Pulse-coupled neural network performance for real-time identification of vegetation during forced landing

Safety concerns in the operation of autonomous aerial systems require safe-landing protocols be followed during situations where the a mission should be aborted due to mechanical or other failure. On-board cameras provide information that can be used in the determination of potential landing sites, which are continually updated and ranked to prevent injury and minimize damage. Pulse Coupled Neural Networks have been used for the detection of features in images that assist in the classification of vegetation and can be used to minimize damage to the aerial vehicle. However, a significant drawback in the use of PCNNs is that they are computationally expensive and have been more suited to off-line applications on conventional computing architectures. As heterogeneous computing architectures are becoming more common, an OpenCL implementation of a PCNN feature generator is presented and its performance is compared across OpenCL kernels designed for CPU, GPU and FPGA platforms. This comparison examines the compute times required for network convergence under a variety of images obtained during unmanned aerial vehicle trials to determine the plausibility for real-time feature detection.

Educating for digital futures : what the learning strategies of digital content professionals can teach higher education

Beginning in the second half of the 20th century, ICTs transformed many societies from industrial societies in which manufacturing was the central focus, into knowledge societies in which dealing effectively with data and information has become a central element of work (Anderson, 2008). To meet the needs of the knowledge society, universities must reinvent their structures and processes, their curricula and pedagogic practices. In addition to this, of course higher education is itself subject to the sweeping influence of ICTs. But what might effective higher education look like in the 21st century? In designing higher education systems and learning experiences which are responsive to the learning needs of the future and exploit the possibilities offered by ICTs, we can learn much from the existing professional development strategies of people who are already successful in 21st century fields, such as digital media. In this study, I ask: (1) what are the learning challenges faced by digital media professionals in the 21st century? (2) what are the various roles of formal and informal education in their professional learning strategies at present? (3) how do they prefer to acquire needed capabilities? In-depth interviews were undertaken with successful Australian digital media professionals working in micro businesses and SMEs to answer these questions. The strongest thematic grouping that emerged from the interviews related to the need for continual learning and relearning because of the sheer rate of change in the digital media industries. Four dialectical relationships became apparent from the interviewees’ commentaries around the learning imperatives arising out of the immense and continual changes occurring in the digital content industries: (1) currency vs best practice (2) diversification vs specialisation of products and services (3) creative outputs vs commercial outcomes (4) more learning opportunities vs less opportunity to learn. These findings point to the importance of ‘learning how to learn’ as a 21st century capability. The interviewees were ambivalent about university courses as preparation for professional life in their fields. Higher education was described by several interviewees as having relatively little value-add beyond what one described as “really expensive credentialling services.” For all interviewees in this study, informal learning strategies were the preferred methods of acquiring the majority of knowledge and skills, both for ongoing and initial professional development. Informal learning has no ‘curriculum’ per se, and tends to be opportunistic, unstructured, pedagogically agile and far more self-directed than formal learning (Eraut, 2004). In an industry impacted by constant change, informal learning is clearly both essential and ubiquitous. Inspired by the professional development strategies of the digital media professionals in this study, I propose a 21st century model of the university as a broad, open learning ecology, which also includes industry, professionals, users, and university researchers. If created and managed appropriately, the university learning network becomes the conduit and knowledge integrator for the latest research and industry trends, which students and professionals alike can access as needed.

Building adaptable agent-based models – application to the electricity distribution network

