A gray-box neural network-based model identification and fault estimation scheme for nonlinear dynamic systems

A novel gray-box neural network model (GBNNM), including multi-layer perception (MLP) neural network (NN) and integrators, is proposed for a model identification and fault estimation (MIFE) scheme. With the GBNNM, both the nonlinearity and dynamics of a class of nonlinear dynamic systems can be approximated. Unlike previous NN-based model identification methods, the GBNNM directly inherits system dynamics and separately models system nonlinearities. This model corresponds well with the object system and is easy to build. The GBNNM is embedded online as a normal model reference to obtain the quantitative residual between the object system output and the GBNNM output. This residual can accurately indicate the fault offset value, so it is suitable for differing fault severities. To further estimate the fault parameters (FPs), an improved extended state observer (ESO) using the same NNs (IESONN) from the GBNNM is proposed to avoid requiring the knowledge of ESO nonlinearity. Then, the proposed MIFE scheme is applied for reaction wheels (RW) in a satellite attitude control system (SACS). The scheme using the GBNNM is compared with other NNs in the same fault scenario, and several partial loss of effect (LOE) faults with different severities are considered to validate the effectiveness of the FP estimation and its superiority.

Toward an object-based semantic memory for long-term operation of mobile service robots

Throughout a lifetime of operation, a mobile service robot needs to acquire, store and update its knowledge of a working environment. This includes the ability to identify and track objects in different places, as well as using this information for interaction with humans. This paper introduces a long-term updating mechanism, inspired by the modal model of human memory, to enable a mobile robot to maintain its knowledge of a changing environment. The memory model is integrated with a hybrid map that represents the global topology and local geometry of the environment, as well as the respective 3D location of objects. We aim to enable the robot to use this knowledge to help humans by suggesting the most likely locations of specific objects in its map. An experiment using omni-directional vision demonstrates the ability to track the movements of several objects in a dynamic environment over an extended period of time.

A framework for model integration and holistic modelling of socio-technical systems

This paper presents a layered framework for the purposes of integrating different Socio-Technical Systems (STS) models and perspectives into a whole-of-systems model. Holistic modelling plays a critical role in the engineering of STS due to the interplay between social and technical elements within these systems and resulting emergent behaviour. The framework decomposes STS models into components, where each component is either a static object, dynamic object or behavioural object. Based on existing literature, a classification of the different elements that make up STS, whether it be a social, technical or a natural environment element, is developed; each object can in turn be classified according to the STS elements it represents. Using the proposed framework, it is possible to systematically decompose models to an extent such that points of interface can be identified and the contextual factors required in transforming the component of one model to interface into another is obtained. Using an airport inbound passenger facilitation process as a case study socio-technical system, three different models are analysed: a Business Process Modelling Notation (BPMN) model, Hybrid Queue-based Bayesian Network (HQBN) model and an Agent Based Model (ABM). It is found that the framework enables the modeller to identify non-trivial interface points such as between the spatial interactions of an ABM and the causal reasoning of a HQBN, and between the process activity representation of a BPMN and simulated behavioural performance in a HQBN. Such a framework is a necessary enabler in order to integrate different modelling approaches in understanding and managing STS.

Multi-level neural modelling : an application of object-oriented software engineering

Neu-Model, an ongoing project aimed at developing a neural simulation environment that is extremely computationally powerful and flexible, is described. It is shown that the use of good Software Engineering techniques in Neu-Model’s design and implementation is resulting in a high performance system that is powerful and flexible enough to allow rigorous exploration of brain function at a variety of conceptual levels.

Improving robot vision models for object detection through interaction

We propose a method for learning specific object representations that can be applied (and reused) in visual detection and identification tasks. A machine learning technique called Cartesian Genetic Programming (CGP) is used to create these models based on a series of images. Our research investigates how manipulation actions might allow for the development of better visual models and therefore better robot vision. This paper describes how visual object representations can be learned and improved by performing object manipulation actions, such as, poke, push and pick-up with a humanoid robot. The improvement can be measured and allows for the robot to select and perform the `right' action, i.e. the action with the best possible improvement of the detector.

From ImageNet to mining: Adapting visual object detection with minimal supervision

This paper presents visual detection and classification of light vehicles and personnel on a mine site.We capitalise on the rapid advances of ConvNet based object recognition but highlight that a naive black box approach results in a significant number of false positives. In particular, the lack of domain specific training data and the unique landscape in a mine site causes a high rate of errors. We exploit the abundance of background-only images to train a k-means classifier to complement the ConvNet. Furthermore, localisation of objects of interest and a reduction in computation is enabled through region proposals. Our system is tested on over 10km of real mine site data and we were able to detect both light vehicles and personnel. We show that the introduction of our background model can reduce the false positive rate by an order of magnitude.

