John Mullane

A Random-Finite-Set Approach to Bayesian SLAM

This paper proposes an integrated Bayesian frame work for feature-based simultaneous localization and map building (SLAM) in the general case of uncertain feature number and data association. By modeling the measurements and feature map as random finite sets (RFSs), a formulation of the feature-based SLAM problem is presented that jointly estimates the number and location of the features, as well as the vehicle trajectory. More concisely, the joint posterior distribution of the set-valued map and vehicle trajectory is propagated forward in time as measurements arrive, thereby incorporating both data association and feature management into a single recursion. Furthermore, the Bayes optimality of the proposed approach is established. A first-order solution, which is coined as the probability hypothesis density (PHD) SLAM filter, is derived, which jointly propagates the posterior PHD of the map and the posterior distribution of the vehicle trajectory. A Rao-Blackwellized (RB) implementation of the PHD-SLAM filter is proposed based on the Gaussian-mixture PHD filter (for the map) and a particle filter (for the vehicle trajectory). Simulated and experimental results demonstrate the merits of the proposed approach, particularly in situations of high clutter and data association ambiguity.

SLAM Gets a PHD: New Concepts in Map Estimation

Having been referred to as the Holy Grail of autonomous robotics research, simultaneous localization and mapping (SLAM) lies at the core of most the autonomous robotic applications. This article explains the recent advances in the representations of robotic sensor measurements and the map itself as well as their consequences on the robustness of SLAM. Fundamentally, the concept of a set-based measurement and map state representation allows all of the measurement information, spatial and detection, to be incorporated into joint Bayesian SLAM frameworks. Modeling measurements and the map state as random finite sets (RFSs) rather than the traditionally adopted random vectors is not merely a triviality of notation. It will be demonstrated that a set-based framework circumvents the necessity for any fragile data association and map management heuristics, which are necessary in vector-based solutions.

Random finite sets for robot mapping and SLAM : new concepts in autonomous robotic map representations

The monograph written by John Mullane, Ba-Ngu Vo, Martin Adams and Ba-Tuong Vo is devoted to the field of autonomous robot systems, which have been receiving a great deal of attention by the research community in the latest few years. The contents are focused on the problem of representing the environment and its uncertainty in terms of feature based maps. Random Finite Sets are adopted as the fundamental tool to represent a map, and a general framework is proposed for feature management, data association and state estimation. The approaches are tested in a number of experiments on both ground based and marine based facilities.

A random set formulation for Bayesian SLAM

This paper presents an alternative formulation for the Bayesian feature-based simultaneous localisation and mapping (SLAM) problem, using a random finite set approach. For a feature based map, SLAM requires the joint estimation of the vehicle location and the map. The map itself involves the joint estimation of both the number of features and their states (typically in a 2D Euclidean space), as an a priori unknown map is completely unknown in both landmark location and number. In most feature based SLAM algorithms, so-called dasiafeature managementpsila algorithms as well as data association hypotheses along with extended Kalman filters are used to generate the joint posterior estimate. This paper, however, presents a recursive filtering algorithm which jointly propagates both the estimate of the number of landmarks, their corresponding states, and the vehicle pose state, without the need for explicit feature management and data association algorithms. Using a finite set-valued joint vehicle-map state and set-valued measurements, the first order statistic of the set, called the intensity, is propagated via the probability hypothesis density (PHD) filter, from which estimates of the map and vehicle can be jointly extracted. Assuming a mildly non-linear Gaussian system, an extended-Kalman Gaussian Mixture implementation of the recursion is then tested for both feature-based robotic mapping (known location) and SLAM. Results from the experiments show promising performance for the proposed SLAM framework, especially in environments of high spurious measurements.

Including probabilistic target detection attributes into map representations

Abstract Range measuring sensors can play an extremely important role in robot navigation. All range measuring devices rely on a ‘detection criterion’ made in the presence of noise, to determine when the transmitted signal is considered detected and hence a range reading is obtained. In commonly used sensors, such as laser range finders and polaroid sonars, the criterion under which successful detection is assumed, is kept hidden from the user. However, ‘detection decisions’ on the presence of noise still take place within the sensor. This paper integrates signal detection probabilities into the map building process which provides the most accurate interpretation of such sensor data. To facilitate range detection analysis, map building with a frequency modulated continuous wave millimetre wave radar (FMCW MMWR), which is able to provide complete received power-range spectra for multiple targets down range is considered. This allows user intervention in the detection process and although not directly applicable to the commonly used ‘black-box’ type range sensors, provides insight as to how not only range values, but received signal strength values should be incorporated into the map building process. This paper presents two separate methods of map building with sensors which return both range and received signal power information. The first is an algorithm which uses received signal-to-noise power to make an estimates of the range to multiple targets down range, without any signal distribution assumptions. We refer to this as feature detection based on target presence probability (TPP). In contrast to the first method, the second method does use assumptions on the statistics of the signal in target presence and absence scenarios to formulate a probabilistic likelihood detector. This allows for an increased rate of convergence to ground truth. Evidence theory is then introduced to model and update successive observations in a recursive fashion. Both methods are then compared using real MMWR data sets from indoor and outdoor experiments.

Rao-Blackwellised PHD SLAM

This paper proposes a tractable solution to feature-based (FB) SLAM in the presence of data association uncertainty and uncertainty in the number of features. By modeling the feature map as a random finite set (RFS), a rigorous Bayesian formulation of the FB-SLAM problem that accounts for uncertainty in the number of features and data association is presented. As such, the joint posterior distribution of the set-valued map and vehicle trajectory is propagated forward in time as measurements arrive. A first order solution, coined the PHD-SLAM filter, is derived, which jointly propagates the posterior PHD or intensity function of the map and the posterior distribution of the trajectory of the vehicle. A Rao-Blackwellised implementation of the PHD-SLAM filter is proposed based on the Gaussian mixture PHD filter for the map and a particle filter for the vehicle trajectory. Simulated results demonstrate the merits of the proposed approach, particularly in situations of high clutter and data association ambiguity.

