A statistical framework for natural feature representation

This paper presents a robust stochastic framework for the incorporation of visual observations into conventional estimation, data fusion, navigation and control algorithms. The representation combines Isomap, a non-linear dimensionality reduction algorithm, with expectation maximization, a statistical learning scheme. The joint probability distribution of this representation is computed offline based on existing training data. The training phase of the algorithm results in a nonlinear and non-Gaussian likelihood model of natural features conditioned on the underlying visual states. This generative model can be used online to instantiate likelihoods corresponding to observed visual features in real-time. The instantiated likelihoods are expressed as a Gaussian mixture model and are conveniently integrated within existing non-linear filtering algorithms. Example applications based on real visual data from heterogenous, unstructured environments demonstrate the versatility of the generative models.

Applying ISOMAP to the learning of hyperspectral image

In this paper, we present the application of a non-linear dimensionality reduction technique for the learning and probabilistic classification of hyperspectral image. Hyperspectral image spectroscopy is an emerging technique for geological investigations from airborne or orbital sensors. It gives much greater information content per pixel on the image than a normal colour image. This should greatly help with the autonomous identification of natural and manmade objects in unfamiliar terrains for robotic vehicles. However, the large information content of such data makes interpretation of hyperspectral images time-consuming and userintensive. We propose the use of Isomap, a non-linear manifold learning technique combined with Expectation Maximisation in graphical probabilistic models for learning and classification. Isomap is used to find the underlying manifold of the training data. This low dimensional representation of the hyperspectral data facilitates the learning of a Gaussian Mixture Model representation, whose joint probability distributions can be calculated offline. The learnt model is then applied to the hyperspectral image at runtime and data classification can be performed.

Predicting single-vehicle fatal crashes for two-lane rural highways in Southeastern United States

The rural two-lane highway in the southeastern United States is frequently associated with a disproportionate number of serious and fatal crashes and as such remains a focus of considerable safety research. The Georgia Department of Transportation spearheaded a regional fatal crash analysis to identify various safety performances of two-lane rural highways and to offer guidance for identifying suitable countermeasures with which to mitigate fatal crashes. The fatal crash data used in this study were compiled from Alabama, Georgia, Mississippi, and South Carolina. The database, developed for an earlier study, included 557 randomly selected fatal crashes from 1997 or 1998 or both (this varied by state). Each participating state identified the candidate crashes and performed physical or video site visits to construct crash databases with enhance site-specific information. Motivated by the hypothesis that single- and multiple-vehicle crashes arise from fundamentally different circumstances, the research team applied binary logit models to predict the probability that a fatal crash is a single-vehicle run-off-road fatal crash given roadway design characteristics, roadside environment features, and traffic conditions proximal to the crash site. A wide variety of factors appears to influence or be associated with single-vehicle fatal crashes. In a model transferability assessment, the authors determined that lane width, horizontal curvature, and ambient lighting are the only three significant variables that are consistent for single-vehicle run-off-road crashes for all study locations.

Computed success rates of various carrier phase integer estimation solutions and their comparison with statistical success rates

The success rate of carrier phase ambiguity resolution (AR) is the probability that the ambiguities are successfully fixed to their correct integer values. In existing works, an exact success rate formula for integer bootstrapping estimator has been used as a sharp lower bound for the integer least squares (ILS) success rate. Rigorous computation of success rate for the more general ILS solutions has been considered difficult, because of complexity of the ILS ambiguity pull-in region and computational load of the integration of the multivariate probability density function. Contributions of this work are twofold. First, the pull-in region mathematically expressed as the vertices of a polyhedron is represented by a multi-dimensional grid, at which the cumulative probability can be integrated with the multivariate normal cumulative density function (mvncdf) available in Matlab. The bivariate case is studied where the pull-region is usually defined as a hexagon and the probability is easily obtained using mvncdf at all the grid points within the convex polygon. Second, the paper compares the computed integer rounding and integer bootstrapping success rates, lower and upper bounds of the ILS success rates to the actual ILS AR success rates obtained from a 24 h GPS data set for a 21 km baseline. The results demonstrate that the upper bound probability of the ILS AR probability given in the existing literatures agrees with the actual ILS success rate well, although the success rate computed with integer bootstrapping method is a quite sharp approximation to the actual ILS success rate. The results also show that variations or uncertainty of the unit–weight variance estimates from epoch to epoch will affect the computed success rates from different methods significantly, thus deserving more attentions in order to obtain useful success probability predictions.