Predicting spin of compact objects from their QPOs: A global QPO model

We establish a unified model to explain Quasi-Periodic-Oscillation (QPO) observed from black hole and neutron star systems globally. This is based on the accreting systems thought to be damped harmonic oscillators with higher order nonlinearity. The model explains multiple properties parallelly independent of the nature of the compact object. It describes QPOs successfully for several compact sources. Based on it, we predict the spin frequency of the neutron star Sco X-1 and the specific angular momentum of black holes GRO J1655-40, GRS 1915+105.

A model based factorization approach for dense 3D recovery from monocular video

Feature track matrix factorization based methods have been attractive solutions to the Structure-front-motion (Sfnl) problem. Group motion of the feature points is analyzed to get the 3D information. It is well known that the factorization formulations give rise to rank deficient system of equations. Even when enough constraints exist, the extracted models are sparse due the unavailability of pixel level tracks. Pixel level tracking of 3D surfaces is a difficult problem, particularly when the surface has very little texture as in a human face. Only sparsely located feature points can be tracked and tracking error arc inevitable along rotating lose texture surfaces. However, the 3D models of an object class lie in a subspace of the set of all possible 3D models. We propose a novel solution to the Structure-from-motion problem which utilizes the high-resolution 3D obtained from range scanner to compute a basis for this desired subspace. Adding subspace constraints during factorization also facilitates removal of tracking noise which causes distortions outside the subspace. We demonstrate the effectiveness of our formulation by extracting dense 3D structure of a human face and comparing it with a well known Structure-front-motion algorithm due to Brand.

On the Benefits and Problems of the Object-Oriented Paradigm including a Finnish Study

Several researchers are of the opinion that there are many benefits in using the object-oriented paradigm in information systems development. If the object-oriented paradigm is used, the development of information systems may, for example, be faster and more efficient. On the other hand, there are also several problems with the paradigm. For example, it is often considered complex, it is often difficult to make use of the reuse concept and it is still immature in some areas. Although there are several interesting features in the object-oriented paradigm, there is still little comprehensive knowledge of the benefits and problems associated with it. The objective of the following study was to investigate and to gain more understanding of the benefits and problems of the object-oriented paradigm. A review of previous studies was made and twelve benefits and twelve problems were established. These benefits and problems were then analysed, studied and discussed. Further a survey and some case studies were made in order to get some knowledge on what benefits and problems with the object-oriented paradigm Finnish software companies had experienced. One hundred and four companies answered the survey that was sent to all Finnish software companies with five or more employees. The case studies were made with six large Finnish software companies. The major finding was that Finnish software companies were exceptionally positive towards the object-oriented information systems development and had experienced very few of the proposed problems. Finally two models for further research were developed. The first model presents connections between benefits and the second between problems.

Model-theoretic approach to clustering

The paper deals with a model-theoretic approach to clustering. The approach can be used to generate cluster description based on knowledge alone. Such a process of generating descriptions would be extremely useful in clustering partially specified objects. A natural byproduct of the proposed approach is that missing values of attributes of an object can be estimated with ease in a meaningful fashion. An important feature of the approach is that noisy objects can be detected effectively, leading to the formation of natural groups. The proposed algorithm is applied to a library database consisting of a collection of books.

A model for heat transfer in a honey bee swarm

A swarm is a temporary structure formed when several thousand honey bees leave their hive and settle on some object such as the branch of a tree. They remain in this position until a suitable site for a new home is located by the scout bees. A continuum model based on heat conduction and heat generation is used to predict temperature profiles in swarms. Since internal convection is neglected, the model is applicable only at low values of the ambient temperature T-a. Guided by the experimental observations of Heinrich (1981a-c, J. Exp. Biol. 91, 25-55; Science 212, 565-566; Sci. Am. 244, 147-160), the analysis is carried out mainly for non-spherical swarms. The effective thermal conductivity is estimated using the data of Heinrich (1981a, J. Exp. Biol. 91, 25-55) for dead bees. For T-a = 5 and 9 degrees C, results based on a modified version of the heat generation function due to Southwick (1991, The Behaviour and Physiology of Bees, PP 28-47. C.A.B. International, London) are in reasonable agreement with measurements. Results obtained with the heat generation function of Myerscough (1993, J. Theor. Biol. 162, 381-393) are qualitatively similar to those obtained with Southwick's function, but the error is more in the former case. The results suggest that the bees near the periphery generate more heat than those near the core, in accord with the conjecture of Heinrich (1981c, Sci. Am. 244, 147-160). On the other hand, for T-a = 5 degrees C, the heat generation function of Omholt and Lonvik (1986, J. Theor. Biol. 120, 447-456) leads to a trivial steady state where the entire swarm is at the ambient temperature. Therefore an acceptable heat generation function must result in a steady state which is both non-trivial and stable with respect to small perturbations. Omholt and Lonvik's function satisfies the first requirement, but not the second. For T-a = 15 degrees C, there is a considerable difference between predicted and measured values, probably due to the neglect of internal convection in the model.