Agent-based modelling (ABM), like other modelling techniques, is used to answer specific questions from real world systems that could otherwise be expensive or impractical. Its recent gain in popularity can be attributed to some degree to its capacity to use information at a fine level of detail of the system, both geographically and temporally, and generate information at a higher level, where emerging patterns can be observed. This technique is data-intensive, as explicit data at a fine level of detail is used and it is computer-intensive as many interactions between agents, which can learn and have a goal, are required. With the growing availability of data and the increase in computer power, these concerns are however fading. Nonetheless, being able to update or extend the model as more information becomes available can become problematic, because of the tight coupling of the agents and their dependence on the data, especially when modelling very large systems. One large system to which ABM is currently applied is the electricity distribution where thousands of agents representing the network and the consumers’ behaviours are interacting with one another. A framework that aims at answering a range of questions regarding the potential evolution of the grid has been developed and is presented here. It uses agent-based modelling to represent the engineering infrastructure of the distribution network and has been built with flexibility and extensibility in mind. What distinguishes the method presented here from the usual ABMs is that this ABM has been developed in a compositional manner. This encompasses not only the software tool, which core is named MODAM (MODular Agent-based Model) but the model itself. Using such approach enables the model to be extended as more information becomes available or modified as the electricity system evolves, leading to an adaptable model. Two well-known modularity principles in the software engineering domain are information hiding and separation of concerns. These principles were used to develop the agent-based model on top of OSGi and Eclipse plugins which have good support for modularity. Information regarding the model entities was separated into a) assets which describe the entities’ physical characteristics, and b) agents which describe their behaviour according to their goal and previous learning experiences. This approach diverges from the traditional approach where both aspects are often conflated. It has many advantages in terms of reusability of one or the other aspect for different purposes as well as composability when building simulations. For example, the way an asset is used on a network can greatly vary while its physical characteristics are the same – this is the case for two identical battery systems which usage will vary depending on the purpose of their installation. While any battery can be described by its physical properties (e.g. capacity, lifetime, and depth of discharge), its behaviour will vary depending on who is using it and what their aim is. The model is populated using data describing both aspects (physical characteristics and behaviour) and can be updated as required depending on what simulation is to be run. For example, data can be used to describe the environment to which the agents respond to – e.g. weather for solar panels, or to describe the assets and their relation to one another – e.g. the network assets. Finally, when running a simulation, MODAM calls on its module manager that coordinates the different plugins, automates the creation of the assets and agents using factories, and schedules their execution which can be done sequentially or in parallel for faster execution. Building agent-based models in this way has proven fast when adding new complex behaviours, as well as new types of assets. Simulations have been run to understand the potential impact of changes on the network in terms of assets (e.g. installation of decentralised generators) or behaviours (e.g. response to different management aims). While this platform has been developed within the context of a project focussing on the electricity domain, the core of the software, MODAM, can be extended to other domains such as transport which is part of future work with the addition of electric vehicles.

Traversability estimation for a planetary rover via experimental kernel learning in a Gaussian process framework

A critical requirement for safe autonomous navigation of a planetary rover is the ability to accurately estimate the traversability of the terrain. This work considers the problem of predicting the attitude and configuration angles of the platform from terrain representations that are often incomplete due to occlusions and sensor limitations. Using Gaussian Processes (GP) and exteroceptive data as training input, we can provide a continuous and complete representation of terrain traversability, with uncertainty in the output estimates. In this paper, we propose a novel method that focuses on exploiting the explicit correlation in vehicle attitude and configuration during operation by learning a kernel function from vehicle experience to perform GP regression. We provide an extensive experimental validation of the proposed method on a planetary rover. We show significant improvement in the accuracy of our estimation compared with results obtained using standard kernels (Squared Exponential and Neural Network), and compared to traversability estimation made over terrain models built using state-of-the-art GP techniques.

An Artificial Neutral Network (ANN) model for predicting biodiesel kinetic viscosity as a function of temperature and chemical compositions

An Artificial Neural Network (ANN) is a computational modeling tool which has found extensive acceptance in many disciplines for modeling complex real world problems. An ANN can model problems through learning by example, rather than by fully understanding the detailed characteristics and physics of the system. In the present study, the accuracy and predictive power of an ANN was evaluated in predicting kinetic viscosity of biodiesels over a wide range of temperatures typically encountered in diesel engine operation. In this model, temperature and chemical composition of biodiesel were used as input variables. In order to obtain the necessary data for model development, the chemical composition and temperature dependent fuel properties of ten different types of biodiesels were measured experimentally using laboratory standard testing equipments following internationally recognized testing procedures. The Neural Networks Toolbox of MatLab R2012a software was used to train, validate and simulate the ANN model on a personal computer. The network architecture was optimised following a trial and error method to obtain the best prediction of the kinematic viscosity. The predictive performance of the model was determined by calculating the absolute fraction of variance (R2), root mean squared (RMS) and maximum average error percentage (MAEP) between predicted and experimental results. This study found that ANN is highly accurate in predicting the viscosity of biodiesel and demonstrates the ability of the ANN model to find a meaningful relationship between biodiesel chemical composition and fuel properties at different temperature levels. Therefore the model developed in this study can be a useful tool in accurately predict biodiesel fuel properties instead of undertaking costly and time consuming experimental tests.

