Search

Pooling is an important component in convolutional neural networks (CNNs) for aggregating features and reducing computational burden. Compared with other components such as convolutional layers and fully connected layers which are completely learned from data, the pooling component is still handcrafted such as max pooling and average pooling. This paper proposes a learnable pooling function using recurrent neural networks (RNN) so that the pooling can be fully adapted to data and other components of the network, leading to an improved performance. Such a network with learnable pooling function is referred to as a fully trainable network (FTN). Experimental results demonstrate that the proposed RNN based pooling can well approximate the existing pooling functions with just one neuron, thus making it appropriate to be used as pooling function in a network with its rich representation capability. Furthermore, experiments have shown that the proposed FTN can achieve better performance than the existing pooling methods under similar network architectures for image classification.

Detecting camouflaged moving foreground objects has been known to be difficult due to the similarity between the foreground objects and the background. Conventional methods cannot distinguish the foreground from background due to the small differences between them and thus suffer from under-detection of the camouflaged foreground objects. In this paper, we present a fusion framework to address this problem in the wavelet domain. We first show that the small differences in the image domain can be highlighted in certain wavelet bands. Then the likelihood of each wavelet coefficient being foreground is estimated by formulating foreground and background models for each wavelet band. The proposed framework effectively aggregates the likelihoods from different wavelet bands based on the characteristics of the wavelet transform. Experimental results demonstrated that the proposed method significantly outperformed existing methods in detecting camouflaged foreground objects. Specifically, the average F-measure for the proposed algorithm was 0.87, compared to 0.71 to 0.8 for the other state-of-the-art methods., 13 pages, accepted by IEEE TIP

Human activity recognition based on the smartphone sensors has the potential to impact a wide range of applications such as healthcare, smart home, and remote monitoring. For simple activities like "Sit" and "Walk", it can be distinguished relatively easily. However, for similar activities with respect to transportations, such as "Train", "Bus" and "Subway", it still remains an open problem. To tackle this problem, this paper uses the recently proposed Independently Recurrent Neural Network (IndRNN) to process data of different lengths in order to capture the temporal patterns at different granularities. The unique property of IndRNN in terms of the gradient propagation allows it to construct deep RNN networks and extract good features. The proposed method has been applied and submitted to SHL recognition challenge as "UOW_AMRL".

Recurrent neural networks (RNNs) have been widely used for processing sequential data. However, RNNs are commonly difficult to train due to the well-known gradient vanishing and exploding problems and hard to learn long-term patterns. Long short-term memory (LSTM) and gated recurrent unit (GRU) were developed to address these problems, but the use of hyperbolic tangent and the sigmoid action functions results in gradient decay over layers. Consequently, construction of an efficiently trainable deep network is challenging. In addition, all the neurons in an RNN layer are entangled together and their behaviour is hard to interpret. To address these problems, a new type of RNN, referred to as independently recurrent neural network (IndRNN), is proposed in this paper, where neurons in the same layer are independent of each other and they are connected across layers. We have shown that an IndRNN can be easily regulated to prevent the gradient exploding and vanishing problems while allowing the network to learn long-term dependencies. Moreover, an IndRNN can work with non-saturated activation functions such as relu (rectified linear unit) and be still trained robustly. Multiple IndRNNs can be stacked to construct a network that is deeper than the existing RNNs. Experimental results have shown that the proposed IndRNN is able to process very long sequences (over 5000 time steps), can be used to construct very deep networks (21 layers used in the experiment) and still be trained robustly. Better performances have been achieved on various tasks by using IndRNNs compared with the traditional RNN and LSTM. The code is available at https://github.com/Sunnydreamrain/IndRNN_Theano_Lasagne., CVPR 2018

Foreground detection has been widely studied for decades due to its importance in many practical applications. Most of the existing methods assume foreground and background show visually distinct characteristics and thus the foreground can be detected once a good background model is obtained. However, there are many situations where this is not the case. Of particular interest in video surveillance is the camouflage case. For example, an active attacker camouflages by intentionally wearing clothes that are visually similar to the background. In such cases, even given a decent background model, it is not trivial to detect foreground objects. This paper proposes a texture guided weighted voting (TGWV) method which can efficiently detect foreground objects in camouflaged scenes. The proposed method employs the stationary wavelet transform to decompose the image into frequency bands. We show that the small and hardly noticeable differences between foreground and background in the image domain can be effectively captured in certain wavelet frequency bands. To make the final foreground decision, a weighted voting scheme is developed based on intensity and texture of all the wavelet bands with weights carefully designed. Experimental results demonstrate that the proposed method achieves superior performance compared to the current state-of-the-art results., Comment: IEEE International Conference on Image Processing, 2017

