Increasingly it is data, vast amounts of data, that drives scientific discovery. At the heart of this so-called "fourth paradigm of science" is the rapid development of large scale statistical data fusion and machine learning methods. While these developments in "big data" methods are largely driven by commercial applications such as internet search or customer modelling, the opportunity for applying these to scientific discovery is huge. This talk will describe a number of applied machine learning projects addressing real-world inference problems in physical, life and social science areas. In particular, I will describe a major Science and Industry Endowment Fund (SIEF) project, in collaboration with the NICTA and Macquarie University, looking to apply machine learning techniques to discovery in the natural sciences. This talk will look at the key methods in machine learning that are being applied to the discovery process, especially in areas like geology, ecology and biological discovery.

The panel session 'Data Driven Science' discusses application and use of knowledge discovery, machine learning and data analytics in science disciplines; in natural, physical, medical and social science; from physics to geology, and from neuroscience to population health. Knowledge discovery methods are finding broad application in all areas of scientific endeavor, to explore experimental data, to discover new models, to propose new scientific theories and ideas. In addition, the availability of ever larger scientific data sets is driving a new data-driven paradigm for modeling of complex phenomena in physical, natural and social sciences.

This paper presents a distributed algorithm for performing joint localisation of a team of robots. The mobile robots have heterogeneous sensing capabilities, with some having high quality inertial and exteroceptive sensing, while others have only low quality sensing or none at all. By sharing information, a combined estimate of all robot poses is obtained. Inter-robot range-bearing measurements provide the mechanism for transferring pose information from well-localised vehicles to those less capable. In our proposed formulation, high frequency egocentric data (e.g., odometry, IMU, GPS) is fused locally on each platform. This is the distributed part of the algorithm. Inter-robot measurements, and accompanying state estimates, are communicated to a central server, which generates an optimal minimum mean-squared estimate of all robot poses. This server is easily duplicated for full redundant decentralisation. Communication and computation are efficient due to the sparseness properties of the information-form Gaussian representation. A team of three indoor mobile robots equipped with lasers, odometry and inertial sensing provides experimental verification of the algorithms effectiveness in combining location information.

Model uncertainty complicates most kinodynamic motion planning and control approaches due to their reliance on accurate forward prediction. If the model uncertainty is significant, a generated path or control strategy based on forward simulation of this model is potentially invalid and expensive to track (if possible). This paper explores the use of system identification/estimation to tune model parameters. Framed as an extension to rapidly exploring random tree (RRT) methods, it updates the model so that reachable actions added to the tree have more fidelity. This can be viewed as a mixture of a model predictive control (MPC) for local planning with an approximate-model global planner providing sub-goals and thus overcoming the limited lookahead caused by model uncertainty. The benefits of this approach are illustrated for a 3 DOF serial manipulator controlled by computed torque control operating under large external disturbances. In this case, the approach provides operation under intermittent feedback and disturbance observation. Tracking and actuator utilization are also improved over solutions found via conventional methods.

Obtaining a comprehensive model of large and complex terrain typically entails the use of both multiple sensory modalities and multiple data sets. This paper demonstrates the use of dependent Gaussian processes for data fusion in the context of large scale terrain modeling. Specifically, this paper derives and demonstrates the use of a non-stationary kernel (Neural Network) in this context. Experiments performed on multiple large scale (spanning about 5 sq km) 3D terrain data sets obtained from multiple sensory modalities (GPS surveys and laser scans) demonstrate the approach to data fusion and provide a preliminary demonstration of the superior modeling capability of Gaussian processes based on this kernel.

This paper develops an information (inverse-covariance) based method for efficient fusion and distributed estimation of large scale terrain. The output resembles a standard triangulated irregular network (TIN) terrain representation. However the proposed method uses distributed information fusion to estimate the elevations of the mesh vertices. This terrain mapping system is intended to use multiple scanning vehicles for online monitoring of the terrain for automated mining operations or other multi-vehicle field robotics systems. The method is based on a pre-specified regular finite-element mesh to define the set of estimated state variables. The method maintains a joint Gaussian distribution of the mesh vertices' elevations, in the information (inverse-covariance) form. The mesh elevations are estimated jointly given the irregular terrain observations, together with smoothness terms. The smoothness terms enable interpolation into unobserved regions as well as reducing noise. In the information form, the observations and smoothness terms are additive and the information matrix remains sparse in a fixed pattern, enabling constant-memory fusion of observations, efficient distribution among multiple sensing platforms and efficient solving for the estimates and uncertainty. Results show the reduction in data size for the fused observations compared to the raw observations, whilst still obtaining large scale high quality terrain maps. This paper focuses on a hierarchical distributed system in which each node estimates a subset of its parent's region, with the top-level node estimating a terrain map of the whole area. This paper compares two methods for the distributed communication from parent to child: An exact but expensive method, and an approximate fast method. Results compare the communication cost and resulting level of estimation approximation, showing that the marginalised information is expensive and the approximation is acceptable without it. This paper is applied to the estimation of large scale surface terrain from a distributed network of multiple sensors, such as 3D laser scanners, for automated terrain mapping for large scale mining applications.

This paper presents a new technique for sparsification of the information matrix of a multi-dimensional Gaussian distribution. We call this technique Conservative Sparsification (CS) and show that it produces estimates which are consistent with respect to an optimal filter. This technique was applied to the Simultaneous Localisation and Mapping (SLAM) problem, and compared with two existing sparsification approaches; the Sparse Extended Information Filter (SEIF) and the Data Discarding Sparse Extended Information Filter (DDSEIF). Simulation demonstrates that CS is a consistent approach and provides a tighter upper bound than existing conservative methods.

Trajectory generation and control of large equipment in open field environments involves systematically and robustly operating in uncertain and dynamic terrain. This paper presents an integrated motion planning and control system for tracked vehicles. Flexible path-end adjustments and adaptive look-ahead are introduced to a state lattice planning approach with waypoint control. For a given processing horizon, this increases search coverage and reduces planning error. This tramming approach has been successfully fielded on a 98-ton autonomous blast hole drill rig used in iron ore mining in Western Australia. The system has undergone extensive testing and is now integrated into a production environment. This work is a key element in a larger program aimed at developing a fully autonomous, remotely operated mine.

This paper presents the parallelization of grid-based recursive Bayesian estimation (RBE) using a graphics processing unit (GPU) for real-time control of autonomous vehicles. Although the grid-based method has been effectively used for autonomous search due to its ability to represent search space explicitly, heavy computational load has been a bottleneck for real-time application similarly to other non-Gaussian RBE techniques. The proposed RBE, which parallelizes grid-wise computations using GPU upon the analysis of mathematical operations, removes sequential processes and accelerates RBE significantly. Numerical examples have first demonstrated the validation of the proposed RBE and investigated its performance through parametric studies. The proposed RBE was then applied to the cooperative search by autonomous unmanned ground vehicles (UGVs), and its real-time capability has been demonstrated.

This paper describes a method of incorporating sensor and localisation uncertainty into contextual occupancy maps to provide for robust mapping. This paper builds on a recently proposed application of the Gaussian process (GP) to occupancy mapping. An extension of GPs is employed which incorporates uncertain inputs into the covariance function. In turn, this allows statistically consistent, multi-resolution maps to be constructed which exploit the spatial inference properties of GPs while correctly accounting for sensor and localisation errors. Experiments are described, with both synthetic and real data, which show the benefits of complete uncertainty modeling and how contextual occupancy maps may be constructed by fusing data from different sensors on different robots in a common probabilistic representation.

The need for efficient monitoring of spatio-temporal dynamics in large environmental surveillance applications motivates the use of robotic sensors to achieve sufficient spatial and temporal coverage. A common approach in machine learning to model spatial dynamics is to use the nonparametric Bayesian framework known as Gaussian Processes (GPs) (c.f., [1]) which are fully specified by a mean and a covariance function. However, defining suitable covariance functions that are able to appropriately model complex space-time dependencies in the environment is a challenging task. In this paper, we develop a generic approach for constructing several classes of covariance functions for spatio-temporal GP modeling. The GP models are then extended to perform efficient path planning in continuous space while maximizing the information gain. Extensive empirical evaluation for the different classes of covariance functions using real world sensing datasets is discussed, including experiments on a tethered robotic system - Networked Info Mechanical System (NIMS).

