1,720,984 research outputs found
Image processing and analysis for autonomous grapevine pruning.
Full text linkIn recent years, efforts are made to automate vineyard operations to cap the ever increasing labour cost. However, one of the operations that have not been completely automated is grapevine pruning. A robotic machine for grapevine pruning needs to respond to the changing physical characteristics of the environment, and to date, no algorithm can accurately identified appropriate positions for grapevine pruning in a variety of environmental conditions. The aim of this research was therefore to develop a new algorithm using image processing, image analysis and stereo vision system to determine pruning positions and making automatic grapevine pruning possible. In order to get the pruning positions accurately and automatically, images taken from two cameras are processed and analysed. Utilizing the latest computer vision techniques, the algorithm takes three steps before the final cutting positions are derived. First, the uploaded images are pre-processed by the so called image processing phase during which binary image is obtained from the original image. Second, image analysis technique is employed to identify different parts of grape vine and obtain the 2D positions of the cutting points. Novel algorithms are proposed to locate the cordon, the branch and finally the nodes step by step with high accuracy. Both the locating precision and computing complexity of the algorithms are considered, so that the images could be processed incessantly. Finally, cameral calibration and stereo vision were also performed after that to determine the 3D pruning point, so the system can inform the robotic machine to perform the pruning. For the experiments, images were taken from the vineyard which located in the Adelaide National Wine Centre and the University of Adelaide Waite Campus. Ten images full of canes were analysed, and an 85% success rate for pruning positions was achieved. 85% of the cutting positions found by the proposed method were consistent with those found by experts, while the rest 15% of the cutting positions found by the proposed method would not affect the growth of the grapevine significantly in the next year although they don‘t match the positions found by experts perfectly. The experiment verified the accuracy and application potential of proposed algorithm. The contribution of this research has three folds:
a) The latest research and application of automatic grapevine pruning is reviewed;
b) A novel automatic grapevine pruning algorithm using computer vision technique to automatically identify important grapevine features, locate the pruning points, and derive their three dimensional coordinates is proposed, researched and developed;
c) The proposed algorithm can be also applicable to other agricultural operation automation other than grapevine pruning.Thesis (M.Eng.Sc.) -- University of Adelaide, School of Mechanical Engineering, 201
Modelling of Magnetic Fields of Permanent Magnets with Diametrical Magnetization
Full text linkCylindrical/ring-shaped permanent magnets with diametrical magnetization can be found in many applications, ranging from electrical motors to position sensory systems. In order to calculate the magnetic field generated by a permanent magnet of this kind correctly and with low computational cost, several studies have been reported in the literature providing analytical expressions. However, these analytical expressions are either limited for an infinite cylinder or for computing the magnetic field only on the central axis of a finite cylinder. The others are derived to calculate the magnetic field at any point in three-dimensional (3D) space but only with low accuracy. This thesis presents an exact analytical model of the magnetic field generated by a diametrically magnetized cylindrical/ring-shaped permanent magnet with a limited length, which can be used to calculate the magnetic field of any point in 3D space fast and with very high accuracy. The expressions were analytically derived, based on geometrical analysis without calculating the magnetic scalar potential. Also, there is no approximation in the derivation steps that yields the exact analytical model. Three components of the magnetic field are analytically represented using complete and incomplete elliptical integrals, which are robust and have low computational cost. The accuracy and efficiency of the developed analytical model was validated using Finite Element Analysis and compared against existing models.Thesis (MPhil) -- University of Adelaide, School of Mechanical Engineering, 201
The modelling and optimal design of a three degree-of-freedom XYθz micro-motion stage
Full text linkThis thesis presents an investigation of the modelling and optimal design of a particular 3-degree-of-freedom (DOF) XYθz micro-motion stage. This stage provides micron-scale motion in X and Y directions and a rotation about the Z-axis. Such a stage can be used for applications where positioning of components with micrometre, or even nanometre positioning accuracy is required. Some applications are; the positioning of samples in a scanning-electron-microscope; the positioning of masks in lithography; aligning fibre-optics and lasers; and manipulation of micro-scale objects in micro-biology or micro-systems assembly. The XYθz micro-motion stage investigated in this study uses a particular topology of monolithic compliant mechanism and three stack piezoelectric actuators. The compliant mechanism used is a 3RRR (three revolute-revolute-revolute) parallel compliant mechanism using flexure hinges. This parallel mechanism uses three RRR linkages. Each of the three RRR linkages uses three circular profile flexure hinges. Each flexure hinge provides predominantly rotational motion about one axis. This topology of mechanism has a symmetrical structure and