The interaction of the rebar bolt profile with the surrounding medium is investigated analytically. The grout/bolt interface failure is categorised into dilational slip and parallel shear failure based on experimental observations. For each failure mode, an analytical expression for the failure is derived after introducing Mohr–Coulomb׳s failure criterion. The theoretical predictions are compared to, and showed good agreement with, experimental data. After introducing the concept of the most vulnerable slip angle in dilational slip failure, the role of the rebar profile is identified quantitatively with respect to bond failure. It is found that the optimal profile geometry is related to the mechanical properties of the grout and the confining pressure when failure occurs. Therefore, bolt profile studies must take into account the mechanical properties of the grout and confine pressure. In addition, conclusions drawn from laboratory pull out tests should be evaluated properly considering both the testing and in-situ conditions before any application. A case study has been included to demonstrate the application of the developed theory. In a coal mine roadway primary reinforcement design, rock bolts have to be selected out of the four most popular types currently used in the Australian mining industry. The advantages and disadvantages of each bolt are discussed analytically for each case, providing a way for engineers to select the optimal bolt in their application.

Short encapsulated pullout tests are frequently used in the laboratory and in the field to examine the bond performance of rock bolting. Because of the restriction of sample conditions and of testing apparatus, push testing is usually used instead in the laboratory. Rock bolts have a better performance in push testing and this is a thought to be attributed to Poisson's ratio effect of the steel rebar. In this paper Poisson's ratio effect in short encapsulated push/pull tests is evaluated analytically. The percentage difference of axial load in push and pull tests is calculated based on the failure mode analysis and the corresponding mechanical models of rock bolting mechanisms. Analytical results show that Poisson's ratio effect is closely related to the radial stiffness and embedded length. The theoretical predictions were compared with experimental data and showed good agreement, and so can be used as a guideline in laboratory testing design.

The applications of high-temperature superconductors (HTS) in electric power components have been widely reported and various studies have been made to define their alternating current (AC) losses-a key design parameter for many practical high power electrical engineering applications. However, very few studies over the range 25 to 45 K have been conducted even though this is one of the favored temperature ranges for cost-effective applications of HTS. Methods and techniques used to characterize and measure these losses have been so far grouped into `electrical' and `calorimetric' approaches with external conditions set to resemble the application conditions. In this paper, we present an approach using the calorimetric method to accurately determine losses in the superimposed AC and DC fields likely to be experienced in practical devices such as Fault Current Limiters. This technique provides great simplification compared to pick-up coil and lock-in amplifier methods and is applied to a lower temperature range. The preliminary loss data at 40 K will be presented in applied AC magnetic fields with DC fields up to 1T. The data of losses obtained on this sample will allow the estimation and minimization of losses in practical high power HTS coils and will be used in the verification of numerical coil models.

Wave variables guarantee stability for delayed teleoperation. They also introduce reflections which can prove very disorientating for an operator. This paper presents a new method for reducing the wave reflections when operating in unknown environments. The new method does not only guarantee stability but also laboratory experiments show it also reduces wave reflections allowing for improved velocity tracking and force feedback in complex and unknown operating environments.

A typical Australian 3–5 Mt/year longwall requires 9 to 15 km of roadway to be developed every year (G. Lewis and G. Gibson: Proc. 13th Austr. Tunnelling Conf., Melbourne, Australia, May 2008, ATS, 317–326.). One of the major challenges faced by the Australian longwall mining industry is that roadway development has been unable to keep pace with improvements in longwall extraction rates. This has affected the profitability, viability and future of underground coal mines (M. Kelly: 'Coordination of roadway development strategy', ACARP report no. C5013, ACARP, Brisbane, Australia, 1997.). Gibson (G. Gibson: 'Australian roadway development– current practices', ACARP report no. C15005, ACARP, Brisbane, Australia, 2005.) proposed the development of an integrated suite of management tools and procedures to study how roadway development rates can be improved. A computer based process mapping, management and monitoring system would help the adoption of a systems/process control approach. The objective of ...

Conducting polymer bending actuators show potential for unique manipulation devices, particularly at the microscale, given low actuation voltages, controllable manufacture, biocompatibility, and ability to operate in either air or liquid environments; however, the impracticalities of implementing feedback in these environments and at these scales can impede positional control of the actuator. This paper presents an application of inversion-based feedforward positional control to a trilayer bender actuator, which is shown to improve the performance without the use of feedback or adjustments to the chemistry of the device. The step and dynamic displacement responses have all been improved under the feedforward control system, while the response does not change significantly under large increases in external loads. This study contributes the first implementation of inversion-based feedforward control to the emerging area of conducting polymer actuators, paving the way toward their use in functional devices, particularly where the implementation of feedback is difficult.

