially banking,
has been under-researched. In Ghana, the banking sector caters for about half (58.0%) of the bankable
population (World Bank, 2018). However, applying the appropriate antecedents of consumer pre-banking behaviour has been a perennial challenge for the sector. Consequently, it is costing banks
more to attract prospective consumers. This study explored the determinants of consumer pre-banking
behaviour in the Ghanaian banking sector.
The study developed a tri-component model that explained consumer pre-purchasing behaviour by
extending the constructs of the Theory of Planned Behaviour (TPB) to include affective and conative
components. This has addressed the theoretical gaps in former research that cognitively applied the
TPB. The study sampled 210 retail banking consumers who have recently (in the past six months)
purchased a banking product in the Ghanaian banking sector. The positivist paradigm was used by
adopting the exploratory sequential design of the mixed-methods approach. Thus, the study
qualitatively identified the key indicators making up the constructs of consumer pre-banking
behaviour, and empirically tested the identified latent variables in a subsequent main quantitative
survey. The model was estimated using a PLS-SEM path analysis through the SmartPLS 3.0. A
substantial amount (41.0%) of the variation in consumer pre-banking behaviour was explained by the
model. Affective, conative and cognitive attitude, and perceived trust were significant antecedents of
consumer pre-banking behaviour in the Ghanaian banking sector. Also, consumer emotions and habits
can exist along with their cognition throughout the pre-banking decision-making process. Perceived
risk partially mediated the relationship between perceived trust and consumer pre-banking behaviour.
The study has advanced the theoretical knowledge in decision heuristics and cognitive bias regarding
consumer pre-banking decision-making. The study also provides practical proposition of how
marketing practitioners in other financial sectors such as insurance can emotionally connect with
consumers and develop targeted communication strategies with consumers at the pre-purchasing stage
Lateral bending liquid crystal elastomer beams for microactuators and microgrippers
Get PDFWith the rapid development of microsystems in the last few decades, there is a
requirement for high precision tools for micromanipulation and transportation of micro-objects, such as microgrippers, for applications in microassembly, microrobotics, life
sciences and biomedicine. Polymer based microgrippers and microrobots executing
various tasks have been of significant interest as an alternative to the traditional silicon
and metal based counterparts due to the advantages of low cost fabrication, low
actuation temperature, biocompatibility, and sensitivity to various stimuli. The
exceptional actuation properties of liquid crystal elastomers (LCE) have made these
materials highly attractive for various emerging applications in the last two decades.
Large programmable deformations and the benefits offered by the elastic, thermal and
optical properties of LCEs are suitable for implementing stimuli-responsive
microgrippers as well as various biomimetic motion in soft robots.
In this thesis, a method and the associated processes for fabrication and molecular
alignment in LCE were developed, which enabled new functionality and improved
performance of the LCE based microactuators and microgrippers, providing controlled
response by thermal and remote photothermal actuation, and allowing easy integration
of the LCE end-effectors into robotic systems for automated operation. Lateral bending
actuation has been demonstrated in LCE microbeams of 900 µm of length and 40 µm of
thickness, owing to the new monolithic micromolding technique using vertical patterned
walls for alignment. The effects of parameters such as the beam width, the size of the
microgrooves, and the surface treatment method on the behavior of the microactuators
were studied; the internal alignment pattern of liquid crystals in the structure was
investigated by different microscopy methods. An efficient method for finite element
modeling of the bending LCE actuators was developed and experimentally verified,
based on the gradient of equivalent thermal expansion in the multi-layer structure,
which was able to predict the bending behavior of the actuators in a large range of
thicknesses as well as rolling behavior of the actuators of tapered thickness. The novel
LCE microgripper with in-plane operation showed efficient thermal and photothermal
actuation, achieving the gripping stroke of 64 µm under the light intensity of 239
mW/cm2
for the gripper length of 900 µm, which is more efficient than the typical SU-8
polymer based microgrippers of the same dimensions. The LCE gripper was
successfully demonstrated for the application in manipulation of the objects of tens to
hundreds of micrometers in size. Therefore, the novel LCE microgripper bridges the gap in the LCE-based gripper technologies for typical object size in applications for
systems microassembly, biological and cell micromanipulation. The lateral bending
functionality enabled by the proposed method expands design opportunities for thermal
and photothermal LCE microactuators, providing an effective route toward realization
of new modes of gripping, locomotion, and cargo transportation in soft microrobotics
and micromanipulation
High-speed imaging with optical encoding and compressive sensing
Get PDFImaging instruments can be used to obtain a series of frames in domains such as
frequency and time. Recent advancements in applications such as medical astronomical, scientic and the consumer application, demand overall improvements in
these imaging systems. Many current imaging methods rely on the well-known
Shannon-Nyquist theorem where sustaining this conventional model increases the
system complexity, data rate, storage and processing power as well as the overall
build costs of these units. Recent investigations based on the mathematical theory
of compressed sensing (CS) have broken the traditional sampling mechanisms and
introduces alternative methods of data sampling.
This dissertation investigates the current advancements in the high-speed imaging schemes and proposes new methods and optical designs to improve the spatial
and temporal resolution as well as the required transmission and storage capacity
of the imaging systems. First, we investigate the current mathematical models of
CS based algorithms in video acquisition systems and propose an improved adapted
technique for data reconstruction. Then we investigate the state-of-the-art high-speed imaging methods and introduce optical encoding techniques that enable the
current high-speed imaging systems to reach 10 times faster frame rates whilst preserving the spatial resolution of the existing systems. Second, we develop a novel
high-speed imaging system that implements CS based optical imaging technique
and experimentally demonstrate the operation of this novel imaging system. The
proposed compressive coded rotating mirror (CCRM) camera benefits from noticeably improved physical dimensions, highly reduced build costs and the significantly
simplified operation compared to the other high-speed cameras. Due to the built-in
optical encoding and on-the-fly compression functionalities of CCRM camera, it becomes a viable option for the fields such as the medical and military based imaging
applications where the security of the data remains one of the top priorities in the
imaging instruments. Finally, we discuss the potential improvements on the CCRM
camera and propose several advancement plans for the future of this system
Access Repository Dashboard
Do you manage ROS: The Research Output Service. Heriot-Watt University Edinburgh ? Access insider analytics, issue reports and manage access to outputs from your repository in the CORE Repository Dashboard!