This paper proposes a novel method for computing robot motion uncertainty from ranging sensor data. The method utilises the recently proposed CRF-Matching procedure which matches laser scans based on shape descriptors. Motion estimates are computed in a probabilistic framework by performing inference on a probabilistic graphical model. We propose an efficient sampling procedure for obtaining probable association hypothesis of the probabilistic graphical model. The hypothesis are used to generate estimates on the uncertainty of translational and rotational movements of the robot. Experiments demonstrate the benefits of the approach on simulated data sets and on laser scans from an urban environment. The approach is also combined with the well-established delayed-state information filter for a large-scale outdoor simultaneous localisation and mapping task.

This paper presents a novel approach to data fusion for stochastic processes that model spatial data. It addresses the problem of data fusion in the context of large scale terrain modeling for a mobile robot. Building a model of large scale and complex terrain that can adequately handle uncertainty and incompleteness in a statistically sound manner is a very challenging problem. To obtain a comprehensive model of such terrain, typically, multiple sensory modalities as well as multiple data sets are required. This work uses Gaussian processes to model large scale terrain. The model naturally provides a multi-resolution representation of space, incorporates and handles uncertainties appropriately and copes with incompleteness of sensory information. Gaussian process regression techniques are applied to estimate and interpolate (to fill gaps in unknown areas) elevation information across the field. In this work, the GP modeling approach is extended to fuse multiple, multi-modal data sets to obtain a best estimate of the elevation given the individual data sets. The individual data sets are treated as different noisy samples of the same underlying terrain. Experiments performed on sparse GPS based survey data and dense laser scanner data taken at mine-sites are reported.

A new control strategy based on feedback motion planning is presented for solving nonlinear control problems in constrained environments. The algorithm explores the state-space using a bi-directional rapidly exploring random tree (biRRT) in order to find a feasible trajectory between an initial and goal state. By incrementally scheduling LQR controllers, it attempts to connect states so as to link the two trees. These attempts are evaluated by verifying that the connected state is inside the controllable area of an infinite time horizon controller at the goal. This allows for a rapid delineation of equivalent neighborhoods in the state-space. As a result, random exploration is terminated as soon as a feasible solution is made possible by feedback means, avoiding oversampling and partially introducing optimal actions at the neighborhood of the connection. The algorithm is demonstrated and compared against a biRRT using single-link pendulum and cart-pole swing-up tasks amongst obstacles, the latter showing a nearly order of magnitude more efficient search.

Terrain modeling remains a challenging yet key component for the deployment of ground robots to the field. The difficulty arrives from the variability of terrain shapes, sparseness of the data, and high degree uncertainty often encountered in large, unstructured environments. This paper presents significant advances to data fusion for stochastic processes modeling spatial data, demonstrated in large-scale terrain modeling tasks. We explore dependent Gaussian processes to provide a multi-resolution representation of space and associated uncertainties, while integrating sensors from different modalities. Experiments performed on multiple multi-modal datasets (3D laser scans and GPS) demonstrate the approach for terrains of about 5 km2.

In this paper we introduce a new statistical modeling technique for building occupancy maps. The problem of mapping is addressed as a classification task where the robot's environment is classified into regions of occupancy and unoccupancy. Our model provides both a continuous representation of the robot's surroundings and an associated predictive variance. This is obtained by employing a Gaussian process as a non-parametric Bayesian learning technique to exploit the fact that real-world environments inherently possess structure. This structure introduces a correlation between points on the map which is not accounted for by many common mapping techniques such as occupancy grids. Using a trained neural network covariance function to model the highly non-stationary datasets, it is possible to generate accurate representations of large environments at resolutions which suit the desired applications while also providing inferences into occluded regions, between beams, and beyond the range of the sensor, even with relatively few sensor readings. We demonstrate the benefits of our approach in a simulated data set with known ground-truth, and in an outdoor urban environment covering an area of 120,000 m2.

This paper addresses the problem of large scale terrain modeling for a mobile robot. Building a model of large scale terrain data that can adequately handle uncertainty and incompleteness in a statistically sound way is a very challenging problem. This work proposes the use of Gaussian processes as models of large scale terrain. The proposed model naturally provides a multi-resolution representation of space, incorporates and handles uncertainties aptly and copes with incompleteness of sensory information. Gaussian process regression techniques are applied to estimate and interpolate (to fill gaps in unknown areas) elevation information across the field. The estimates obtained are the best linear unbiased estimates for the data under consideration. A single non-stationary (neural network) Gaussian process is shown to be powerful enough to model large and complex terrain, handling issues relating to discontinuous data effectively. A local approximation methodology based on KD-trees is also proposed in order to ensure local smoothness and yet preserve the characteristic features of rich and complex terrain data. The use of the local approximation technique based on KD-trees further addresses concerns relating to the scalability of the proposed approach for large data sets. Experiments performed on sparse GPS based survey data as well as dense laser scanner data taken at different mine-sites are reported in support of these claims.

The impact of model-based design on the software engineering community is impressive, and recent research in model transformations, and elegant behavioral specifications of systems has the potential to revolutionize the way in which systems are designed. Such techniques aim to raise the level of abstraction at which systems are specified, to remove the burden of producing application-specific programs with general-purpose programming. For complex real-time systems, however, the impact of model-driven approaches is not nearly so widespread. In this paper, we present a perspective of model-based design researchers who joined with software experts in robotics to enter the DARPA Urban Challenge, and to what extent model-based design techniques were used. Further, we speculate on why, according to our experience and the testimonies of many teams, the full promises of model-based design were not widely realized for the competition. Finally, we present some thoughts for the future of model-based design in complex systems such as these, and what advancements in modeling are needed to motivate small-scale projects to use model-based design in these domains.

This paper addresses the long-standing problem of feature representation in the natural world for autonomous navigation systems. The proposed representation combines Isomap, which is a nonlinear manifold learning algorithm, with expectation maximization, which is a statistical learning scheme. The representation is computed off-line and results in a compact, nonlinear, non-Gaussian sensor likelihood model. This model can be easily integrated into estimation algorithms for navigation and tracking. The compactness of the model makes it especially attractive for deployment in decentralized sensor networks. Real sensory data from unstructured terrestrial and underwater environments are used to demonstrate the versatility of the computed likelihood model. The experimental results show that this approach can provide consistent models of natural environments to facilitate complex visual tracking and data-association problems.

This paper presents a new model and an extension to an existing algorithm for camera calibration. The main goal of the proposed approach is to calibrate miniature, low-cost, wide-angle fisheye lenses. The model has been verified with a calibration implementation and was tested on real data. Experiments show that the proposed model improves the accuracy compared to the original algorithm. Results show that the extension not only performs well with fisheye lenses but also with omnidirectional catadioptric lenses as well as other less distorted dioptric lenses. 1 Motivation and Related Work Wide-angle, hemispherical or omnidirectional camera systems have become more popular in the last few years. Especially in robotic applications, wide-angle sensors are favourable for perception and navigation problems. A precise calibration is needed in order to infer accurate bearing information for the 2D pixel information. Three different types of calibration methods can be distinguished. The most common

In this paper, simultaneous localisation and mapping (SLAM) is combined with landmark recognition to close large loops in unstructured, outdoor environments. Camera and laser information are fused to recognise and create appearance models for landmarks. The representation is obtained through a non-linear probabilistic regression model encoding a neighbourhood preserving dimensionality reduction. A new data association algorithm is proposed where landmarks are associated based on both position and appearance. The resulting system is more robust and able to recover from possible misassociations. Experiments demonstrate the benefits of this approach in challenging problems involving mapping with large loop closings in irregular terrain, and with dynamic objects.