Conducting polymers are new materials that can be used as low-voltage actuators and active flexure joints, scaling down to the microscale; however, many devices based on these actuators are limited because they can only operate when submerged in an electrolyte. Conducting polymer trilayer actuators are laminated structures with a wide range of potential applications as they are capable of operating both in air and liquid environments, but their dynamic behavior is not yet fully understood. With the view to developing a comprehensive dynamic model of trilayer actuators, this paper presents the experimentally obtained frequency response of the actuator displacement, as measured using a laser displacement sensor. A model of the resonant frequency is also presented and then comprehensively validated using actuators of various geometry and loading. This model can be used to: 1) estimate properties of the trilayer actuator from experimental measurements; 2) quantify and optimize an actuator's dynamic behavior as a function of geometry; and 3) facilitate the use of these actuators as a component in practical applications, such as force and motion control systems.

The performance of conducting polymers as actuators is determined by a complicated series of molecular processes that occur as a result of oxidation and reduction of the polymer. In particular, the amount of actuation strain generated for a given voltage stimulus is determined by the number and type of ions (and solvent) that enter and leave the polymer and changes in the mechanical properties (particularly elastic modulus) of the polymer. In this paper, we present the effects of cyclic voltammetry on the shear modulus and volume of polypyrrole in two different electrolytes: propylene carbonate and an ionic liquid. An electrochemical quartz crystal microbalance has been used to study simultaneous volume and modulus changes occurring during redox cycling of polypyrrole. The results demonstrate that the modulus generally increases due to oxidation of the polymer, although initial oxidation from the fully reduced state first produces a decrease in modulus followed by a larger increase. The modulus shift and volume changes were smaller in the ionic liquid electrolyte, probably because of the absence of solvent. Comparison of the results obtained in the two electrolytes suggest that interchain interactions dominate in the determination of modulus, so that modulus is higher in the oxidised state even when the polymer is swollen with counterions and solvent.

This paper presents a novel Synchronous Reluctance Tubular Linear Motor (SyncRTLM) for high precision motion control applications, such as machine tools. Finite Element Analysis (FEA) of the motor is conducted to simulate the motor performance and to analyse the magnetic fields in the motor. The design and development of the built motor is outlined, where a novel construction technique is described. The position, velocity and force performance is measured using a modified servo current control algorithm.

A 2.79 kJ prototype high transition temperature Superconducting Magnetic Energy Storage (SMES) device has been constructed. The coil for the prototype has been wound using High Temperature Superconducting (HTS) BSCCO-2223 tape. The refrigeration system is a gaseous helium cold head cryocooler used to maintain the SMES coil at a temperature of 30 K, improving the Ic characteristic of the coil by a factor of 4.7 compared to that at 77 K. The SMES device is capable of supplying a 3-phase load during power interruptions, and has been constructed during a program to develop a larger 20 kJ system aimed at industrial applications.

Conducting polymer actuators are emerging new actuators with many promising features suitable to some cutting edge applications ranging from biomedical devices to micro/nano manipulation systems. These features are highly influenced by their interrelated mechanical, electrical and chemical properties. This makes their behavior complex, and difficult to understand and predict. In order to make use of their full potentials, there is an increasing need to investigate into their actuation mechanism in order to provide enhanced degrees of understanding and predictability. With this in mind, it is the object of this paper to characterize and model the bending motion of the strip-type fourth generation polypyrrole polymer (PPy) actuators, which operate in a non-liquid medium, i.e. in the air. After deriving a mathematical model approximately accounting for mechanical, electrical, and chemical properties and geometric parameters of the actuator, the model has been experimentally verified for two actuators with the dimensions of (20 mm × 1 mm × 0.16 mm) and (10 mm × 1 mm × 0.21 mm). Theoretical and experimental results are presented to demonstrate that the model is effective enough to predict the displacement output of the strip type-PPy actuator all along the edge of the actuator as a function of the applied voltage.

The thrust force and torque produced during drilling contain important information related to the quality of the hole and the wear of the drill bit [1] . In this paper, the force and torque produced during drilling of carbon fibre using a ‘one shot’ drill bit is investigated. The signals in the time domain were divided into stages and common problems and defects associated with each stage discussed. It is also shown how tool wear and thickness of the workpiece affect the thrust force and torque throughout the drilling process. The findings of this paper are used to develop a mathematical model of the maximum thrust force and torque as described on Part II of this paper and are a valuable reference for future optimisation of drilling carbon composites with a ‘oneshot’ drill bit.