provides numerous advantages that make it appropriate for use in a micro-motion stage. However, as yet this topology of compliant mechanism has only been investigated by a handful of researchers and it has not been used in any commercially developed systems. The design methodology of a stage using the 3RRR compliant mechanism has not been investigated in detail. In this thesis a study is presented that investigates different approaches to model the 3RRR compliant mechanism and also considers the piezo-actuator modelling, to give the complete XYθz micro-motion stage. Three models are presented and compared; the Pseudo-Rigid-Body Model (PRBM); a two-dimensional Finite-Element-Model (2-D FEM); and a third model is developed that is similar to the PRBM, but uses analytical equations to model the multiple degree-of-freedom compliance of the flexure hinges. The models developed are then used in parametric study so that the relationship between design parameters and output behaviour can be understood. An optimal design approach is then presented to develop an XYθz micro-motion stage for a particular application in a Scanning-Electron-Microscope (SEM). Finally experimental validation of the models is presented. The results of this study indicate which modelling approaches are accurate enough to prove useful for design, while also considering which models are computationally simple enough to be efficient and easy to use. The kinematic and dynamic behaviour of the 3RRR compliant mechanism and XYθz micro-motion stage is discussed in detail. This includes; a comprehensive description of the stage workspace, defining reachable and constant-rotation workspace areas; a discussion of actuator coupling; and in depth investigation of the modes of vibration. The results of the parametric study provide useful insight to aid the design of the XYz micro-motion stage and help simplify optimal design. The parametric study also highlights the difference in trends predicted by different modelling methods, which demonstrates the importance of using an appropriate model in design. The experimental validation demonstrates the accuracy of some modelling approaches while highlighting the limited accuracy of others.Thesis (Ph.D.) -- University of Adelaide, School of Mechanical Engineering, 200
3D real-time stockpile mapping and modelling with accurate quality calculation using voxels
Full text linkStockpile blending is widely accepted as an effective method to reduce the short-term quality variations and optimise the homogeneity of bulk materials, such as iron ore. Currently, both industry practice and academic research focus on planning, scheduling and optimisation algorithms to stack a stockpile that meets the predefined quality requirements. Namely, using ‘selective stacking’ algorithms to optimise the quality of a stockpile and improve the operational efficiency. However, it has been identified that stockpiled products are currently being reclaimed at approximately 50% of their potential engineering productive rates after applying such ‘selective stacking’ methods at most iron ore loading ports in Australia. There is an evident lack of solutions to this issue in the literature. This study focuses on stockpile modelling techniques to estimate the quality of a stockpile in both stacking and reclaiming operations for consistent and efficient product quality planning and control. The main objective of this work is to build an up-to-date geometric model of a stockpile using laser scanning data and apply this model to quality calculations throughout the stacking and reclaiming operations. The significant elements of the proposed research are to: (1) upgrade a stockyard machine used to stack or reclaim the stockpile (i.e. a Bucket Wheel Reclaimer) into a mobile scanning device using Kalman filtering to measure the stockpile surface continuously; (2) build a 3D stockpile model from the measurement data in real time using polynomial and B-spline surface modelling techniques and use this model to calculate the quality of a stockpile with a great degree of accuracy when the quality composition is available; (3) associate the 3D model with the reclaiming machine model to achieve autonomous operation and predict the quality of the reclaimed material through voxelization techniques. In order to validate the developed techniques, several experimental tests were conducted using simulation and real scenarios. It was verified that the proposed 3D stockpile modelling algorithms are adequate to represent the real geometric shape with great accuracy. The percentage error in volume is better than 0.2%. Therefore, the combination of stock pile and BWR (Bucket Wheel Reclaimer) models enables the reclaiming to be conducted automatically. To the best of author’s knowledge, this is the first time that a stockpile is modelled automatically in real-time and the integration of the stockpile and BWR model generates a novel stockpile management model allows true reclaiming automation. Thus, the quality of material composition after every stacking/reclaiming operation is calculated from the geometric shape/volume, density and quality assay results. Through accomplishing this project, the quality of a stockpile and its distribution inside the stockpile can be tracked continuously and the stacking/reclaiming trajectory of the machine can be controlled precisely. By making available such information, it is then possible to develop proactive stacking or reclaiming pattern strategies with more accurate product quality grade planning and control. Therefore, the workload of current selectively stacking and reactive reclaiming algorithms can be relieved, and the production rates can be improved with good output product quality control.Thesis (Ph.D.) -- University of Adelaide, School of Mechanical Engineering, 2016
An investigation into insect chemical plume tracking using a mobile robot.