This paper presents the element-based method, which can be used for recursive Bayesian estimation (RBE) in robotic operations such as search and tracking involving moving targets. The use of shape functions to define a set of irregularly shaped elements allows the target PDF to be continuously, and thus accurately, represented over the target space. A comparison with the grid-based method first shows that the element-based method requires less than 10% of the number of nodes to achieve the same accuracy. The application of the element-based method to marine search-and-rescue (SAR) scenarios then demonstrates its ability for effective SAR whilst maintaining collected information.

There is a growing excitement about the potential application of large scale sensor networks in diverse applications such as precision agriculture, geophysical and environment monitoring, remote health care, and security. Rapid progress in sensing hardware, communications and low-power computing has resulted in a profusion of commercially available sensor nodes. The big challenge now is to develop effective methods for the automatic fusion and interpretation of the information generated by large-scale sensor networks. The success of future applications is predicated on finding solutions to this data fusion challenge. This talk will focus on probabilistic models and Bayesian data fusion methods appropriate to describing and solving sensor network data fusion problems. Over a number of years, we have used such methods to develop decentralised algorithms for point-target estimation in sensor networks. More recently, other researchers have developed distributed Bayesian algorithms for sensor networks to estimate field properties, such as temperature or humidity over an area. The prospect of developing holistic Bayesian methods for fusion in sensor networks now looks to be a real and exciting possibility. This talk will describe these potential developments. Notions of divergence and information appear naturally in probabilistic fusion algorithms. In turn these provide a handle on two other sensor network data fusion problems; how to model and manage sensor node performance and how to learn patterns in or interpret sensor network information. These ideas will also be developed in this talk. Interspersed amongst these developments we will describe some real sensor network applications, some solved, some currently being addressed, and some that remain as a challenge to network data fusion methods.

This paper develops a general theory of validation gating for non-linear non-Gaussian models. Validation gates are used in target tracking to cull very unlikely measurement-to-track associations, before remaining association ambiguities are handled by a more comprehensive (and expensive) data association scheme. The essential property of a gate is to accept a high percentage of correct associations, thus maximising track accuracy, but provide a sufficiently tight bound to minimise the number of ambiguous associations. For linear Gaussian systems, the ellipsoidal validation gate is standard, and possesses the statistical property whereby a given threshold will accept a certain percentage of true associations. This property does not hold for non-linear non-Gaussian models. As a system departs from linear-Gaussian, the ellipsoid gate tends to reject a higher than expected proportion of correct associations and permit an excess of false ones. In this paper, the concept of the ellipsoidal gate is extended to permit correct statistics for the non-linear non-Gaussian case. The new gate is demonstrated by a bearing-only tracking example

This paper describes the general multi-platform reconnaissance or information gathering task. It is shown that the general problem must be solved in a centralised manner using information supplied by every platform. However, if for a specific problem the objective function is partially separable it is shown the optimal control problem can be solved in a decentralised fashion for an arbitrary sized group. This is demonstrated for scenarios with Gaussian probabilities and an indoor mapping problem

In this paper, we combined the application of a non-linear dimensionality reduction technique, isomap, with expectation maximisation in graphical probabilistic models for learning and classification of hyperspectral image. Hyperspectral image spectroscopy 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 man-made objects in unfamiliar terrains for robotic vehicles. However, the large information content of such data makes interpretation of hyperspectral images time-consuming and user-intensive. 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 mixture of linear models representation similar to a mixture of factor analysers, the joint probability distributions of the model can be calculated offline. The learnt model is then applied to the hyperspectral image at run-time and data classification can be performed. We also show the comparison with results from standard techniques

This paper presents a unified framework for online visual perception. The twin problems of visual feature extraction and representation are explicitly addressed. Simple paradigms for supervised and unsupervised feature extraction are presented to represent the extremes in visual perception system design. Visual feature representation is addressed through a combination of isomap, a non-linear dimensionality reduction algorithm, and expectation maximization (EM), a statistical learning scheme. A joint probability distribution of this representation is computed offline based on existing training data. Example applications based on real visual data from heterogenous, unstructured environments demonstrate the versatility of the generative models

This paper presents a coordinated control technique that allows heterogeneous vehicles to autonomously search for and track multiple targets using recursive Bayesian filtering. A unified sensor model and a unified objective function are proposed to enable search-and-tracking (SAT) within the recursive Bayesian filter framework. The strength of the proposed technique is that a vehicle can switch its task mode between search and tracking while maintaining and using information collected during the operation. Numerical results first show the effectiveness of the proposed technique when a found target becomes lost and must be searched for again. The proposed technique was then applied to a practical marine search-and-rescue (SAR) scenario where heterogeneous vehicles coordinated to search for and track multiple targets. The result demonstrates the applicability of the technique to real search world scenarios

This paper presents a decentralised particle filtering algorithm that enables multiple vehicles to jointly track 3D features under limited communication bandwidth. This algorithm, applied within a decentralised data fusion (DDF) framework, deals with correlated estimation errors due to common past information when fusing two discrete particle sets. Our solution is to transform the particles into Gaussian mixture models (GMMs) for communication and fusion. Not only can decentralised fusion be approximated by GMMs, but this representation also provides summaries of the particle set. Less bandwidth per communication step is required to communicate a GMM than the particle set itself hence conversion to GMMs for communication is an advantage. Real airborne data is used to demonstrate the accuracy of our decentralised particle filtering algorithm for airborne tracking and mapping

This paper presents an algorithm for recognition and segmentation of natural features in unstructured environments. By providing a Bayesian solution for the density estimation problem, the algorithm needs significantly less training data than conventional techniques and is applicable to different environments. The algorithm is based on colour and wavelet convolution of image patches to model the information contained in natural features. Dimensionality reduction techniques are applied to map data points to a lower dimensional space where Bayesian density estimation is computed. Experiments were performed in underwater, aerial and terrestrial domains demonstrating the accuracy and generalisation properties of the algorithm for recognition and segmentation. Comparisons with conventional density estimation techniques are provided to illustrate the benefits of the new approach

A revised car-like vehicle model with forwards and backwards motion, an upper bound on the derivative of curvature and a pause for steering at cusp points is presented. This model has improved performance near cusp points and is shown to converge toward the optimal "Reeds and Shepp" paths under two conditions: 1) when the upper bound of the curvature derivative bound tends to infinity and 2) when the pause at the cusp is sufficient for the vehicle to completely re-orientate its steering before departing the cusp point. This motion model was applied to the general problem of turning a vehicle around between two parallel lines. This required the development of a specialised path planner and a method to search for near-time optimal solutions.

Summary form only given. There is a growing excitement about the potential application of large scale sensor networks in diverse applications such as precision agriculture, geophysical and environment monitoring, remote health care, and security. Rapid progress in sensing hardware, communications and low-power computing has resulted in a profusion of commercially available sensor nodes. The big challenge now is to develop effective methods for the automatic fusion and interpretation of the information generated by large-scale sensor networks. The success of future applications is predicated on finding solutions to this data fusion challenge. This talk will focus on probabilistic models and Bayesian data fusion methods appropriate to describing and solving sensor network data fusion problems. Over a number of years, we have used such methods to develop decentralised algorithms for point-target estimation in sensor networks. More recently, other researchers have developed distributed Bayesian algorithms for sensor networks to estimate field properties, such as temperature or humidity over an area. The prospect of developing holistic Bayesian methods for fusion in sensor networks now looks to be a real and exciting possibility. This talk will describe these potential developments. Notions of divergence and information appear naturally in probabilistic fusion algorithms. In turn these provide a handle on two other sensor network data fusion problems; how to model and manage sensor node performance and how to learn patterns in or interpret sensor network information. These ideas will also be developed in this talk. Interspersed amongst these developments we will describe some real sensor network applications, some solved, some currently being addressed, and some that remain as a challenge to network data fusion methods.

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.

This paper addresses key issues regarding the feasibility of providing micro unmanned air vehicles (micro-UAVs) with a miniature hemispherical eye using the latest CMOS sensor technology. Key specifications of the visual system of such visually guided drones are low power consumption, adaptive resolution and sensitivity, packaging and manufacturing complexity, ultra-wide field-of-view (FOV), lightness and small overall size. We describe a compact pinhole imaging system that yields near hemispherical FOV, and we investigate its theoretical performance limits to ego- and visual acuity.