A mass boil-off measurement system has been used to accurately measure and characterize the ac loss of high-temperature superconductor (HTS) coils at frequencies between 50 and 200 Hz, and in applied ac fields of up to 0.04 T. The mass boil-off calorimeter incorporated a glass cryostat, a copper field coil, and two mass flow meters. The response of the gas flow to a step change in the applied magnetic field was found to have a time constant of about 600 s. Under suitable experimental conditions, it was possible to measure the ac losses of coils with an accuracy of /spl plusmn/0.3 W. The ac loss characteristics of an HTS pancake and two HTS solenoid coils are presented and the accuracy of mass flow calorimetry in liquid nitrogen is reported on.

The UOWShuttle * is a transport app that provides users with real-time bus location and predictive information from Automatic Vehicle Location and Automatic Vehicle Counting systems. The system anonymously tracks interactions within the transport network. This paper is a briefing document on the analytics of the Connected Mobility Digital Ecosystem project, of which the UOWShuttle forms an important part. It describes its overall system architecture and gives some of the statistics of some of the data collected to date.

This paper presents the results of magnetic loss tests on twisted and untwisted 37 multifilament tapes exposed to an alternating field at power frequencies. The losses are measured using calorimetric methods, which are capable of measuring losses in short tapes with the accuracy of several microwatts per centimeter of tape. The losses of the tape due to the longitudinal field of the untwisted tape agree well with theoretical calculations. The measurement results on the twisted filament tape, with 50-mm twist pitch, show that the losses are slightly lower than that in the untwisted tape. This is due to the reduction of the coupling losses between the filaments in the tape. It is also shown that in this presence of normal fields, the losses are about one order higher than for longitudinal fields because of the anisotropic properties of the superconducting parts and because of the existence of eddy current loss in the silver sheath.

One of the factors that concerns designers of machine tool servo systems is the mismatch that often exists between motor and load inertias. This paper provides a thorough mathematical analysis of the factors that interact to introduce stability problems in the presence of a high motor/load inertia mismatch. As a result of this analysis, the relative effects of inertia mismatch are discussed with reference to three commonly used servo system configurations. Methods of improving the system response using fast feedback control are also presented and experimentally verified.

In response to the increasing need for accurate motion control in mechatronic systems, this paper presents a method of increasing accuracy by reducing the detrimental effects of friction at very low speeds. A controller is presented which merges the technique of impulsive friction compensation with conventional linear control. The result is a controller that provides a much smoother output than a pure impulsive controller. This controller is analysed and then applied to an XY table machine tool test bed both in simulation and experimentation; results are presented showing improved performance.

High T/sub c/ superconducting (HTS) (Bi,Pb)/sub 2/Sr/sub 2/Ca/sub 2/ Cu/sub 3/O/sub 10+x/ Ag-clad wire has potential for practical applications in the form of a coil or a winding. This HTS wire has strong magnetic field-dependent and mechanical strain-dependent critical currents, consequently it has a severe problem of critical current degradation when it is used in the form of a coil. To design a winding with this conductor, the critical current degradation has to be identified with relation to the magnetic fields and the winding process. This is important to optimize an appropriate coil winding procedure using this HTS wire. A specially designed noninductive sample has been made with a (Bi,Pb)/sub 2/Sr/sub 2/Ca/sub 2/ Cu/sub 3/O/sub 10+x/ Ag-clad 27-filament wire, and the critical current degradation, which is caused by the magnetic field generated and the mechanical winding procedure used to form a coil, has been separated into two factors accordingly, and discussed in this paper.

Calorimetric methods for AC loss measurement for short superconducting wires have been investigated. The design, operation and results obtained from an experimental calorimeter are described. With these methods the total loss of a short superconducting sample exposed to a 50 Hz alternating field, both perpendicular and axial, have been measured with an accuracy of microwatts per centimeter. The sample is a 6 centimeter long (Bi,Pb)SrCaCuO-2223 silver-sheathed multifilamentary wire prepared by powder-in-tube techniques. The hysteresis part of the loss may be obtained by taking the eddy current component of the silver sheath from the total loss and by neglecting the coupling loss in the silver matrices. It is shown that the hysteresis losses are dominant in this frequency and its values correspond to the theoretical approximation.