A recurrent network in the lateral amygdala : a mechanism for coincidence detection

Synaptic changes at sensory inputs to the dorsal nucleus of the lateral amygdala (LAd) play a key role in the acquisition and storage of associative fear memory. However, neither the temporal nor spatial architecture of the LAd network response to sensory signals is understood. We developed a method for the elucidation of network behavior. Using this approach, temporally patterned polysynaptic recurrent network responses were found in LAd (intra-LA), both in vitro and in vivo, in response to activation of thalamic sensory afferents. Potentiation of thalamic afferents resulted in a depression of intra-LA synaptic activity, indicating a homeostatic response to changes in synaptic strength within the LAd network. Additionally, the latencies of thalamic afferent triggered recurrent network activity within the LAd overlap with known later occurring cortical afferent latencies. Thus, this recurrent network may facilitate temporal coincidence of sensory afferents within LAd during associative learning.

Estimation of biodiesel properties from chemical composition – An artificial neural network (ANN) approach

Biodiesel, produced from renewable feedstock represents a more sustainable source of energy and will therefore play a significant role in providing the energy requirements for transportation in the near future. Chemically, all biodiesels are fatty acid methyl esters (FAME), produced from raw vegetable oil and animal fat. However, clear differences in chemical structure are apparent from one feedstock to the next in terms of chain length, degree of unsaturation, number of double bonds and double bond configuration-which all determine the fuel properties of biodiesel. In this study, prediction models were developed to estimate kinematic viscosity of biodiesel using an Artificial Neural Network (ANN) modelling technique. While developing the model, 27 parameters based on chemical composition commonly found in biodiesel were used as the input variables and kinematic viscosity of biodiesel was used as output variable. Necessary data to develop and simulate the network were collected from more than 120 published peer reviewed papers. The Neural Networks Toolbox of MatLab R2012a software was used to train, validate and simulate the ANN model on a personal computer. The network architecture and learning algorithm were optimised following a trial and error method to obtain the best prediction of the kinematic viscosity. The predictive performance of the model was determined by calculating the coefficient of determination (R2), root mean squared (RMS) and maximum average error percentage (MAEP) between predicted and experimental results. This study found high predictive accuracy of the ANN in predicting fuel properties of biodiesel and has demonstrated the ability of the ANN model to find a meaningful relationship between biodiesel chemical composition and fuel properties. Therefore the model developed in this study can be a useful tool to accurately predict biodiesel fuel properties instead of undertaking costly and time consuming experimental tests.

Pulse-coupled neural network performance for real-time identification of vegetation during forced landing

Safety concerns in the operation of autonomous aerial systems require safe-landing protocols be followed during situations where the mission should be aborted due to mechanical or other failure. This article presents a pulse-coupled neural network (PCNN) to assist in the vegetation classification in a vision-based landing site detection system for an unmanned aircraft. We propose a heterogeneous computing architecture and an OpenCL implementation of a PCNN feature generator. Its performance is compared across OpenCL kernels designed for CPU, GPU, and FPGA platforms. This comparison examines the compute times required for network convergence under a variety of images to determine the plausibility for real-time feature detection.

Predicting freeway pavement construction cost using a back-propagation neural network : a case study in Henan, China

The objective of this research was to develop a model to estimate future freeway pavement construction costs in Henan Province, China. A comprehensive set of factors contributing to the cost of freeway pavement construction were included in the model formulation. These factors comprehensively reflect the characteristics of region and topography and altitude variation, the cost of labour, material, and equipment, and time-related variables such as index numbers of labour prices, material prices and equipment prices. An Artificial Neural Network model using the Back-Propagation learning algorithm was developed to estimate the cost of freeway pavement construction. A total of 88 valid freeway cases were obtained from freeway construction projects let by the Henan Transportation Department during the period 1994−2007. Data from a random selection of 81 freeway cases were used to train the Neural Network model and the remaining data were used to test the performance of the Neural Network model. The tested model was used to predict freeway pavement construction costs in 2010 based on predictions of input values. In addition, this paper provides a suggested correction for the prediction of the value for the future freeway pavement construction costs. Since the change in future freeway pavement construction cost is affected by many factors, the predictions obtained by the proposed method, and therefore the model, will need to be tested once actual data are obtained.