Full text linkInsects are confronted with the problem of locating food, mates, prey and hosts for their young over long distances, which they often overcome using chemical plume tracking. Tracking a plume of chemical back to its source is made difficult due to the complexity of plume structure. Turbulence and shifts in the wind direction prevail over diffusion in the spreading of an airborne chemical from a point in most cases, producing intricate plumes consisting of filaments of high chemical concentration interspersed with regions of clean air. It has been proposed that insects achieve plume tracking in this environment through variations of anemotaxis, which involves travelling upwind when an attractive chemical is perceived. This study aimed to investigate anemotaxis through the use of a mobile robot to test the efficacy of algorithms which mimic the way insects achieve plume tracking and also to determine whether these algorithms are an effective means of plume tracking for a mobile robot under a range of conditions.
To achieve the aims of this study, various plume-tracking algorithms were implemented on a mobile robot built to model a plume-tracking insect and their performance was compared under a range of wind conditions. The algorithms tested were based upon a range of plume-tracking hypotheses. The simplest algorithm was surge anemotaxis, where the robot surged upwind in the presence of an attractive chemical and performed crosswind casting (back and forth motion) in the absence of chemical. The other algorithms tested were the counterturner, where the robot zigzagged upwind, and two bounded search methods. To allow these algorithms to be appropriately implemented, a robot model was constructed that could move in two dimensions and sense the wind velocity and ion level at a point in space. An ion plume was used instead of a chemical plume in each test as it behaves in a similar manner to a chemical plume, but ion sensors have response and recovery times far more rapid than conventional chemical sensors, similar to insects.
The plume-tracking robot was tested in three series of tests. Initially, the entire range of plume-tracking algorithms was tested in a wind tunnel with fixed wind direction for a range of wind speeds and release positions. The second series of tests compared the performance of the surge anemotaxis and bounded search algorithms, again in a wind tunnel, but with a wind shift of 20° during some of the tests. The algorithms were tested with and without a direct crosswind surge response to detected wind shifts. The third set of tests examined the performance of the simple and wind shift response algorithms outdoors using natural wind to produce the plume.
