Bayesian optimization (BO) is a popular method for optimizing expensive black-box functions. BO has several well-documented shortcomings, including computational slowdown with longer optimization runs, poor suitability for non-stationary or ill-conditioned objective functions, and poor convergence characteristics. Several algorithms have been proposed that incorporate local strategies, such as trust regions, into BO to mitigate these limitations; however, none address all of them satisfactorily. To address these shortcomings, we propose the LABCAT algorithm, which extends trust-region-based BO by adding a rotation aligning the trust region with the weighted principal components and an adaptive rescaling strategy based on the length-scales of a local Gaussian process surrogate model with automatic relevance determination. Through extensive numerical experiments using a set of synthetic test functions and the well-known COCO benchmarking software, we show that the LABCAT algorithm outperforms several state-of-the-art BO and other black-box optimization algorithms.

Simultaneous Localization and Mapping (SLAM) is a crucial component to the push towards full autonomy of robotic systems, yet it is computationally expensive and can rarely achieve real-time execution speeds on embedded platforms. Therefore, a need exists to evaluate the performance of SLAM algorithms in practical embedded environments – this paper addresses this need by creating prediction models to estimate the performance that ORB-SLAM3 can achieve on embedded platforms. The paper uses three embedded platforms: Nvidia Jetson TX2, Raspberry Pi 3B+ and the Raspberry Pi 4B, to generate a dataset that is used in training and testing performance prediction models. The process of profiling ORB-SLAM3 aids in the selection of inputs to the prediction model as well as benchmarking the embedded platforms’ performances by using PassMark. The EuRoC micro aerial vehicle (MAV) dataset is used to generate the average tracking time that the embedded platforms can achieve when executing ORB-SLAM3, which is the target of the prediction model. The best-performing model has the following results 2.84%, 3.93%, and 0.95 for MAE, RMSE and R2 score respectively. The results show the feasibility of predicting the performance that SLAM applications can achieve on embedded platforms.

Simultaneous localisation and mapping (SLAM) play a vital role in autonomous robotics. Robotic platforms are often resource-constrained, and this limitation motivates resource-efficient SLAM implementations. While sparse visual SLAM algorithms offer good accuracy for modest hardware requirements, even these more scalable sparse approaches face limitations when applied to largescale and long-term scenarios. A contributing factor is that the point clouds resulting from SLAM are inefficient to use and contain significant redundancy. This paper proposes the use of subset selection algorithms to reduce the map produced by sparse visual SLAM algorithms. Information-theoretic techniques have been applied to simpler related problems before, but they do not scale if applied to the full visual SLAM problem. This paper proposes a number of novel information-theoretic utility functions for map point selection and optimises these functions using greedy algorithms. The reduced maps are evaluated using practical data alongside an existing visual SLAM implementation (ORB-SLAM 2). Approximate selection techniques proposed in this paper achieve trajectory accuracy comparable to an offline baseline while being suitable for online use. These techniques enable the practical reduction of maps for visual SLAM with competitive trajectory accuracy. Results also demonstrate that SLAM front-end performance can significantly impact the performance of map point selection. This shows the importance of testing map point selection with a front-end implementation. To exploit this, this paper proposes an approach that includes a model of the front-end in the utility function when additional information is available. This approach outperforms alternatives on applicable datasets and highlights future research directions. & COPY; 2023 Elsevier B.V. All rights reserved.

Conflict prediction is a vital component of path planning for autonomous vehicles. Prediction methods must be accurate for reliable navigation, but also computationally efficient to enable online path planning. Efficient prediction methods are especially crucial when testing large sets of candidate trajectories. We present a prediction method that has the same accuracy as existing methods, but up to an order of magnitude faster. This is achieved by rewriting the conflict prediction problem in terms of the first-passage time distribution using a dimension-reduction transform. First-passage time distributions are analytically derived for a subset of Gaussian processes describing vehicle motion. The proposed method is applicable to 2-D stochastic processes where the mean can be approximated by line segments, and the conflict boundary can be approximated by piece-wise straight lines. The proposed method was tested in simulation and compared to two probability flow methods, as well as a recent instantaneous conflict probability method. The results demonstrate a significant decrease of computation time.