Continuous appearance-based trajectory SLAM

This paper describes a novel probabilistic approach to incorporating odometric information into appearance-based SLAM systems, without performing metric map construction or calculating relative feature geometry. The proposed system, dubbed Continuous Appearance-based Trajectory SLAM (CAT-SLAM), represents location as a probability distribution along a trajectory, and represents appearance continuously over the trajectory rather than at discrete locations. The distribution is evaluated using a Rao-Blackwellised particle filter, which weights particles based on local appearance and odometric similarity and explicitly models both the likelihood of revisiting previous locations and visiting new locations. A modified resampling scheme counters particle deprivation and allows loop closure updates to be performed in constant time regardless of map size. We compare the performance of CAT-SLAM to FAB-MAP (an appearance-only SLAM algorithm) in an outdoor environment, demonstrating a threefold increase in the number of correct loop closures detected by CAT-SLAM.

Reliability analysis and economic equipment replacement appraisal for substation and sub-transmission systems with explicit inclusion of non-repairable failures

The modern society has come to expect the electrical energy on demand, while many of the facilities in power systems are aging beyond repair and maintenance. The risk of failure is increasing with the aging equipments and can pose serious consequences for continuity of electricity supply. As the equipments used in high voltage power networks are very expensive, economically it may not be feasible to purchase and store spares in a warehouse for extended periods of time. On the other hand, there is normally a significant time before receiving equipment once it is ordered. This situation has created a considerable interest in the evaluation and application of probability methods for aging plant and provisions of spares in bulk supply networks, and can be of particular importance for substations. Quantitative adequacy assessment of substation and sub-transmission power systems is generally done using a contingency enumeration approach which includes the evaluation of contingencies, classification of the contingencies based on selected failure criteria. The problem is very complex because of the need to include detailed modelling and operation of substation and sub-transmission equipment using network flow evaluation and to consider multiple levels of component failures. In this thesis a new model associated with aging equipment is developed to combine the standard tools of random failures, as well as specific model for aging failures. This technique is applied in this thesis to include and examine the impact of aging equipments on system reliability of bulk supply loads and consumers in distribution network for defined range of planning years. The power system risk indices depend on many factors such as the actual physical network configuration and operation, aging conditions of the equipment, and the relevant constraints. The impact and importance of equipment reliability on power system risk indices in a network with aging facilities contains valuable information for utilities to better understand network performance and the weak links in the system. In this thesis, algorithms are developed to measure the contribution of individual equipment to the power system risk indices, as part of the novel risk analysis tool. A new cost worth approach was developed in this thesis that can make an early decision in planning for replacement activities concerning non-repairable aging components, in order to maintain a system reliability performance which economically is acceptable. The concepts, techniques and procedures developed in this thesis are illustrated numerically using published test systems. It is believed that the methods and approaches presented, substantially improve the accuracy of risk predictions by explicit consideration of the effect of equipment entering a period of increased risk of a non-repairable failure.

Hierarchical Bayesian models for estimating the extent of plant pest invasions

Plant biosecurity requires statistical tools to interpret field surveillance data in order to manage pest incursions that threaten crop production and trade. Ultimately, management decisions need to be based on the probability that an area is infested or free of a pest. Current informal approaches to delimiting pest extent rely upon expert ecological interpretation of presence / absence data over space and time. Hierarchical Bayesian models provide a cohesive statistical framework that can formally integrate the available information on both pest ecology and data. The overarching method involves constructing an observation model for the surveillance data, conditional on the hidden extent of the pest and uncertain detection sensitivity. The extent of the pest is then modelled as a dynamic invasion process that includes uncertainty in ecological parameters. Modelling approaches to assimilate this information are explored through case studies on spiralling whitefly, Aleurodicus dispersus and red banded mango caterpillar, Deanolis sublimbalis. Markov chain Monte Carlo simulation is used to estimate the probable extent of pests, given the observation and process model conditioned by surveillance data. Statistical methods, based on time-to-event models, are developed to apply hierarchical Bayesian models to early detection programs and to demonstrate area freedom from pests. The value of early detection surveillance programs is demonstrated through an application to interpret surveillance data for exotic plant pests with uncertain spread rates. The model suggests that typical early detection programs provide a moderate reduction in the probability of an area being infested but a dramatic reduction in the expected area of incursions at a given time. Estimates of spiralling whitefly extent are examined at local, district and state-wide scales. The local model estimates the rate of natural spread and the influence of host architecture, host suitability and inspector efficiency. These parameter estimates can support the development of robust surveillance programs. Hierarchical Bayesian models for the human-mediated spread of spiralling whitefly are developed for the colonisation of discrete cells connected by a modified gravity model. By estimating dispersal parameters, the model can be used to predict the extent of the pest over time. An extended model predicts the climate restricted distribution of the pest in Queensland. These novel human-mediated movement models are well suited to demonstrating area freedom at coarse spatio-temporal scales. At finer scales, and in the presence of ecological complexity, exploratory models are developed to investigate the capacity for surveillance information to estimate the extent of red banded mango caterpillar. It is apparent that excessive uncertainty about observation and ecological parameters can impose limits on inference at the scales required for effective management of response programs. The thesis contributes novel statistical approaches to estimating the extent of pests and develops applications to assist decision-making across a range of plant biosecurity surveillance activities. Hierarchical Bayesian modelling is demonstrated as both a useful analytical tool for estimating pest extent and a natural investigative paradigm for developing and focussing biosecurity programs.