All algorithms tested achieved successful plume tracking in some conditions. The surge anemotaxis and triangular bounded search algorithms were particularly successful. The tests also showed that the paths obtained from tests undertaken in natural outdoor wind conditions varied greatly from those undertaken in a wind tunnel. This indicates the need to test plume-tracking algorithms in natural environments. This is vital both in the investigation of insect plume-tracking behaviour, as insects navigate in these environments, and in the process of producing plume-tracking robots that are capable of operating effectively in these conditions.Thesis (Ph.D.) -- University of Adelaide, School of Mechanical Engineering, 200
Going Beyond Counting First Authors in Author Co-citation Analysis
Full text linkThe present study examines one of the fundamental aspects of author co-citation analysis (ACA) - the way co-citation
counts are defined. Co-citation counting provides the data on which all subsequent statistical analyses and mappings
are based, and we compare ACA results based on two different types of co-citation counting - the traditional type that
only counts the first one among a cited work's authors on the one hand and a non-traditional type that takes into
account the first 5 authors of a cited work on the other hand. Results indicate that the picture produced through this non-traditional author co-citation counting contains more coherent author groups and is therefore considerably clearer. However, this picture represents fewer specialties in the research field being studied than that produced through the traditional first-author co-citation counting when the same number of top-ranked authors is selected and analyzed. Reasons for these effects are discussed
Variations on the Author
Full text link“Variations on the Author” discusses two of Eduardo Coutinho’s recent films (Um Dia na Vida, from 2010, and Últimas Conversas, posthumously released in 2015) and their contribution to the general question of documentary authorship. The director’s filmography is characterized by a consistent yet self-effacing form of authorial self-inscription: Coutinho often features as an interviewer that rather than express opinions propels discourses; an interviewer that is good at listening. This mode of self-inscription characterizes him as an author who is not expressive but who is nonetheless markedly present on the screen. In Um Dia na Vida, however, Coutinho is completely absent form the image, while Últimas Conversas, on the contrary, includes a confessional prologue that moves the director from the margins to the center of his films. This article examines the ways in which these works stand out in the filmography of a director who offers new insights into the notion of cinematic authorship
Indoor environments robot localization by employing 802.11 wireless network one way propagation time measurements.
Full text linkIndoor service robots have been under research and development for decades, but still can not be widely applied. Lack of autonomy is one of the critical factors which limit the application of indoor service robots. Unlike industry production line robots, indoor service robots require highly accurate localization ability for navigation and automation. Commonly used GPS localization algorithms can not be applied in indoor
environments, because the indoor satellite signals are too weak for localization. Different technologies have been employed to provide robots with indoor localization abilities. Infrared sensors, Radio Frequency ID (RFID), ultrasonic sensors, laser range finders and visual cameras all have been successfully developed to provide indoor localization solutions. Except the concerns like measurements’ accuracy, size, cost to deploy, most of these localization solutions can only locate robots relevant to indoor objects, for instance walls, doors and cabinets. The localization information lacks a reference to the whole building or at least the floor where robots are working. This thesis aims to develop an indoor localization system which can provide cheap, easy to access, reliable, accurate localization information with the reference to the whole floorplan.
With the development of Wireless Local Area Network (WLAN), many researchers have developed the indoor localization systems based on WLAN. Angle of Arrival (AOA), Received Signal Strength (RSS) and Time of Arrival (TOA) are three commonly used WLAN indoor localization technologies. With the consideration of accuracy and difficulty to apply among different WLAN localization methods, One Way Propagation Time (OWPT), which is a sub-category of TOA, was developed to an accurate and low cost indoor localization system. In all the reviewed literatures, no research was found which aimed to develop an OWPT localization system for WLAN. The OWPT measurement algorithms in this thesis provide a potential for achieving sub-meter indoor localization accuracy, which is more precise than most other localization methods. Based on OWPT localization algorithms, a WLAN OWPT
indoor localization system is presented at the end of the thesis which can continuously provide robots with accurate and robust localization information in indoor environments.
The developed OWPT algorithms contain two parts: one is Synchronization between Access Points (AP) and Mobile Station (MS); another is OWPT Measurement Precision Improvement. According to 802.11 protocols, every AP in a WLAN transmits beacon frames periodically. A beacon frame contains its transfer time from an AP and that AP’s identification: Basic Service Set ID (BSSID). A MS can record the beacon frame’s arrival time. The beacon frame’s OWPT is calculated from these two recorded times. There are two challenges in the application of OWPT measurements in a WLAN. Firstly, the AP clock and the MS clock need to be highly synchronized before calculating OWPT. Secondly, an ordinary 802.11 WLAN APs’ time precision is 1 microsecond (μs) which corresponds to 30 meters (m) localization precision. The proposed OWPT algorithms utilize μs precision transfer times and
nanosecond (ns) precision arrival times to synchronize the MS and AP clocks and measure a beacon frame’s OWPT with ns precision.