Engaging creativity through an action learning and action research process to develop an Indigenous art exhibition

In most art exhibitions, the creative part of the exhibition is assumed to be the artworks on display. But for the Capricornia Arts Mob’s first collective art exhibition in Rockhampton during NAIDOC Week in 2012, the process of developing the exhibition became the focus of creative action learning and action research. In working together to produce a multi-media exhibition, we learned about the collaborative processes and time required to develop a combined exhibition. We applied Indigenous ways of working – including yarning, cultural respect, cultural protocols, mentoring young people, providing a culturally safe working environment and sharing both time and food – to develop our first collective art exhibition. We developed a process that allowed us to ask deep questions, engage in a joint journey of learning, and develop our collective story. This paper explores the processes that the Capricornia Arts Mob used to develop the exhibition for NAIDOC 2012.

Skill development in first year law students : an evaluation of peer assisted learning across two law units

Universities no longer equip graduates solely with the content knowledge of their discipline, but also with prospective employment skills. Professions also seek graduates who can ‘collaborate, share skills and knowledge, and communicate their ideas effectively’ (Kruck and Reif, 2001, p 37). However, as admission to university does not always guarantee that one is well equipped for the task, first year students also need guidance in the development of academic skills. This session describes two models of peer assisted learning embedded within the Torts and Legal Foundations B units at the Faculty of Law, Queensland University of Technology, and how they are used to supplement student understanding of substantive law with the development of academic and work-related skills. Student perceptions of the programs developed are considered, together with the challenges faced. Session participants will be asked to contribute to a discussion of these challenges and to offer ideas on their redress.

The QUT innovation space : a trans-disciplinary learning environment for entrepreneurship education : final report

In the current climate of global economic volatility, there are increasing calls for training in enterprising skills and entrepreneurship to underpin the systemic innovation required for even medium-term business sustainability. The skills long-recognised as the essential for entrepreneurship now appear on the list of employability skills demanded by industry. The QUT Innovation Space (QIS) was an experiment aimed at delivering entrepreneurship education (EE), as an extra-curricular platform across the university, to the undergraduate students of an Australian higher education institute. It was an ambitious project that built on overseas models of EE studied during an Australian Learning and Teaching Council (ALTC) Teaching Fellowship (Collet, 2011) and implemented those approaches across an institute. Such EE approaches have not been attempted in an Australian university. The project tested resonance not only with the student population, from the perspective of what worked and what didn’t work, but also with every level of university operations. Such information is needed to inform the development of EE in the Australian university landscape. The QIS comprised a physical co-working space, virtual sites (web, Twitter and Facebook) and a network of entrepreneurial mentors, colleagues, and students. All facets of the QIS enabled connection between like-minded individuals that underpins the momentum needed for a project of this nature. The QIS became an innovation community within QUT. This report serves two purposes. First, as an account of the QIS project and its evolution, the report serves to identify the student demand for skills and training as well as barriers and facilitators of the activities that promote EE in an Australian university context. Second, the report serves as a how-to manual, in the tradition of many tomes on EE, outlining the QIS activities that worked as well as those that failed. The activities represent one measure of QIS outcomes and are described herein to facilitate implementation in other institutes. The QIS initially aimed to adopt an incubation model for training in EE. The ‘learning by doing’ model for new venture creation is a highly successful and high profile training approach commonly found in overseas contexts. However, the greatest demand of the QUT student population was not for incubation and progression of a developed entrepreneurial intent, but rather for training that instilled enterprising skills in the individual. These two scenarios require different training approaches (Fayolle and Gailly, 2008). The activities of the QIS evolved to meet that student demand. In addressing enterprising skills, the QIS developed the antecedents of entrepreneurialism (i.e., entrepreneurial attitudes, motivation and behaviours) including high-level skills around risk-taking, effective communication, opportunity recognition and action-orientation. In focusing on the would-be entrepreneur and not on the (initial) idea per se, the QIS also fostered entrepreneurial outcomes that would never have gained entry to the rigid stage-gated incubation model proposed for the original QIS framework. Important lessons learned from the project for development of an innovation community include the need to: 1. Evaluate the context of the type of EE program to be delivered and the student demand for the skills training (as noted above). 2. Create a community that builds on three dimensions: a physical space, a virtual environment and a network of mentors and partners. 3. Supplement the community with external partnerships that aid in delivery of skills training materials. 4. Ensure discovery of the community through the use of external IT services to deliver advertising and networking outlets. 5. Manage unrealistic student expectations of billion dollar products. 6. Continuously renew and rebuild simple activities to maintain student engagement. 7. Accommodate the non-university end-user group within the community. 8. Recognise and address the skills bottlenecks that serve as barriers to concept progression; in this case, externally provided IT and programming skills. 9. Use available on-line and published resources rather than engage in constructing project-specific resources that quickly become obsolete. 10. Avoid perceptions of faculty ownership and operate in an increasingly competitive environment. 11. Recognise that the continuum between creativity/innovation and entrepreneurship is complex, non-linear and requires different training regimes during the different phases of the pipeline. One small entity, such as the QIS, cannot address them all. The QIS successfully designed, implemented and delivered activities that included events, workshops, seminars and services to QUT students in the extra-curricular space. That the QIS project can be considered successful derives directly from the outcomes. First, the QIS project changed the lives of emerging QUT student entrepreneurs. Also, the QIS activities developed enterprising skills in students who did not necessarily have a business proposition, at the time. Second, successful outcomes of the QIS project are evidenced as the embedding of most, perhaps all, of the QIS activities in a new Chancellery-sponsored initiative: the Leadership Development and Innovation Program hosted by QUT Student Support Services. During the course of the QIS project, the Brisbane-based innovation ecosystem underwent substantial change. From a dearth of opportunities for the entrepreneurially inclined, there is now a plethora of entities that cater for a diversity of innovation-related activities. While the QIS evolved with the landscape, the demand endpoint of the QIS activities still highlights a gap in the local and national innovation ecosystems. The freedom to experiment and to fail is not catered for by the many new entities seeking to build viable businesses on the back of the innovation push. The onus of teaching the enterprising skills, which are the employability skills now demanded by industry, remains the domain of the higher education sector.