Power engineering applications of HTSC tapes are often restricted by AC losses. Both direct and indirect methods have been used to measure AC losses in short samples. This paper reports on the development of a direct calorimetric method for AC loss measurement techniques for short tapes based on measurements of slight temperature increases of the sample. This method has the advantage of measuring total losses in the HTSC sample. The test rig discussed here consists of a calorimeter and associated data collection and analysis equipment. The calorimeter uses a sensitive temperature sensor and an ohmic heater, needed for the calibration of the system, surrounded by a thermal insulator. A platinum sensor is used which works at 77 K and may be used in a magnetic field. Sensors and controlled variables are interfaced to the controlling computer via IEEE interfaces. The power loss of a sample carrying alternating current has been measured with a high accuracy using this experimental set-up. The sample is a seven-filament BSCCO-2223 tape. In addition, the effect of the frequency and amplitude of the signal has easily and quickly been analysed.

This paper investigates the effect of using force impulses of different shapes to control machines with friction to achieve more precise positioning and better low speed velocity control. It is shown that by shaping the pulses the smallest available movement of the servomechanism, can be reduced, and so it's precision is enhanced. Several new force impulse controllers are investigated and their benefits, including improved precision and reduced machine vibration, described and experimentally verified. These benefits are achieved without mechanical modifications or any requirement for higher tolerance mechanical components, and the control scheme does not have to replace, but can be used to augment, a conventional PID controller.

In this paper, a monocular approach for generation of the depth maps of an object using computer vision is proposed. In this technique the object is immersed in a colour liquid and a single plan view is captured by a CCD camera. Due to the absorption property of the colour liquid, the intensity of the pixels of the captured image is modulated by the depths of the object along the camera axis. The depth maps are generated from the grey level image by consecutive segmentation of the image into binary images at various thresholds. The depth maps produced are parallel layers located in planes normal to the camera axis. The 2-D features of the contours contained in each layer are derived. The features of the layers are then combined to produce the 3-D model of the object. In order to employ the technique for object recognition a matching algorithm based on 3-D features of the object is developed. The technique is then verified by an experiment to recognize a set of objects. The approach can be used as long as the object is not damaged when placed in the colour liquid.

Recent efforts on manufacturing high temperature superconductors (HTS) have resulted in the production of coils which promise significant benefits for power engineering applications. Losses in two HTS pancake coils carrying AC current are measured and the results are analysed. It is shown that a thin pancake coil with relatively large diameter can be treated as straight tapes arranged in a face-to-face stack with the spacing between the tapes chosen to be relatively small. Losses in the tapes become those predicted using Norris's equation for an infinite slab.

Permanent magnet tubular linear motors have shown potential for use as direct drives in high precision machine tools. These devices are capable of replacing the traditional ball screw, given their ability to overcome or reduce many of the fundamental mechanical properties that limit the precision performance of the ball screw. However, given the desire from industry to increase speed and acceleration, the physical size of the motor must be minimised while maintaining high force output to increase bandwidth in high precision applications. This paper presents a new optimal tooth design to reduce the frame size of a slotted tubular linear motor. Several design aspects of the optimal tooth are considered and general solutions are derived. Data from finite element analysis and from an experimental machine is presented to validate the new designs.

While robots have proved to be ideal for the manipulation of single articles accurately in three dimensions, they are not as cost‐effective in industry for tasks which require simultaneous manipulations of several parts. In addition, changes in part geometry often necessitate the redesign of the gripping device, modification to the sensors and re‐examination of the grasping strategies to be used. These tasks are time consuming and add to the inefficiency of the approach.

Conducting polymers have a wide range of reversibly controllable properties, leading to a number of potentially useful devices for robotic applications, including actuators, sensors and batteries. Conducting polymers have the advantages of low weight, low cost, flexibility, small activation potentials (≪2V), biocompatibility and the ability to be manufactured using relatively straightforward techniques. Trilayer bending actuators which utilise the controllable change in volume of conducting polymers are potentially useful devices, but their speed and positioning ability must be improved. Significant research effort has been directed towards improving conducting polymer actuator performance through its chemistry, but the use of compensating control systems has had relatively little focus. This paper experimentally investigates three potential control systems for a trilayer bending actuator - feedforward gain control, feedforward inversion-based control and inversion-based PI control. It was found that the inversion-based PI control system provided the best performance, but the implementation utilised a large laser displacement sensor. To limit the requirement for such a large feedback sensor, a trilayer bending actuator with an integrated displacement sensor is proposed, exploiting the additional sensing capability of conducting polymers.

Wave variables guarantee stability for delayed teleoperation. They also introduce reflections which can prove very disorientating for an operator. This paper presents a new method for reducing the wave reflections when operating in nonlinear and unknown environments. Simulations and laboratory experiments shows the new method does not only guarantee stability but also reduce wave reflections allowing for improved velocity tracking and force feedback.