A One-AP-One-MS simulation model was built to simulate transmission of beacon frames from an AP to a MS. The viability of the proposed OWPT algorithms was verified in the simulation model. Simulation results indicated that the proposed algorithms could refine WLAN OWPT measurements to meet the ns precision required. When the proposed algorithms were further verified in experiments, hardware limitations were exposed. Due to the lack of highly accurate onboard timers, the ordinary wireless adapter wireless adapters could not record beacon frame arrival times with ns precision. To test the proposed algorithms experimentally, an software timer and oscilloscope were employed to record beacon frames arrival time with ns precision respectively. Test results are detailed and analysed in the thesis.
The presented OWPT measurement algorithms will not only benefit indoor service robots, but also provide a low cost indoor localization solution for other indoor services which based on locations, for example warehouse arrangement, customer information access, entertainment etc. Further research is needed to improve hardware timers’ precision on either wireless adapters or computers to record beacon frame arrival time with ns precision. The combination of a ns precision hardware clock and the proposed OWPT measurement algorithms will allow effective implementation of high precision OWPT measurement in WLAN for indoor localization.Thesis (M.Eng.Sc.) -- University of Adelaide, School of Mechanical Engineering, 201
Constructive spiking neural networks for simulations of neuroplasticity
Full text linkArtificial neural networks are important tools in machine learning and neuroscience;
however, a difficult step in their implementation is the selection of the neural network size and
structure. This thesis develops fundamental theory on algorithms for constructing neurons in
spiking neural networks and simulations of neuroplasticity. This theory is applied in the
development of a constructive algorithm based on spike-timing- dependent plasticity (STDP) that
achieves continual one-shot learning of hidden spike patterns through neuron construction.
The theoretical developments in this thesis begin with the proposal of a set of definitions of
the fundamental components of constructive neural networks. Disagreement in terminology across the
literature and a lack of clear definitions and requirements for constructive neural networks is a
factor in the poor visibility and fragmentation of research. The proposed definitions are used as
the basis for a generalised methodology for decomposing constructive neural networks into
components to perform comparisons, design and analysis.
Spiking neuron models are uncommon in constructive neural network literature; however, spiking
neurons are common in simulated studies in neuroscience. Spike- timing-dependent construction is
proposed as a distinct class of constructive algorithm for spiking neural networks. Past algorithms
that perform spike-timing-dependent construction are decomposed into defined components for a
detailed critical comparison and found to have limited applicability in simulations of biological
neural networks.
This thesis develops concepts and principles for designing constructive algorithms that are
compatible with simulations of biological neural networks. Simulations often have orders of
magnitude fewer neurons than related biological neural systems; there- fore, the neurons in a
simulation may be assumed to be a selection or subset of a larger neural system with many neurons
not simulated. Neuron construction and pruning may therefore be reinterpreted as the transfer of
neurons between sets of simulated neurons and hypothetical neurons in the neural system.
Constructive algorithms with a functional equivalence to transferring neurons between sets allow
simulated neural networks to maintain biological plausibility while changing size.
The components of a novel constructive algorithm are incrementally developed from the principles
for biological plausibility. First, processes for calculating new synapse weights from observed
simulation activity and estimates of past STDP are developed and analysed. Second, a method for
predicting postsynaptic spike times for synapse weight calculations through the simulation of a proxy for hypothetical neurons is developed. Finally, spike-dependent conditions for neuron construction and pruning are developed and
the processes are combined in a constructive algorithm for simulations of STDP.
Repeating hidden spike patterns can be detected by neurons tuned through STDP; this result is
reproduced in STDP simulations with neuron construction. Tuned neurons become unresponsive to other
activity, preventing detuning but also preventing neurons from learning new spike patterns.
Continual learning is demonstrated through neuron construction with immediate detection of new
spike patterns from one-shot predictions of STDP convergence.
Future research may investigate applications of the developed constructive algorithm in
neuroscience and machine learning. The developed theory on constructive neural networks and
concepts of selective simulation of neurons also provide new directions for future research.Thesis (Ph.D.) -- University of Adelaide, School of Mechanical Engineering, 201
- …