Simultaneous localisation and mapping (SLAM) play a vital role in autonomous robotics. Robotic platforms are often resource-constrained, and this limitation motivates resource-efficient SLAM implementations. While sparse visual SLAM algorithms offer good accuracy for modest hardware requirements, even these more scalable sparse approaches face limitations when applied to large-scale and long-term scenarios. A contributing factor is that the point clouds resulting from SLAM are inefficient to use and contain significant redundancy. This paper proposes the use of subset selection algorithms to reduce the map produced by sparse visual SLAM algorithms. Information-theoretic techniques have been applied to simpler related problems before, but they do not scale if applied to the full visual SLAM problem. This paper proposes a number of novel information\hyp{}theoretic utility functions for map point selection and optimises these functions using greedy algorithms. The reduced maps are evaluated using practical data alongside an existing visual SLAM implementation (ORB-SLAM 2). Approximate selection techniques proposed in this paper achieve trajectory accuracy comparable to an offline baseline while being suitable for online use. These techniques enable the practical reduction of maps for visual SLAM with competitive trajectory accuracy. Results also demonstrate that SLAM front-end performance can significantly impact the performance of map point selection. This shows the importance of testing map point selection with a front-end implementation. To exploit this, this paper proposes an approach that includes a model of the front-end in the utility function when additional information is available. This approach outperforms alternatives on applicable datasets and highlights future research directions.

A dynamic environment can be challenging for a robot to navigate; it should avoid collisions with objects while determining its position in its environment (localisation). Thus, it is necessary for a mobile robot to take measurements of its environment, such as features from camera images, to determine whether objects are static or dynamic (motion segmentation). This is difficult to do as knowledge of static objects is required for localisation which is then used to track the trajectories of dynamic objects. This paper proposes a motion segmentation technique that classifies objects as static or dynamic by measuring the change in distance between them across many time steps; this removes the need for localisation information. The technique is adapted from a probabilistic method for outlier removal and existing motion segmentation techniques. A simple, 1D environment is simulated to show proof of concept. Additionally, a few strategies for PGM model construction are investigated where the results show a clear relationship between accuracy and computational times.

Stereo vision systems rely on accurate feature matching to provide valid stereo reconstruction and pose estimation. This accuracy is achieved through outlier removal techniques, such as RANSAC. However, images also contain semantic information, which can be extracted using neural networks. This paper proposes an additional outlier removal method, where the images are semantically segmented using a neural network, before the features identified are assigned semantic identifiers using a probabilistic data association technique, and matches are evaluated based on this added semantic information. This blending of feature-based techniques with dense semantic maps allows for more information to be tied to each feature, not just its position in the image. This opens paths to applications like class-based clustering. The approach proposed is compared to a traditional outlier removal system by comparing the produced disparity values to known ground truth measurements, and assessed for accuracy and execution speed. It is shown how the addition of semantic segmentation does improve the accuracy of disparity measurements in stereo images, with a loss in processing speed. However, this loss can be mitigated by utilising more specialised hardware.

In this paper, we propose two efficient, approximate formulations of the multi-sensor labelled multi-Bernoulli (LMB) filter, which both allow the sensors' measurement updates to be computed in parallel. Our first filter is based on the direct mathematical manipulation of the multi-sensor, multi-object Bayes filter's posterior distribution. Unfortunately, it requires the division of probability distributions and its extension beyond linear Gaussian applications is not obvious. Our second filter is based on geometric average fusion and it approximates the multi-sensor, multi-object Bayes filter's posterior distribution using the geometric average of each sensor's measurement-updated distribution. This filter can be used under non-linear conditions; however, it is not as accurate as our first filter. In both cases, we approximate the LMB filter's measurement update using an existing loopy belief propagation algorithm. Both filters have a constant complexity in the number of sensors, and linear complexity in both number of measurements and objects. This is an improvement on an iterated-corrector LMB (IC-LMB) filter, which has linear complexity in the number of sensors. The proposed filters are of interest when tracking many objects using several sensors, where filter run-time is more important than filter accuracy. We evaluate both filters' performances on simulated data and the results indicate that the filters' loss of accuracy compared to the IC-LMB filter is not significant.

This paper proposes a novel approach of Gaussian process observation set compression based on a squared difference measure. It is used to discard observations to speed up Gaussian process prediction while retaining the information encoded in the full set of observations. Furthermore, this paper compares the regression performance of a compressed Gaussian process to its uncompressed version and to a randomly downsampled Gaussian process for a standard two-dimensional test function. The empirical results of this paper show that this is an effective algorithm for Gaussian process compression, speeding up prediction while maintaining predictive accuracy with regards to the predicted means.