This paper deals with the problem of coordinating a team of mobile sensor platforms searching for a single mobile non-evading target. It follows the general Bayesian active sensor network approach introduced in [2] where each decision maker plans locally based on an equivalent representation of the target state probability density function (PDF). This paper focuses on the prediction stage of the decentralized Bayesian filter. It looks at how different types of realistic external constraints may affect the target motion and how they may be taken into account in the process model. Two general classes of constraints are identified soft and hard. A few constraint examples from each class are given to illustrate their impact on the evolution of the target state PDF. Multiple constraints of various types can be combined to increase the accuracy of the predicted PDF estimate, thus affecting the individual trajectories of the search platforms. The effectiveness of the framework is demonstrated for a team of airborne search vehicles looking for a drifting target lost in a storm at sea.

This paper presents the real-time control of cooperative unmanned air vehicles (UAV) that dynamically engage multiple targets in a time-optimal manner. Techniques to dynamically allocate vehicles to targets and to subsequently find the time-optimal control actions are proposed. The decentralization of the proposed control strategy is further presented such that the vehicles can be controlled in real-time without significant time delay. The proposed strategy is men applied to various practical battlefield problems, and numerical results show the efficiency of the proposed strategy.

Perception by autonomous systems, in unstructured dynamic worlds, is one of the significant research challenges in the development of effective intelligent systems. Nonlinear dimensionality reduction techniques have been extensively utilized within the artificial intelligence community to devise compact representations of high dimensional data. These techniques display great promise in yielding low dimensional, meaningful representations of an unstructured environment in real time from raw sensory information. Two such techniques, the kernel principal component analysis method and locally linear embedding (LLE) are evaluated herein, with respect to their ability to generate compact and physically reasonable embeddings of an unstructured environment. The LLE technique shows great potential in the computation of low dimensional and perceptually meaningful embeddings of natural environments.

This paper presents a control strategy for the pursuer in the pursuit-evasion game problem when the evader behaves intelligently. The pursuer in the proposed technique does not try to react to the evader's behavior instantaneously. The proposed technique therefore does not yield instantaneous optimality but capture the evader in a time-efficient and robust fashion even when the evader is intelligent. The proposed technique was applied to two numerical examples and the results were compared to those by the conventional motion tracking algorithms. The results and comparison show that the proposed technique could capture the evader faster than the conventional motion tracking algorithms in both the examples.

The paper presents a decentralized approach to the solution of the distributed information gathering problem. The main design objectives are scalability with the number of network components, maximum flexibility in implementation and deployment, and robustness to component and communication failure. The design approach emphasizes interactions between components rather than the definition of the components themselves. The architecture specifies a small set of interfaces sufficient to implement a wide range of information gathering systems. The results of two deployment scenarios on an indoor sensor network are presented.

This paper addresses the problem of coordinating a team of multiple heterogeneous sensing platforms searching for a single lost target. In this approach, the utility of a control sequence is a function of the probability density function (PDF) of the target state. Each decision maker builds an equivalent estimate of this PDF by communicating and fusing the information from the other sensor nodes. Coupled utilities incite the agents to collaborate and to agree on the next best set of actions. Decentralized cooperative planning is achieved via anonymous negotiation based on communication of expected observed information. Simulation results demonstrate the efficiency of the cooperative trajectories for a team of autonomous airborne search vehicles.

This paper presents a formulation and solution of the time-optimal control of multiple cooperative robot vehicles. In particular, a group of robot vehicles reaching a terminal state in absolute and/or relative formations in minimum time is addressed. A canonical formulation of the problem is first derived and a numerical technique, which can effectively solve this class of problems, is then proposed. Numerical results are then presented to demonstrate the efficacy of the proposed formulation and method of solution. The techniques described offer a practical solution to the problem of building and controlling formations of cooperative autonomous vehicles in real-time.

This paper describes an information-theoretic approach to distributed and coordinated control of a multi-robot sensor system. The approach is based on techniques long established for the related problem of decentralised data fusion (DDF). The DDF architecture uses information measures to communicate state estimates in a network of sensors. For coordinated control of robot sensors, the control objective becomes maximisation of these information measures. This yields platform trajectories, which maximise the total information, gained by the system. This approach inherits the many benefits of the DDF method including scalability, robustness to sub-system failure and addition, and interoperability among heterogeneous systems. The approach is applied to a practical bearings-only multi-feature localisation problem.

This paper describes a decentralized Bayesian approach to coordinating multiple autonomous sensor platforms searching for a single non-evading target. In this architecture, each decision maker builds an equivalent representation of the target state PDF through a Bayesian DDF network enabling him or her to coordinate their actions without exchanging any information about their plans. The advantage of the approach is that a high degree of scalability and real time adaptability can be achieved. The effectiveness of the approach is demonstrated in different scenarios by implementing the framework for a team of airborne search vehicles looking for a stationary, and a drifting target lost at sea.

This paper presents a solution to the real-time control of cooperative unmanned air vehicles (UAVs) that engage multiple targets in a time-optimal manner. Techniques to dynamically allocate vehicles to targets and to find the time-optimal control actions of vehicles are proposed. The effectiveness of the time-optimal control technique is first demonstrated through numerical examples. The proposed strategy is then applied to a practical battlefield problem where ten vehicles are required to engage four targets, and numerical results show the efficiency of the proposed strategy.

This work is concerned with mapping of indoor environment by a team of robots who share the information they gather to build a global representation of the environment in a consistent, efficient, and scalable way. The well known approach of Decentralized Data Fusion is applied to a widely used representation of Certainty Grid maps. The result is a scalable and intuitive algorithm for combining observations from multiple heterogeneous sensing platforms with built-in processing and communication capabilities. The platforms are assumed to have access to external localization. Results of realistic simulations are presented.

This paper examines the role played by vehicle models and their impact on the performance of sensor-based navigation systems for autonomous land vehicles. In a navigation system, information from internal and external vehicle sensors is combined to estimate the motion of the vehicle. However, while the issue of sensing and effects of sensor accuracy have been widely studied, there are few results or insights into the complementary role played by the vehicle model. This paper has two main contributions: a theoretical analysis of the role of the vehicle model in navigation system performance, and an empirical study of three models of increasing complexity, used in a navigation system for a conventional road vehicle. The theoretical analysis focuses on understanding the effect of estimation errors caused by approximations to the "true" vehicle model. It shows that while substantial performance improvements can be obtained from better vehicle modeling, there is, in general, no definitive "best" model for such complex nonlinear estimation problems. The empirical study shows that an appropriate choice of a higher order model can lead to significant improvements in the performance of the navigation system. However, the highest order model suffers from problems related to the observability of some of its parameters. We show how this problem can be overcome through the imposition of weak constraints.

This paper presents a method for the fusion of millimetre wave radar and nightvision sensors to generate an information-rich representation of the environment. The data from each of the sensors is divided into unstructured spatial objects according to the data available to that sensor. A hyperdimensional representation then constructed from these objects, with the observable characteristics providing the axes. This representation can then be provided to target extraction, classification and tracking algorithms to achieve advanced machine sensing and perception tasks. This differs significantly from the traditional approaches to this problem in which target identification is completed for each individual sensor and these estimates are then combined. The results of initial field trials are used to demonstrate the feasibility of this approach.

The paper describes a terrain-aided navigation system that employs points of maximum curvature extracted from laser scan data as primary landmarks. A scale space method is used to extract points of maximum curvature from laser range scans of unmodified outdoor environments. This information is then fused with odometric information to provide localization information for an outdoor vehicle. The method described is invariant to the size and orientation of the range images under consideration (with respect to rotation and translation), is robust to noise, and can reliably detect and localize naturally occurring landmarks in the operating environment. The algorithm is demonstrated in the application of a road vehicle in an unmodified operating domain.