Brain decoding based on functional magnetic resonance imaging using machine learning : a comparative study

Brain decoding of functional Magnetic Resonance Imaging data is a pattern analysis task that links brain activity patterns to the experimental conditions. Classifiers predict the neural states from the spatial and temporal pattern of brain activity extracted from multiple voxels in the functional images in a certain period of time. The prediction results offer insight into the nature of neural representations and cognitive mechanisms and the classification accuracy determines our confidence in understanding the relationship between brain activity and stimuli. In this paper, we compared the efficacy of three machine learning algorithms: neural network, support vector machines, and conditional random field to decode the visual stimuli or neural cognitive states from functional Magnetic Resonance data. Leave-one-out cross validation was performed to quantify the generalization accuracy of each algorithm on unseen data. The results indicated support vector machine and conditional random field have comparable performance and the potential of the latter is worthy of further investigation.

Standards that we can 'own' : Threshold Learning Outcomes for Journalism, Public Relations, and Media & Communication Studies

This paper outlines the progress by the JoMeC (Journalism, Media & Communication) Network in developing TLO (Threshold Learning Outcome) statements for Bachelor-level university programs in the disciplines of Journalism, Public Relations and Media & Communications Studies. The paper presents the finalised TLO statement for Journalism, and outlines moves to engage discipline-based groups to further develop preliminary TLOs for Public Relations and Media & Communication Studies. The JoMeC Network was formed in 2011, in response to requirements that from 2014 all degrees and qualifications at Australian universities would be able to demonstrate that they comply with the threshold learning standards set by the Australian Qualifications Framework (AQF). The AQF’s threshold standards define the minimum types and levels of knowledge, skills and capabilities that a student must demonstrate in order to graduate. The Tertiary Education Quality and Standards Agency (TEQSA) will use the AQF’s threshold standards as a key tool in recording and assessing the performance of higher educational institutions, and determining whether they should be registered as Australian Higher Education Providers under the Higher Education Standards Framework. The Office of Learning & Teaching (OLT) places the onus on discipline communities to collaborate in order to develop and ‘own’ the threshold learning standards that can be considered the minimum learning outcomes of university-level programs in that field. With the support of an OLT Grant, the JoMeC Network’s prime goal has been to develop three sets of discipline-specific TLOs – one each for the Journalism, Public Relations, and Media & Communications Studies disciplines. This paper describes the processes of research, consultation, drafting and ongoing revision of the TLO for Journalism. It outlines the processes that the JoMeC Network has taken in developing a preliminary TLO draft to initiate discussion of Public Relations and Media & Communication Studies. The JoMeC Network plans to hand management of further development of these TLOs to scholars within the discipline who will engage with academics and other stakeholders to develop statements that the respective disciplines can embrace and ‘own’.