Investigation into travel modes of TOD users : impacts of personal and transit characteristics

Transit Oriented Developments (TODs) are often designed to promote the use of sustainable modes of transport and reduce car usage. This paper investigates the effect of personal and transit characteristics on travel choices of TOD users. Binary logistic regression models were developed to determine the probability of choosing sustainable modes of transport including walking, cycling and public transport. Kelvin Grove Urban Village (KGUV) located in Brisbane, Australia was chosen as case study TOD. The modal splits for employees, students, shoppers and residents showed that 47% of employees, 84% of students, 71% of shoppers and 56% of residents used sustainable modes of transport.

Identifying differences in wet and dry road crashes using data mining

It is commonly accepted that wet roads have higher risk of crash than dry roads; however, providing evidence to support this assumption presents some difficulty. This paper presents a data mining case study in which predictive data mining is applied to model the skid resistance and crash relationship to search for discernable differences in the probability of wet and dry road segments having crashes based on skid resistance. The models identify an increased probability of wet road segments having crashes for mid-range skid resistance values.

Review of the under the limit drink driving rehabilitation program

The combination of alcohol and driving is a major health and economic burden to most communities in industrialised countries. The total cost of crashes for Australia in 1996 was estimated at approximately 15 billion dollars and the costs for fatal crashes were about 3 billion dollars (BTE, 2000). According to the Bureau of Infrastructure, Transport and Regional Development and Local Government (2009; BITRDLG) the overall cost of road fatality crashes for 2006 $3.87 billion, with a single fatal crash costing an estimated $2.67 million. A major contributing factor to crashes involving serious injury is alcohol intoxication while driving. It is a well documented fact that consumption of liquor impairs judgment of speed, distance and increases involvement in higher risk behaviours (Waller, Hansen, Stutts, & Popkin, 1986a; Waller et al., 1986b). Waller et al. (1986a; b) asserts that liquor impairs psychomotor function and therefore renders the driver impaired in a crisis situation. This impairment includes; vision (degraded), information processing (slowed), steering, and performing two tasks at once in congested traffic (Moskowitz & Burns, 1990). As BAC levels increase the risk of crashing and fatality increase exponentially (Department of Transport and Main Roads, 2009; DTMR). According to Compton et al. (2002) as cited in the Department of Transport and Main Roads (2009), crash risk based on probability, is five times higher when the BAC is 0.10 compared to a BAC of 0.00. The type of injury patterns sustained also tends to be more severe when liquor is involved, especially with injuries to the brain (Waller et al., 1986b). Single and Rohl (1997) reported that 30% of all fatal crashes in Australia where alcohol involvement was known were associated with Breadth Analysis Content (BAC) above the legal limit of 0.05gms/100ml. Alcohol related crashes therefore contributes to a third of the total cost of fatal crashes (i.e. $1 billion annually) and crashes where alcohol is involved are more likely to result in death or serious injury (ARRB Transport Research, 1999). It is a major concern that a drug capable of impairment such as is the most available and popular drug in Australia (Australian Institute of Health and Welfare, 2007; AIHW). According to the AIHW (2007) 89.9% of the approximately 25,000 Australians over the age of 14 surveyed had consumed at some point in time, and 82.9% had consumed liquor in the previous year. This study found that 12.1% of individuals admitted to driving a motor vehicle whilst intoxicated. In general males consumed more liquor in all age groups. In Queensland there were 21503 road crashes in 2001, involving 324 fatalities and the largest contributing factor was alcohol and or drugs (Road Traffic Report, 2001). 23438 road crashes in 2004, involving 289 fatalities and the largest contributing factor was alcohol and or drugs (DTMR, 2009). Although a number of measures such as random breath testing have been effective in reducing the road toll (Watson, Fraine & Mitchell, 1995) the recidivist drink driver remains a serious problem. These findings were later supported with research by Leal, King, and Lewis (2006). This Queensland study found that of the 24661 drink drivers intercepted in 2004, 3679 (14.9%) were recidivists with multiple drink driving convictions in the previous three years covered (Leal et al., 2006). The legal definition of the term “recidivist” is consistent with the Transport Operations (Road Use Management) Act (1995) and is assigned to individuals who have been charged with multiple drink driving offences in the previous five years. In Australia relatively little attention has been given to prevention programs that target high-risk repeat drink drivers. However, over the last ten years a rehabilitation program specifically designed to reduce recidivism among repeat drink drivers has been operating in Queensland. The program, formally known as the “Under the Limit” drink driving rehabilitation program (UTL) was designed and implemented by the research team at the Centre for Accident Research and Road Safety in Queensland with funding from the Federal Office of Road Safety and the Institute of Criminology (see Sheehan, Schonfeld & Davey, 1995). By 2009 over 8500 drink-drivering offenders had been referred to the program (Australian Institute of Crime, 2009).