This paper presents a novel approach to the multi-vehicle simultaneous localisation and mapping (SLAM) problem that exploits the manner in which observations are fused into the global map of the environment to manage the computational complexity of the algorithm and improve the data association process. Rather than incorporating every observation directly into the global map of the environment, the constrained local submap filter (CLSF) relies on creating an independent, local submap of the features in the immediate vicinity of the vehicle. This local submap is then periodically fused into the global map of the environment. This representation is shown to reduce the computational complexity of maintaining the global map estimates as well as improving the data association process. This paper examines the prospect of applying the CLSF algorithm to the multi-vehicle SLAM problem.

This paper presents a novel approach to the simultaneous localisation and mapping algorithm that exploits the manner in which observations are fused into the global map of the environment to manage the computational complexity of the algorithm and improve the data association process. Rather than incorporating every observation directly into the global map of the environment, the constrained local submap filter relies on creating an independent, local submap of the features in the immediate vicinity of the vehicle. This local submap is then periodically fused into the global map of the environment using appropriately formulated constraints between the common feature estimates. This approach is shown to be effective in reducing the computational complexity of maintaining the global map estimates as well as improving the data association process.

Exploration involving mapping and concurrent localization in an unknown environment is a pervasive task in mobile robotics. In general, the accuracy of the mapping process depends directly on the accuracy of the localization process. This paper address the problem of maximizing the accuracy of the map building process during exploration by adaptively selecting control actions that maximize localisation accuracy. The map building and exploration task is modeled using an Occupancy Grid (OG) with concurrent localisation performed using a feature-based Simultaneous Localisation And Mapping (SLAM) algorithm. Adaptive sensing aims at maximizing the map information by simultaneously maximizing the expected Shannon information gain (Mutual Information) on the OG map and minimizing the uncertainty of the vehicle pose and map feature uncertainty in the SLAM process. The resulting map building system is demonstrated in an indoor environment using data from a laser scanner mounted on a mobile platform.

Integrated exploration strategy advocated in this paper refers to a tight coupling between the tasks of localization, mapping, and motion control and the effect of this. coupling on the overall effectiveness of an exploration strategy. Our approach to exploration calls for a balanced evaluation of alternative motion actions from the point of view of information gain, localization quality, and navigation cost. To provide a uniform basis of comparison of localization quality between different locations, a "localizability" metric is introduced. It is based on the estimate of the lowest vehicle pose covariance attainable from a given location.

The simultaneous localization and map building (SLAM) problem asks if it is possible for an autonomous vehicle to start in an unknown location in an unknown environment and then to incrementally build a map of this environment while simultaneously using this map to compute absolute vehicle location. Starting from estimation-theoretic foundations of this problem, the paper proves that a solution to the SLAM problem is indeed possible. The underlying structure of the SLAM problem is first elucidated. A proof that the estimated map converges monotonically to a relative map with zero uncertainty is then developed. It is then shown that the absolute accuracy of the map and the vehicle location reach a lower bound defined only by the initial vehicle uncertainty. Together, these results show that it is possible for an autonomous vehicle to start in an unknown location in an unknown environment and, using relative observations only, incrementally build a perfect map of the world and to compute simultaneously a bounded estimate of vehicle location. The paper also describes a substantial implementation of the SLAM algorithm on a vehicle operating in an outdoor environment using millimeter-wave radar to provide relative map observations. This implementation is used to demonstrate how some key issues such as map management and data association can be handled in a practical environment. The results obtained are cross-compared with absolute locations of the map landmarks obtained by surveying. In conclusion, the paper discusses a number of key issues raised by the solution to the SLAM problem including suboptimal map-building algorithms and map management.

This paper presents a new method for improving the accuracy of inertial measurement units (IMUs) mounted on land vehicles. The algorithm exploits nonholonomic constraints that govern the motion of a vehicle on a surface to obtain velocity observation measurements which aid in the estimation of the alignment of the IMU as well as the forward velocity of the vehicle. It is shown that this can be achieved without any external sensing provided that certain observability conditions are met. A theoretical analysis is provided together with a comparison of experimental results between a nonlinear implementation of the algorithm and an IMU/GPS navigation system. This comparison demonstrates the effectiveness of the algorithm. The real time implementation is also addressed through a multiple observation inertial aiding algorithm based on the information filter. The results show that the use of these constraints and vehicle speed guarantees the observability of the velocity and the attitude of the inertial unit, and hence bounds the errors associated with these states. The strategies proposed provides a tighter navigation loop which can sustain outages of GPS for a greater amount of time as compared to when the inertial unit is used with standard integration algorithms.

In this paper, a fuzzy global controller based on sliding mode control is proposed for a reasonable amalgamation of local controllers to meet various control requirements for complex systems operating in highly uncertain environments. Also, a fuzzy logic approach is developed for the smooth allocation of several algorithms of a controller to gain the advantages, and at the same time to alleviate the disadvantages of each control algorithm. Illustrative examples are provided.

Data association is the process of relating features observed in the environment to features viewed previously or to features in a map. This paper presents a graph theoretic method that is applicable to data association problems where the features are observed via a batch process. Batch observations detect a set of features simultaneously or with sufficiently small temporal difference that, with motion compensation, the features can be represented with precise relative coordinates. This data association method is described in the context of two possible navigation applications: metric map building with simultaneous localisation, and topological map based localisation. Experimental results are presented using an indoor mobile robot with a 2D scanning laser sensor. Given two scans from different unknown locations, the features common to both scans are mapped to each other and the relative change in pose (position and orientation) of the vehicle between the two scans is obtained.

The theoretical basis of the solution to the simultaneous localisation and map building (SLAM) problem where an autonomous vehicle starts in an unknown location in an unknown environment and then incrementally build a map of landmarks present in this environment while simultaneously using this map to compute absolute vehicle location is well understood. Although a number of SLAM implementations have appeared in the literature, the need to maintain the knowledge of the relative relationships between all the landmark location estimates contained in the map makes SLAM computationally intractable in implementations containing more than few tens of landmarks. In this paper, the theoretical basis and a practical implementation of a computationally efficient solution to SLAM is presented. The paper shows that it is indeed possible to remove a large percentage of the landmarks from the map without making the map-building process statistically inconsistent. Furthermore, it is shown that the efficiency of the SLAM can be maintained by judicious selection of landmarks, to be preserved in the map, based on their information content.

Compensating for friction is considered in the paper using the variable structure systems approach. First, variable structure-based observers are developed for friction estimation in mechanical systems with or without information of velocity. The estimates are then used for a model-based feedforward compensation for friction. For a non-model based approach, a robust sliding mode controller can also be used to cancel the influence of friction. Sigmoidal functions are used in lieu of signum functions to reduce chattering. Simulation results verify the validity of the proposed technique to compensate for friction of both static and dynamic models.

We present a system for behavior-based control of an autonomous underwater vehicle for the purpose of inspection of coral reefs, a task currently performed by divers holding a video camera while following a rope. Using sonar and vision-based approaches, behaviors have been developed for guiding the robot along its intended course, for maintaining a constant height above the sea floor, and for avoiding obstacles. A task-level controller selects which behaviors should be active according to user-defined plans and in response to system failures. Behavior arbitration has been implemented using both fuzzy logic and utility fusion. Initial experiments have been conducted in a natural coastal inlet, and the system is to be soon demonstrated in the coral reef environment.

We present results of the application of a simultaneous localisation and map building (SLAM) algorithm to estimate the motion of a submersible vehicle. Scans obtained from an on-board sonar are processed to extract stable point features in the environment. These point features are then used to build up a map of the environment while simultaneously providing estimates of the vehicle location. Results are shown from deployment in a swimming pool at the University of Sydney as well as from field trials in a natural environment along Sydney's coast. This work represents the first instance of a deployable underwater implementation of the SLAM algorithm.

Describes a real time, on-the-fly, roll and pitch alignment algorithm for inertial measurement units (IMUs) mounted on land vehicles. Unlike conventional strategies, the alignment is achieved without external observations. This is achieved by exploiting the nonholonomic constraints that govern the motion of a vehicle on a surface to obtain the roll and pitch of the IMU. The position of the vehicle along its path is still unobservable and the algorithm is effective only when there is sufficient excitation in all degrees of freedom. However, with the proposed alignment algorithm, the IMU is able to provide sufficiently accurate position information for substantially longer periods of time compared with conventional methods. Experimental results are provided along with a comparison of an IMU/GPS navigation loop showing the effectiveness of the algorithm.