This paper presents a modified impulse controller to improve the steady state positioning of a SCARA robot having characteristics of high nonlinear friction. A hybrid control scheme consisting of a conventional PID part and an impulsive part is used as a basis to the modified controller. The impulsive part uses short width torque pulses to provide small impacts of force to overcome static fiction and move a robot manipulator towards its reference position. It has been shown that this controller can greatly improve a robot’s accuracy in position tracking. However, the system in attempting to reach steady state will inevitably enter into a small limit cycle whose amplitude of oscillation is related to the smallest usable impulse. It is shown in this paper that by modifying the impulse controller to adjust the width of successive pulses, the limit cycle can be shifted up or down in position so that the final steady state error can be even further reduced.

Conducting polymer trilayers are attractive for use in functional devices, given low actuation voltages, operation in air and potentially useful stresses and strains; however, their dynamic behavior must be understood from an engineering perspective before they can be effectively incorporated into a design. As a step towards the identification of the actuator dynamics, frequency response analysis has been performed to identify the magnitude and phase shift of displacement in response to a sinusoidal voltage input. The low damping of the trilayer operating in air and the use of a laser displacement sensor has allowed the frequency response to be continuously identified up to 100Hz, demonstrating a resonant peak at 80Hz for a 10mm long actuator. Two linear transfer function models have been fitted to the frequency response of the trilayer displacement (i) a 3 rd order model to represent the dynamics below 20Hz and (ii) a higher complexity 6 th order model to also include the resonant peak. In response to a random input signal, the 3 rd order model coarsely follows the experimental identified displacement, while the 6 th order model is able to fully simulate the real trilayer movement. Step responses have also been obtained for the 3 rd and 6 th order transfer functions, with both models capable of following the first 4 seconds of experimental displacement. The application of empirical transfer function models will facilitate accurate simulation and analysis of trilayer displacement, and will lead to the design of accurate positional control systems.

With power systems subject to an ever-increasing variety of load types, power quality remains of concern to both utilities and consumers alike. Some of the most concerning issues to users of sensitive equipment include voltage sags, supply interruption and current harmonics. This paper describes a comprehensive power quality device that can be designed to compensate for harmonic load currents and mains voltage variations in a system, as well as providing mains failure backup for smaller critical system loads. The design process described integrates high capacity high temperature superconducting magnetic energy storage (SMES) technology with active power filter (APF) switching technology to achieve the power quality device. The sizing of higher capacity SMES devices to suit critical load power requirements in a system and the complete design philosophy that maximises the usefulness of the whole power conditioning system are both discussed in the paper.

Application of conducting polymers has been growing widely in different fields such as batteries, solar cells, capacitors and actuators. Mechanical properties of conducting polymers like flexibility, high power to mass ratio and high active strain make them potentially applicable to robotic and automation industries. Obviously, a dynamic model of the actuation phenomenon in conducting polymers is needed to study its controllability and also to optimize the mechanical performance. De Rossi and colleagues suggest treating the mechanical behaviour of conducting polymers separately from the viscoelastic structural model and electrochemical actuation[1]. But it has been observed that the effects of electrochemical actuation and diffusion of ions on the viscoelastic coefficients cannot be neglected in some conducting polymer actuators, as shown in[1]. In this paper, we present the effects of cyclic voltammetry actuation on shear modulus of polypyrrole in propylene carbonate and EMI.TSFI as measured by an electrochemical Quartz Crystal Microbalance (eQCM). The QCM consists basically of an AT-cut piezoelectric quartz crystal disc with metallic electrode films deposited on its faces. One face is exposed to the active medium. A driver circuit applies an AC signal to the electrodes, causing the crystal to oscillate in a shear mode, at a given resonance frequency. QCM has been routinely used for the determination of mass changes. Measured resonance frequency shifts are converted into mass changes by the wellknown Sauerbrey's equation. In this paper, we correlate the admittance output of QCM to the real shear modulus of polypyrrole. Then the results of the correlation which contains mechanical data are presented during actuation using two different types of electrolyte.