Robotic Mapping Using Measurement Likelihood Filtering

The classical occupancy grid formulation requires the use of a priori known measurement likelihoods whose values are typically either assumed or learned from training data. Furthermore, in previous approaches, the likelihoods used to propagate the occupancy map variables are, in fact, independent of the state of interest and are derived from the spatial uncertainty of the detected point. This allows the use of a discrete Bayes filter as a solution to the problem, as discrete occupancy measurement likelihoods are used. In this paper, we first shown that once the measurement space is redefined, theoretically accurate and state-dependant measurement likelihoods can be obtained and used in the propagation of the occupancy random variable. The required measurement likelihoods for occupancy filtering are, in fact, those commonly encountered in both the landmark detection and data association hypotheses decisions. However, the required likelihoods are generally a priori unknown as they are a highly non-linear function of the landmarks signal-to-noise ratio and the surrounding environment. The probabilistic occupancy mapping problem is therefore reformulated as a continuous joint estimation problem where the measurement likelihoods are treated as continuous random states which must be jointly estimated with the map. In particular, this work explicitly considers the sensors detection and false-alarm probabilities in the occupancy mapping formulation. A particle solution is proposed which recursively estimates both the posterior on the map and the measurement likelihoods. The ideas presented in this paper are demonstrated in the field robotics domain using a millimeter wave radar sensor and comparisons with previous approaches, using constant discrete measurement likelihoods, are shown. A manually constructed ground-truth map and satellite imagery are also provided for map validation.

Evidential versus Bayesian Estimation for Radar Map Building

This paper discusses the role played by signal detection algorithms in the mobile robot map building problem. Typical mapping techniques make the assumption that the internal signal detection, which is required to produce an (r, rho) point estimate, is ideal. That is, the probability of detecting the signal is unity, and the probabilities of a false alarm or missed detection are zero. In the case of grid mapping, this allows for the occupancy probability to be distributed under the constraint of a unity summation amongst affected cells. In the case of SLAM, this allows for a features (x,y) coordinates to be modeled with (Gaussian) probability density functions. This paper shows that typical signal detection algorithms contain all the necessary measurement models to exactly calculate the map occupancy estimates. Furthermore, once restrictive signal assumptions are relaxed, its shown that evidence theory and not Bayesian theory should be used in the combination and updating of the map estimates. The ideas presented in this paper are demonstrated in the field robotics domain using a millimeter wave radar sensor. Target presence and absence beliefs are derived directly from signal likelihood ratios as opposed to a priori assigned constants as is typical for mapping algorithms. Results obtained from outdoor sensing experiments, show the improvement of this new model, given targets of fluctuating radar cross section (RCS)

Circumventing the Feature Association Problem in SLAM

In autonomous applications, a vehicle requires reliable estimates of its location and information about the world around it. To capture prior knowledge of the uncertainties in a vehicles motion response to input commands and sensor measurements, this fundamental task has been cast as probabilistic Simultaneous Localization and Map building (SLAM). SLAM has been investigated as a stochastic filtering problem in which sensor data is compressed into features, which are consequently stacked in a vector, referred to as the map. Inspired by developments in the tracking literature, recent research in SLAM has recast the map as a Random Finite Set (RFS) instead of a random vector, with huge mathematical consequences. With the application of recently formulated Finite Set Statistics (FISST), such a representation circumvents the need for fragile feature management and association routines, which are often the weakest component in vector based SLAM algorithms. This tutorial demonstrates that true sensing uncertainty lies not only in the spatial estimates of a feature, but also in its existence. This gives rise to sensor probabilities of detection and false alarm, as well as spatial uncertainty values. By re-addressing the fundamentals of SLAM under an RFS framework, it will be shown that it is possible to estimate the map in terms of true feature number, as well as location. The concepts are demonstrated with short range radar, which detects multiple features, but yields many false measurements. Comparison of vector, and RFS SLAM algorithms shows the superior robustness of RFS based SLAM to such realistic sensing defects.

X-band radar based SLAM in Singapore's off-shore environment

This paper presents a simultaneous localisation and mapping (SLAM) algorithm implemented on an autonomous sea kayak with a commercial off-the-shelf X-band marine radar mounted. The Autonomous Surface Craft (ASC) was driven in an off-shore test site in Singapores southern Selat Puah marine environment. Data from the radar, GPS and an inexpensive single-axis gyro data were logged by an on-board processing unit as the ASC traversed the environment, which comprised geographical and surface vessel landmarks. An automated feature extraction routine is presented, based on a probabilistic landmark detector, followed by a clustering and centroid approximation approach. With restrictive feature modeling, and a lack of vehicle control input information, it is demonstrated that via the novel RB-PHD-SLAM Filter, useful results can be obtained, despite an actively rolling and pitching ASC on the sea surface. In addition, the merits of investigating ASC SLAM are demonstrated, particularly with respect to the map estimation, obstacle avoidance and target tracking problems. Despite the presence of GPS and gyro data, heading information on such small ASCs is greatly compromised which induces large sensing error, further accentuate by the large range of the radar sensor. This work is a step towards realising an ASC capable of performing environmental or security surveillance and reporting a real-time active awareness of the above-water scene.