This work presents a general nonlinear psi-angle approach that does not require coarse alignment. In this psi-angle model, the three misalignment angles are assumed all large. Three states are used to describe three psi-angles rather than the four used in previous works. The approach is identical to the standard small error methods when the process errors diminish to small angles. The position and velocity error models are also presented. Standard extended Kalman filter techniques are used to solve the nonlinear data fusion problem. Experimental results of in-flight inertial navigation systems (INS) alignment and calibration are presented considering total uncertainty in azimuth orientation using a low cost inertial measurement unit aided with a differential global positioning system.

This paper describes a series of trials that were done at an underground mine in New South Wales, Australia. Experimental results are presented from the data obtained during the field trials and suitable sensor suites for an autonomous mining vehicle navigation system are evaluated.

Describes the theoretical development and experimental evaluation of a navigation system for an autonomous load, haul, and dump truck based on the results obtained during extensive in-situ field trials. The particular contributions of the theoretical work are in designing the navigation system to be able to cope with vehicle slip in rough uneven terrain using information from inertial sensors, odometry, and a bearing only laser. Results are presented using data obtained during the field trials.

This paper describes the development and implementation of a high integrity navigation system, based on the combined use of the Global Positioning System (GPS) and an inertial measurement unit (IMU), for autonomous land vehicle applications. The paper focuses on the issue of achieving the integrity required of the navigation loop for use in autonomous systems. The paper highlights the detection of possible faults both before and during the fusion process in order to enhance the integrity of the navigation loop. The implementation of this fault detection methodology considers both low frequency faults in the IMU caused by bias in the sensor readings and the misalignment of the unit, and high frequency faults from the GPS receiver caused by multipath errors. The implementation, based on a low-cost, strapdown IMU, aided by either standard or carrier phase GPS technologies, is described. Results of the fusion process are presented.

This paper reports on the map-building and map-based localization of a load-haul-dump (LHD) truck in an underground mine using statistical pattern-matching techniques utilizing range images obtained from a scanning laser range-finder The map-building approach is based on an extended Kalman filter (EKF) and the resulting map is composed of poly-lines. Three approaches are proposed for the localization of the vehicle, namely the iterative closest point (ICP) approach, a reflective beacon based approach and the combined ICP-EKF approach, wherein, the last two approaches explicitly take into account the uncertainty associated with the observation data. These approaches are then compared using data gathered from an underground mine in Queensland, Australia for their relative merits subject to various factors and the corresponding results are presented.

Presents an efficient method of multi-sensor estimation for the navigation of vehicles. A decentralised architecture is used for the fusion of information obtained from several asynchronous measurements. The issue of the synchronisation of the information, which is critical in the proposed method, is addressed. The information form of the Kalman filter (information filter) is used as the main algorithm for position estimation. The method is implemented as part of the navigation system for an autonomous land vehicle. The navigation system includes two independent loops which communicate through an assimilation unit. One loop incorporates inertial and GPS information and the other uses millimetre wave radar and encoder measurements to obtain local estimates. The information obtained from each measurement is then broadcast to the other loops after being synchronised. This information is used in an assimilation stage to achieve more accurate estimates. The assimilation takes place periodically, where the assimilation frequency can be selected. The performance of the navigation method is examined by comparing the resulting position estimates to those of independent navigation loops.

This paper discusses the use of a 77 GHz millimeter wave radar as a guidance sensor for autonomous land vehicle navigation. A test vehicle has been fitted with a radar and encoders that give steer angle and velocity. An extended Kalman filter optimally fuses the radar range and bearing measurements with vehicle control signals to give estimated position and variance as the vehicle moves around a test site. The effectiveness of this data fusion is compared with encoders alone and with a satellite positioning system. Consecutive scans have been combined to give a radar image of the surrounding environment. Data in this format are invaluable for future work on collision detection and map building navigation.

A crucial problem in crane control is to identify exactly the position and orientation of the load in space. This paper describes a new non-contact method for determining the crane load location by means of an inertial navigation system (INS) and a Kalman filter. The Kalman filter estimates the spreader position, which is not observed by the INS. Experiments were conducted on a 1/15th geometric scale model of a quay-crane. The work has potential application in the development of integrated estimation and control systems for full-scale quay-cranes.

A new approach for the autonomous navigation of a load-haul-dump (LHD) truck in an underground mine is presented. The development of a minimal-structure combined ICP-EKF algorithm utilizing a scanning-laser range-finder for the localization of the vehicle is described. The iterative closest point (ICP) algorithm is employed for matching the scanned data to an existing map in the form of a poly-line. This combined approach efficiently deals with the uncertainty present in the range data. An extended Kalman filter (EKF) algorithm is employed, that exploits a nonlinear kinematic model incorporating the vehicle-slip, a nonlinear observation model based on the vertices of the poly-line map, and the bearing of the laser-observations. This provides reliable vehicle estimates. Real data gathered during a trial run in the mine is employed in testing the efficiency of this approach which is found to be robust with respect to occlusions and outliers, demonstrating the successful navigation of the LHD.

In many ways autonomous navigation of underwater vehicles is a 'holy grail' of subsea robotics. For an AUV to continually estimate its pose (position and orientation) and operate within an initially unknown environment a solution is required to the simultaneous map building and localisation or 'SLAM' problem. The paper describes and investigates an inertial and terrain based approach to this problem. The fusion of body frame inertial and sonar based world frame feature information can form part of a robust navigation algorithm suitable for implementation in both artificial and natural environments. Particular attention is given to the role of the sonar and its ability to detect and track terrain features.

This paper provides an analysis of Kalman filter based systems with respect to fault detection. By using frequency domain techniques, a metric is developed that describes the detectability of a fault by showing how a fault is transmitted to the filter innovations (if at all). Through experiment, it is shown that redundancy must be employed for guaranteed detection of faults, and that unlike sensors should be used. Further, it is shown that modelling errors can be treated within the same framework as "hard" actuator or sensor faults.

The objective of this paper is to introduce and investigate the issue of integrity in an INS/GPS navigation loop for autonomous land vehicle applications. The paper briefly outlines the standard fusion algorithm for the INS/GPS loop, while the focus is on the detection of possible faults both before and during the fusion process. The concept of fault detection focuses on the low frequency faults of the INS, caused by bias and drift, and the high frequency faults of the GPS unit caused by multipath errors and changes in satellite geometry.

This paper presents a novel robust sliding mode controller for a class of nonlinear systems. The control signal consists of: an equivalent control to assign the desired dynamics to the closed-loop system, a switching control to guarantee a sliding mode, and a fuzzy control to enhance fast tracking and to attenuate chattering. Sufficient conditions for asymptotic tracking are achieved when matching conditions are satisfied. The control method is applied to the position control of an electrohydraulic servo system. To verify the potential and effectiveness of the proposed controller, simulation results are given to demonstrate its strong robustness against a large range of parameter variations, load disturbance and nonlinear spring stiffness.

We examine the problem of constructing and maintaining a map of an autonomous vehicle's environment for the purpose of navigation, using evidential reasoning. The inherent uncertainty in the origin of measurements of sensors demands a probabilistic approach to processing, or fusing, the new sensory information to build an accurate map. In the paper, the map is based on a two-dimensional (2-D) occupancy grid. The sensor readings are fused into the map using the Dempster-Shafer inference rule. This evidential approach with its multivalued hypotheses allows quantitative analysis of the quality of the data. The map building system is experimentally evaluated using sonar data from real environments.

This paper describes the design and implementation of a semi-autonomous, and ultimately fully autonomous, container quay-crane. The new crane is based on a novel reeving arrangement which allows both fast and accurate gross motion as well as fine micropositioning. The paper describes the essential theory behind this design and presents experimental results from a 1/15th scale model. The proposed instrumentation of this crane is also briefly described as are key elements of the operator interface.