This paper presents a study that was conducted at Monash University College Gippsland on the effect of a robot’s geometrical parameters on its dynamic performance. This research was undertaken in pursuit of an optimal dynamic performance of industrial robots. In this study an indicator was established to quantitatively measure the dynamic performance of a robot arm with respect to changes in the geometrical parameters. Since the dynamic behaviour of a robot’s arm is largely dependent on its inertia terms; the performance indicator was based on the Logarithmic function of the sensitivity of the inertia matrix’s eigenvalues to changes in a robot’s geometrical parameters. Also, since the inertia matrix is a function of the joint displacement; the performance was examined for a particular parameter over a range of joint angles. The sensitivity of the inertia matrix’s eigenvalues to changes in the non-zero parameters of an articulated robot (PUMA 560)l was examined in the course of this research. This paper presents cases related to the effect of changing the twist angles on the dynamic performance. The results of this study, which are illustrated by 3-dimensional surface plots, are discussed and analysed in this paper. All the simulation software was written in extended ANSI FORTRAN 77 and run on an HP 9000/550 computer using the UNIX operating system.

Conducting polymers (CP) are a promising area in the field of micro actuators, and have potential applications in micro robotics. Their properties are modelled as having an electro-chemical active component and a passive viscoelastic component. Methods exist to model the passive component as a configuration of springs and dashpots and the electro-active effects as a strain generator. Typically, the strain is assumed to be proportional to the charge transferred, and the two components are assumed to be independent. We show that there is a significant interaction between the two components for polypyrrole actuators, by observing the dynamic elastic modulus whilst varying the electric potential. The elastic modulus was measured in-situ by applying a high frequency rectangular isotonic stress input and recording the corresponding strain output. Two separate potential control inputs, vs. a reference electrode, were used. In the first experiment, a triangular voltage signal with variable frequency was applied to the PPy helix tube actuator and in the second experiment; a step voltage signal was applied to the actuator. The value of total real modulus was calculated during both experiments to evaluate the effect of actuation on the mechanical properties of PPy actuator. The performance of the mentioned method was confirmed by comparing its results to that of a sinusoidal stress input during a temperature ramp through the glass transition of polyethylene terephthalate (PET). We show that polypyrrole actuators show a complex change in stiffness with contractile state, which mimic skeletal muscle [1].

Conducting polymer actuators with favourable properties such as linearity, high power density and compliance are of increasing demand in micro applications. These materials generate forces over two times larger than produced by mammalian skeletal muscles. They operate to convert electro chemical energy to mechanical stress and strain. On the other hand, the application of conducting polymers is limited by the lack of a full description of the relation between four essential parameters: stress, strain, voltage and current. In this paper, polypyrrole helix tube micro actuator mechanical characteristics are investigated. The electrolyte is propylene carbonate and the dopant is TBA. PF 6 . The experiments are both in isotonic and isometric conditions and the input parameters are both electrical and mechanical. A dual mode force and length control and potentiostat / galvanostat are utilized for this purpose. Ultimately, the viscoelastic behaviour of the actuator is presented in this paper by a standard stress relaxation test. The effect of electrical stimulus on mechanical parameters is also explored by cyclic voltametry at different scan rates to obtain the best understanding of the actuation mechanism. The results demonstrate that the linear viscoelastic model, which performed well on conducting polymer film actuators, has to be modified to explain the mechanical behaviour of PPy helix tube fibre micro actuators. Secondly, the changes in mechanical properties of PPy need to be considered when modelling electromechanical behaviour.

In grinding operation, on-line process monitoring of dressing will ensure the quality of the grinding wheel and will achieve reproducible surface finish. The results of a study conducted to evaluate the feasibility and effectiveness of Acoustic Emission (AE) sensors in online monitoring of dressing is reported. Two aspects of the process are addressed. The first is the detection of undesired wheel conditions, such as an out-of-round wheel and wheel contour errors. The other is improving the surface finish of the ground workpiece by monitoring the dressing parameters. Experimental results indicate that AE signal can be effectively used to identify the faults that might occur during dressing. This in turn will result in better surface finish in the machined workpiece.

An adaptive control approach is presented for the tracking control of industrial robots. The approach utilizes the fact that a robot model can be described by equations that are linear in the system's unknown parameters. Taking uncertainties into account, the resulting controller has the property of robustness. The proof of stability and analytical results of the boundness of position tracking errors are given. By introducing filter operations in state measurements, the approach avoids the difficulty of measuring the accelerations of the robots' actuators. Simulation results are also presented. >

We study the problem of precise control of machines with static and Coulomb friction. An experimental X-Y testbed built for testing various control methods is described and initial experimental results are presented. A simple pole placement controller is designed for computer control of the X-Y testbed based on a generalised sampled-data hold. Robustness and stability of the controller are analysed. Simulation suggests that this controller is able to achieve precise position control even when parameters of both the dynamics of the machine and the characteristics of friction are unknown.