Construction of solid model from measured point data

This paper describes an algorithm for constructing the solid model (boundary representation) from pout data measured from the faces of the object. The poznt data is assumed to be clustered for each face. This algorithm does not require any compuiier model of the part to exist and does not require any topological infarmation about the part to be input by the user. The property that a convex solid can be constructed uniquely from geometric input alone is utilized in the current work. Any object can be represented a5 a combznatzon of convex solids. The proposed algorithm attempts to construct convex polyhedra from the given input. The polyhedra so obtained are then checked against the input data for containment and those polyhedra, that satisfy this check, are combined (using boolean union operation) to realise the solid model. Results of implementation are presented.

Investigation on the Sources of Reflection at the Junction of Tall Grounded Object and Lightning Channel

During lightning strike to a tall grounded object (TGO), reflections of current waves are known to occur at either ends of the TGO. These reflection modify the channel current and hence, the lightning electromagnetic fields. This study aims to identify the possible contributing factors to reflection at a TGO-channel junction for the current waves ascending on the TGO. Possible sources of reflection identified are corona sheath and discontinuity of resistance and radius. For analyzing the contribution of corona sheath and discontinuity of resistance at the junction, a macroscopic physical model for the return stroke developed in our earlier work is employed. NEC-2D is used for assessing the contribution of abrupt change in radii at a TGO-channel junction. The wire-cage model adopted for the same is validated using laboratory experiments. Detailed investigation revealed the following. The main contributor for reflection at a TGO-channel junction is the difference between TGO and channel core radii. Also, the discontinuity of resistance at a TGO-channel junction can be of some relevance only for the first microsecond regime. Further, corona sheath does not play any significant role in the reflection.

Experiments on three-dimensional wire-frame object recognition

Some experimental results on the recognition of three-dimensional wire-frame objects are presented. In order to overcome the limitations of a recent model, which employs radial basis functions-based neural networks, we have proposed a hybrid learning system for object recognition, featuring: an optimization strategy (simulated annealing) in order to avoid local minima of an energy functional; and an appropriate choice of centers of the units. Further, in an attempt to achieve improved generalization ability, and to reduce the time for training, we invoke the principle of self-organization which utilises an unsupervised learning algorithm.

Monte Carlo simulation of light transport in turbid medium with embedded object-spherical, cylindrical, ellipsoidal, or cuboidal objects embedded within multilayered tissues

Monte Carlo modeling of light transport in multilayered tissue (MCML) is modified to incorporate objects of various shapes (sphere, ellipsoid, cylinder, or cuboid) with a refractive-index mismatched boundary. These geometries would be useful for modeling lymph nodes, tumors, blood vessels, capillaries, bones, the head, and other body parts. Mesh-based Monte Carlo (MMC) has also been used to compare the results from the MCML with embedded objects (MCML-EO). Our simulation assumes a realistic tissue model and can also handle the transmission/reflection at the object-tissue boundary due to the mismatch of the refractive index. Simulation of MCML-EO takes a few seconds, whereas MMC takes nearly an hour for the same geometry and optical properties. Contour plots of fluence distribution from MCML-EO and MMC correlate well. This study assists one to decide on the tool to use for modeling light propagation in biological tissue with objects of regular shapes embedded in it. For irregular inhomogeneity in the model (tissue), MMC has to be used. If the embedded objects (inhomogeneity) are of regular geometry (shapes), then MCML-EO is a better option, as simulations like Raman scattering, fluorescent imaging, and optical coherence tomography are currently possible only with MCML. (C) 2014 Society of Photo-Optical Instrumentation Engineers (SPIE)