Real time estimation of soil parameters is essential in achieving precise, robust autonomous guidance and control of a tracked vehicle. The paper shows that the slip of the tracks over the terrain can be identified from trajectory data using an extended Kalman filter. The use of a suitable soil model can then allow key soil parameters to be estimated as the vehicle passes over the soil. Knowledge of the soil parameters may in turn be used to allow reference trajectories and control algorithms to be adjusted to suit the soil conditions.

Position information obtained from standard GPS receivers is known to be corrupted with coloured (time-correlated) noise. To make effective use of GPS information in a navigation system it is essential to model this coloured noise and to incorporate additional sensing to de-correlate and eliminate its effect. In this paper frequency domain techniques are employed to generate a model for GPS noise sources. This model shows clearly what type and combination of additional sensor information is necessary to de-correlate GPS errors and to make best use of position information in navigation tasks. The frequency-domain methodology proposed has wider application in the design of sensor suites for high-performance navigation systems. Experimental results are presented demonstrating the method in fusing standard GPS latitude and longitude information with information from a velocity sensor.

This paper describes the theoretical development and experimental evaluation of a guidance system for an autonomous load, haul and dump truck (LHD) for use in underground mining. The particular contributions of this paper are in designing the navigation system to be able to cope with vehicle slip in rough uneven terrain using information from an inertial navigation system (INS) and a bearing only laser. Results are presented using data obtained during field trials.

This paper presents the results of an experimental program for evaluating sensors and sensing technologies in an underground mining applications. The objective of the experiments is to infer what combinations of sensors will provide reliable navigation systems for autonomous vehicles operating in a harsh underground environment. Results from a wide range of sensors are presented and analysed. A conclusion as to the best combination of sensors is drawn.

Examines the problem of constructing and maintaining a map of of an autonomous vehicle's environment for the purpose of navigation, using evidential reasoning. The inherent uncertainty in the origin of measurements of sensors demands a probabilistic approach to processing, or fusing, the new sensory information to build an accurate map. In this paper, the map is based on a two-dimensional occupancy grid. The sensor readings are 'fused' into the map using the Dempster-Shafer inference rule. This evidential approach with its multi-valued hypotheses allows quantitative analysis of the quality of the data. The map building system is experimentally evaluated using sonar data from real environments.

This paper describes the design of a navigation system for an autonomous guided vehicle (AGV) system able to transport ISO standard cargo containers in a port environment. The navigation system is based on the use of millimeter wave radar sensors detecting the range and bearing to a number of fixed known beacons. The central navigation algorithm is an extended Kalman filter that exploits a model of the vehicle motion and radar observations to continuously provide estimates of the vehicle location. The main contribution of the system described in this paper lies in the use of a new, and relatively sophisticated process model describing the motion of a large vehicle, and in the incorporation of this with a novel sensing system.

A nonlinear process model for the navigation of a high-speed conventional road vehicle is described. In simulations it is shown to significantly reduce the errors in estimating of vehicle position and orientation. The model also performs limited online estimation of certain critical tyre parameters such as mean radius and stiffness.

A novel Kalman filter based sonar navigation system is presented that utilises directed sensing techniques to achieve continuous indoor localisation of a mobile robot. Data from extensive trials is summarised demonstrating continuous, reliable, localisation using this system in a wide variety of active industrial sites. Full details of experimental method and performance assessment are given.

A low-cost solid-state inertial navigation system (INS) for mobile robotics applications is described. Error models for the inertial sensors are generated and included in an extended Kalman filter (EKF) for estimating the position and orientation of a moving robot vehicle. Two different solid-state gyroscopes have been evaluated for estimating the orientation of the robot. Performance of the gyroscopes with error models is compared to the performance when the error models are excluded from the system. Similar error models have been developed for each axis of a solid-state triaxial accelerometer and for a conducting-bubble tilt sensor which may also be used as a low-cost accelerometer. An integrated inertial platform consisting of three gyroscopes, a triaxial accelerometer and two tilt sensors is described.>

An error model for a solid-state gyroscope developed previously by the authors is included in a Kalman filter for improving the orientation estimate of a mobile robot. Orientation measurement with the error model is compared to the performance when no error model is incorporated in the system. The results demonstrate that without error compensation, the error in localization is between 5-15/spl deg//min but can be improved by a factor of 5 to 7 if an adequate error model is supplied. Results from tests of this gyroscope on a large outdoor mobile robot system are described and compared to the results obtained from the robot's own radar-based guidance system. Like all inertial systems, the platform requires additional information from some absolute position sensing mechanism to overcome long-term drift. However, the results show that with careful and detailed modelling of error sources, low cost inertial devices can provide valuable orientation and position information particularly for outdoor mobile robot applications.>

As autonomous guided vehicles claim wider applications, the need for a flexible navigation system increases. Previously developed systems designed to incorporate localisation and environmental modelling have suffered from poor operation speed, often depending upon a "stop look update" policy. We describe a multi-track extended Kalman filter navigation system which offers real time localisation while simultaneously constructing a map consisting of geometric features, each in turn described by an extended Kalman filter. Dynamic operation is achieved by utilising the increased scan rate and data quality of infrared scanning to overcome the need for a multi-hypothesis framework or other such computationally expensive paradigm.>

This paper describes the development of an extended Kalman filter model for autonomous ground vehicle navigation using integrated civilian-band GPS position measurements and inertial navigation information. The Kalman filter model incorporates a detailed statistical model of all major GPS errors including those associated with selective availability (SA) range errors. The resulting model consists of 14 states. Data showing the accuracy of this filter is presented.>

A low-cost, solid-state inertial navigation system for robotics applications is described. Error models for the inertial sensors are generated and included in an extended Kalman filter (EKF) for estimating the position and orientation of a moving robot vehicle. A solid-state gyroscope and an accelerometer have been evaluated. Without error compensation, the error in orientation is between 5-15/spl deg//min but can be improved at least by a factor of five if an adequate error model is supplied. Similar error models have been developed for each axis of a solid-state triaxial accelerometer. Linear position estimation with accelerometers and tilt sensors is more susceptible to errors due to the double integration process involved in estimating position. WIth the system described here, the position drift rate is 1-8 cm/s, depending on the frequency of acceleration changes. The results show that with careful and detailed modeling of error sources, low cost inertial sensing systems can provide valuable position information.

With the use of a scanning optical ranger, a dense map of an environment has been constructed. From this, line segment targets are extracted and matched against an a priori map to obtain observations for an extended Kalman filter. This filter is then able to update predictions of an autonomous guided vehicle's position (made using odometry) in real time, offering high-speed position estimation using inexpensive sensing techniques.

The kinematics for modular wheeled mobile robots is presented. Basic theory of planar motion is used to derive inverse and forward kinematics. The inverse kinematics is inherently modular, whereas the information form of a least-squares estimator is required to completely modularize the forward kinematics. An implementation is described and results are shown.

This paper addresses the design, verification and implementation of a fully decentralized linear quadratic Gaussian (LQG) control system based on transputer architecture. A decentralized, information filter is developed and extended to solve the decentralized control problem, keeping consistency with information space ideas. Locally processed information is shared by all nodes, so that optimal control is computed using the best, locally determined, global state estimate. There is no central processing site, and hence the system performance degrades gracefully under structural perturbations. The designed system is modeled and verified using Communicating Sequential Process (CSP) algebra. Software is developed in Parallel ANSI C. A simulated test implementation is effected using transputer-based hardware. Validation and performance analysis are carried out. Application to intelligent control of a modular, decentralized navigating robot is considered.