This paper outlines the University's involvement with industry based manufacturing projects, and how this has lead to the recent establishment of postgraduate and undergraduate mechatronics degrees. The nature of the industrial projects is described with examples of specific problems, test equipment and experimental rigs given. This work can be used to explain the reasons for the design of the mechatronics courses at Wollongong. A new teaching methodology particularly suitable for mechatronic's education is also discussed.

There are a number of performance limiting factors that concern the designers of machine tool servo systems. One such factor is the mismatch that often exists between motor and load inertias. This paper briefly discusses the results of a previous study on the factors that interact to introduce stability problems in the presence of a high motor-load inertia mismatch. The effects of such an inertia mismatch, on servo system performance, are then discussed and analysed using closed loop frequency responses. Various methods of improving the system response using fast feedback control are presented, including traditional PI control and modern methods incorporating full order state observation. The effectiveness of these control methods are then compared after comprehensive testing on a specially constructed experimental rig.

There are many robot controllers using advanced control theories proposed in the literature but very few papers that cover the practical ramifications of implementing advanced controllers. Many proposals concentrate on theoretical analysis and simulation of advanced controllers and neglect such practical engineering issues as their performance when current and torque limits or friction exist, their robustness in the presence of noise and modelling inaccuracies, or the trade-off between controller complexity and performance. This paper presents an investigation of a selection of controllers from major classes of control schemes available for robots and quantifies their relative performance and robustness when implemented on a practical robot subjected to a wide range of operating conditions while undertaking different tasks. The results of this work provide the practising engineer with information to make a much more informed choice of robot controller from the enormous array of candidate controllers available from the literature.

The measurement of transport critical current (Ic) for MgB2 wires and tapes has been investigated with two different techniques, the conventional four-probe arrangement with direct current (DC) power source, and a tailored triangle pulse at different rates of current change. The DC method has been widely used and practiced by various groups, but suffers from inevitable heating effects when high currents are used at low magnetic fields. The pulsed current method has no heating effects, but the critical current can depend on the rate of the current change (dI/dt) in the pulse. Our pulsed current measurements with varying dI/dt show that the same values of Ic are obtained as with the DC method, but without the artifacts of heating. Our method is particularly useful at low field regions which are often inaccessible by DC methods. We also performed a finite element method (FEM) analysis to obtain the time dependent heat distribution in MgB2 due to the electric potential produced at the current contacts to the superconducting sample and its gradient around the contacts. This gradient is defined as the current transfer length (CTL) of the samples and leads to Joule heating of the wire near the contacts. The FEM results provide further evidence of the limitation of the DC method in obtaining high transport critical current.

A layer structured V2O5·nH2O xerogel was synthesized via a simple green hydrothermal technique by dissolving commercial V2O5 powder in de-ionized water and hydrogen peroxide. Graphene–V2O5·nH2O xerogel composites were then prepared by mixing and filtration of as-prepared V2O5·nH2O xerogel and graphene in the desired ratio. The method is a cost effective and energy saving way to prepare nanostructured composites. Structure and morphology were investigated by X-ray diffraction, thermogravimetric analysis, field emission scanning electron microscopy, and transmission electron microscopy. Heat treatment at different temperatures could yield V2O5·nH2O xerogels with different amounts of crystal water, and the presence of graphene in the composites enhanced the thermal stability of V2O5·nH2O, in which the phase transformation moved towards higher temperature compared with the sample without graphene. The pristine V2O5·nH2O xerogel consisted of thin layers of ribbons with widths around 100 nm. In the composites, the V2O5·nH2O ribbons were located on the surface of the graphene sheets. Increasing the graphene content in the composites resulted in better cycling stability when the composites were tested as cathodes in different voltage ranges for lithium ion batteries. The initial and the 50th discharge capacities of the composite cathode with 17.8% graphene are 299 and 174 mAh g−1, respectively, when cycled between 1.5 and 4.0 V. The capacities decreased to 227 and 156 mAh g−1, respectively, when cycled between 2.0 and 4.0 V. The initial and the 50th discharge capacities of the composite with 39.6% graphene are 212 and 190 mAh g−1 in the voltage range of 1.5–4.0 V, and the capacities are 143 and 163 mAh g−1 when cycled between 2.0 and 4.0 V, respectively. The outstanding electrochemical performance could be attributed to the graphene induced unique structure and morphology.

The research presented in this study primarily proposes a simple but efficient method for the generation of depth maps for a certain class of industrial parts. The suggested technique is based on physical principle that the light is absorbed in a color liquid as it travels along an optical path. Consequently the grey level image of an object immersed in a color liquid contains information about the optical path of rays reflected from the object. In other words, the intensity of the grey level image of the object produced in this way is modulated by the depth of the object.© (1993) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.