A dynamic approach to forecast rental growth

Numerous econometric models have been proposed for forecasting property market performance, but limited success has been achieved in finding a reliable and consistent model to predict property market movements over a five to ten year timeframe. This research focuses on office rental growth forecasts and overviews many of the office rent models that have evolved over the past 20 years. A model by DiPasquale and Wheaton is selected for testing in the Brisbane, Australia office market. The adaptation of this study did not provide explanatory variables that could assist in developing a reliable, predictive model of office rental growth. In light of this result, the paper suggests a system dynamics framework that includes an econometric model based on historical data as well as user input guidance for the primary variables. The rent forecast outputs would be assessed having regard to market expectations and probability profiling undertaken for use in simulation exercises. The paper concludes with ideas for ongoing research.

Analysis of detection performance in spectrum sensing optimisation for long sensing periods

Spectrum sensing optimisation techniques maximise the efficiency of spectrum sensing while satisfying a number of constraints. Many optimisation models consider the possibility of the primary user changing activity state during the secondary user's transmission period. However, most ignore the possibility of activity change during the sensing period. The observed primary user signal during sensing can exhibit a duty cycle which has been shown to severely degrade detection performance. This paper shows that (a) the probability of state change during sensing cannot be neglected and (b) the true detection performance obtained when incorporating the duty cycle of the primary user signal can deviate significantly from the results expected with the assumption of no such duty cycle.

On the empirical eigenvalue distribution of slotted amplify-and-forward relaying protocol model

In this paper we consider the case of large cooperative communication systems where terminals use the protocol known as slotted amplify-and-forward protocol to aid the source in its transmission. Using the perturbation expansion methods of resolvents and large deviation techniques we obtain an expression for the Stieltjes transform of the asymptotic eigenvalue distribution of a sample covariance random matrix of the type HH† where H is the channel matrix of the transmission model for the transmission protocol we consider. We prove that the resulting expression is similar to the Stieltjes transform in its quadratic equation form for the Marcenko-Pastur distribution.

A Stochastic View of Optimal Regret through Minimax Duality

We study the regret of optimal strategies for online convex optimization games. Using von Neumann's minimax theorem, we show that the optimal regret in this adversarial setting is closely related to the behavior of the empirical minimization algorithm in a stochastic process setting: it is equal to the maximum, over joint distributions of the adversary's action sequence, of the difference between a sum of minimal expected losses and the minimal empirical loss. We show that the optimal regret has a natural geometric interpretation, since it can be viewed as the gap in Jensen's inequality for a concave functional--the minimizer over the player's actions of expected loss--defined on a set of probability distributions. We use this expression to obtain upper and lower bounds on the regret of an optimal strategy for a variety of online learning problems. Our method provides upper bounds without the need to construct a learning algorithm; the lower bounds provide explicit optimal strategies for the adversary. Peter L. Bartlett, Alexander Rakhlin

Making coups history

International support is capable of making the difference between the successful defense of democracy and its ignominious defeat. Indeed, the perceived probability of both support for democratically chosen leaders and opposition to their attackers can fundamentally shift the balance in the domestic struggle between them. Nevertheless, although changes to international law and international relations justify a greater international role in preventing and deterring coups and erosions, not all responsibility for protecting democracy should be assigned to the international community. Indeed, the first line of defense should be a democracy’s own domestic initiatives, with the main role of the international community being to support a domestic response to threats to democracy.