The problem of identification of objects being tracked by a fully decentralized surveillance system is considered. A decentralized multisensor system is used to track targets (people and mobile robots) as they enter and move around a factory assembly room performing tasks. The sensors used in this system (CCD cameras) reveal information about the targets that is sufficiently rich to allow them not only to be tracked, but also identified as a person, a robot, etc. This identity information can be used to aid in man-machine interface design and to facilitate situation assessment. After defining the identification problem, a centralized Bayesian algorithm is developed for determining the identity of each target based on each sensor's information. The algorithm is decentralized and its performance compared to the centralized version. Results of an implementation of the algorithm working on real data from the surveillance system are presented.>

Describes the design and implementation of a task allocation for a human-computer interface that is capable of adapting the human's workload, online, according to his ability to perform tasks for systems, where the computer acts as a backup decision maker. During the operation of the system, the human and the computer's service and error rates are continually measured. An embedded human model estimates the change in the human's ability to service his tasks baled on the measured service and error rates. Future human and computer service rates and task arrival rates in the system are then predicted. Every time a new task arrives in the system the interface decides whether the human is capable of dealing with all current tasks before the next task arrives. If the human is unable to deal with the workload, the computer is switched on to aid him. The interface is broken down into modules, with each module's quantitative behavior defined separately. A probabilistic sensitivity analysis of the task allocation equations develops metrics that can be used to assess the task allocation's robustness with regard to uncertainties in the predicted human and system performances. The interface is implemented in a real time, human operated surveillance system.>

The authors describe an algorithm for implementing a multisensor system in a model-based environment with consideration of the constraints. Based on an environment model, geometric features and constraints are generated from a CAD model database. Sensor models are used to predict sensor response to certain features and to interpret raw sensor data. A constrained MMS (minimum mean squared) estimator is used to recursively predict, match, and update feature location. The effects of applying various constraints in estimation were shown by simulation system mounted on a robot arm for localization of known object features.>

Discusses a significant open problem in mobile robotics: simultaneous map building and localization, which the authors define as long-term globally referenced position estimation without a priori information. This problem is difficult because of the following paradox: to move precisely, a mobile robot must have an accurate environment map; however, to build an accurate map, the mobile robot's sensing locations must be known precisely. In this way, simultaneous map building and localization can be seen to present a question of 'which came first, the chicken or the egg?' (The map or the motion?) When using ultrasonic sensing, to overcome this issue the authors equip the vehicle with multiple servo-mounted sonar sensors, to provide a means in which a subset of environment features can be precisely learned from the robot's initial location and subsequently tracked to provide precise positioning.>

Present two algorithms for tracking and identification in decentralized multi-sensor systems. Decentralized architectures have many benefits in terms of modularity, speed and robustness. The state estimation (tracking) algorithm is a decentralized Kalman filter (DKF) based on the extended Kalman filter. Identification is achieved by the decentralized Bayesian identification (DBI) algorithm, which identifies targets being tracked. For each of the algorithms The authors discuss optimality and the effect of reducing connectivity. The structure of the algorithms leads to the development of an architecture for a modular sensing node based on communication considerations. The authors present example implementations of both algorithms on actual transputer-based sensing nodes. They describe RCD (region of constant depth) tracking and for this develop a monopulse sonar arrangement with which they implement real-time autonomous tracking by a single sensor node. The second example implementation describes identification of targets being tracked by the DKF using the DBI on actual CCD camera-based nodes.>

Describes an implementation of a model based multi-sensor localisation system using high frequency ultrasonic and selective infrared transducers. The sensor system is mounted on the end of an ADEPT robot arm and measurements are taken while the robot arm follows a pre-determined trajectory. The methodology is continuous estimation by tracking through prediction and observation. A sensor model and partial knowledge of the environment geometry is essential. The extended Kalman filter algorithm is used to recursively estimate the location of certain objects relative to the robot arm and the uncertainty of measurements and estimation process are also maintained. Optimal sensing strategies consisting of straight line motions are proposed based on effectiveness of uncertainty reduction. An implementation of the system as a hole finder for automatic freight container handling is investigated. Experimental results are presented together with simulated data. Utilisation of geometric constraints in localisation and estimation is discussed.>

The application of the extended Kaman filter to the problem of mobile robot navigation in a known environment is presented. An algorithm for, model-based localization that relies on the concept of a geometric beacon, a naturally occurring environment feature that can be reliably observed in successive sensor measurements and can be accurately described in terms of a concise geometric parameterization, is developed. The algorithm is based on an extended Kalman filter that utilizes matches between observed geometric beacons and an a priori map of beacon locations. Two implementations of this navigation algorithm, both of which use sonar, are described. The first implementation uses a simple vehicle with point kinematics equipped with a single rotating sonar. The second implementation uses a 'Robuter' mobile robot and six static sonar transducers to provide localization information while the vehicle moves at typical speeds of 30 cm/s.>

A fully decentralized architecture is presented for data fusion problems. This architecture takes the form of a network of sensor nodes, each with its own processing facility, which together do not require any central processor or any central communication facility. In this architecture, computation is performed locally and communication occurs between any two nodes. Such an architecture has many desirable properties, including robustness to sensors failure and flexibility to the addition or loss of one or more sensors. This architecture is appropriate for the class of extended Kalman filter (EKF)-based geometric data fusion problems. The starting point for this architecture is an algorithm which allows the complete decentralization of the multisensor EKF equations among a number of sensing nodes. This algorithm is described, and it is shown how it can be applied to a number of different data-fusion problems. An application of this algorithm to the problem of multicamera, real-time tracking of objects and people moving through a room is described.>

A tactile sensor has been developed which uses a layer of photoelastic material as its primary sensing element. This sensor has several desirable characteristics, including very high resolution, adaptable shape, and edge enhancement. Interestingly, the sensor has a similar sensing mechanism to human skin. A mathematical model has been developed for the sensor in an attempt to better understand its operation and to optimize its performance. The model draws heavily on the established fields of continuum mechanics and photoelastic stress analysis. The results of this analysis provide several important insights into tactile sensing and human touch, including an explanation of the importance of motion to touch, and the enhanced sensitivity of the sensor (and human skin) in detected edges.>

The author suggests that to operate efficiently, a robot system must be able to represent, account for, and reason about the effects of uncertainty in areas in which geometric analysis also plays an important part. He proposed that uncertainty be represented as an intrinsic part of all geometric descriptions. Toward this goal he develops a description of uncertain geometric features as families of parametrized functions together with a distribution function defined on the associated parameter vector. Uncertain points, curves, and surfaces are considered, and it is shown how they can be manipulated and transformed between coordinate frames in an efficient and consistent manner. The effectiveness of these techniques is demonstrated by application to the problem of developing maximal-information sensing strategies.>

Robots must operate in an environment which is inherently uncertain. This uncertainty is important in areas such as modeling, planning and the motion of manipulators and objects; areas where geometric analysis also plays an important part. To operate efficiently, a robot system must be able to represent, account for, and reason about the effects of uncertainty in these geometries in a consistent manner. We maintain that uncertainty should be represented as an intrinsic part of all geometric descriptions. We develop a description of uncertain geometric features as families of parameterized functions together with a distribution function defined on the associated parameter vector. We consider uncertain points, curves and surfaces, and show how they can be manipulated and transformed between coordinate frames in an efficient and consistent manner. The effectiveness of these techniques is demonstrated by application to the problem of developing maximal information sensing strategies.

We present a theory and methodology for integrating and propagating geometric sensor observations. The integration policy takes any number of disparate, partial and uncertain observations and optimaly combines them into a minimum-risk best estimate consensus view of the state of the environment. These consensus observations are considered to be integrated into a geometric model of the world. A methodology is developed that propagates new observations through this world model, maintaining consistency amongst objects and making maximum use of sensor information.

A new model referenced adaptive control scheme for a general actuator-link assembly based on the Lyapunov stability criterion is presented. A general model of the actuator and link dynamics and constraints is formulated as an integral structure. Application of a state feedback gain introduces the adjustable system. The model and control are formulated to address such physical considerations as motor saturation, static friction, unknown loads and uncertain plant parameters. This control scheme is shown to solve several current problems in application of adaptive control to a manipulator system; namely it accounts for the effects of actuator dynamics and constraints, and the static error due to friction and uncertain steady state loads. This scheme has been simulated and demonstrated to be robust over a variety of practical conditions. Finally the practical realization of this algorithm is discussed with reference to the Pennsylvania Anthropomorphic Robot Manipulator currently under construction.