In this paper, we propose a parametrised factor that enables inference on Gaussian networks where linear dependencies exist among the random variables. Our factor representation is effectively a generalisation of traditional Gaussian parametrisations where the positive-definite constraint of the covariance matrix has been relaxed. For this purpose, we derive various statistical operations and results (such as marginalisation, multiplication and affine transformations of random variables) that extend the capabilities of Gaussian factors to these degenerate settings. By using this principled factor definition, degeneracies can be accommodated accurately and automatically at little additional computational cost. As illustration, we apply our methodology to a representative example involving recursive state estimation of cooperative mobile robots.

Path planners based on basic rapidly-exploring random trees (RRTs) are quick and efficient, and thus favourable for real-time robot path planning, but are almost-surely suboptimal. In contrast, the optimal RRT (RRT*) converges to the optimal solution, but may be expensive in practice. Recent work has focused on accelerating the RRT*'s convergence rate. The most successful strategies are informed sampling, path optimisation, and a combination thereof. However, informed sampling and its combination with path optimisation have not been applied to the basic RRT. Moreover, while a number of path optimisers can be used to accelerate the convergence rate, a comparison of their effectiveness is lacking. This paper investigates the use of informed sampling and path optimisation to accelerate planners based on both the basic RRT and the RRT*, resulting in a family of algorithms known as optimised informed RRTs. We apply different path optimisers and compare their effectiveness. The goal is to ascertain if applying informed sampling and path optimisation can help the quick, though almost-surely suboptimal, path planners based on the basic RRT attain comparable or better performance than RRT*-based planners. Analyses show that RRT-based optimised informed RRTs do attain better performance than their RRT*-based counterparts, both when planning time is limited and when there is more planning time.

For mobile robots to operate autonomously in general environments, perception is required in the form of a dense metric map. For this purpose, we present the stochastic triangular mesh (STM) mapping technique: a 2.5-D representation of the surface of the environment using a continuous mesh of triangular surface elements, where each surface element models the mean plane and roughness of the underlying surface. In contrast to existing mapping techniques, an STM map models the structure of the environment by ensuring a continuous model, while also being able to be incrementally updated with linear computational cost in the number of measurements. We reduce the effect of uncertainty in the robot pose (position and orientation) by using landmark-relative submaps. The uncertainty in the measurements and robot pose are accounted for by the use of Bayesian inference techniques during the map update. We demonstrate that an STM map can be used with sensors that generate point measurements, such as stochastic triangular mesh (LiDAR) sensors and stereo cameras. We show that an STM map is a more accurate model than the only comparable online surface mapping technique – a standard elevation map – and we also provide qualitative results on practical datasets.

Acquiring the relative pose between an uncooperative target and a chaser satellite poses a unique problem for any active debris removal mission. A lack of prior knowledge regarding the target's motion, mass distribution and shape limits the possibilities for accurately tracking the target. In this paper, a stereo-camera pair, mounted on a chaser satellite, is used to extract unique features on the surface of an unknown, uncooperative target using the scale invariant feature transform (SIFT). The features are used as measurement input to an extended Kalman Filter (EKF) that makes use of the simultaneous localisation and mapping (SLAM) approach. The orientation and position of the target relative to the chaser is estimated, along with the target's linear and angular velocities. This motion is estimated relative to the camera reference frame (CRF) while the system simultaneously calculates the shape and size of the target. A simulation environment is created to test and verify the estimation algorithm. The integration of the feature extractor with the EKF is tested using real camera data. A laboratory experiment was conducted to capture image sequences of a moving target. A number of practical considerations is highlighted when such a system is to be applied to a mission. Results show that stereo cameras, along with the EKF-SLAM approach is a viable method for fully autonomous space debris mitigation systems.

Each year, latest state of the art technologies and algorithms arise that claim and prove to-shine their predecessors. One such algorithm is the Colour-based Retina Key-point (CREAK) descriptor, which is based on the FAST Retina Key-point (FREAK) descriptor with the included functionality of considering colour information in its Key-point description. This paper explores the implementation of CREAK in a “real-time” visual odometry application by means of a comparative study of the more well-known FREAK algorithm. Although FREAK achieved more accurate odometry when key-points were abundant, this proved to be too computationally expensive. CREAK on the other hand outperformed FREAK when key-points are scarce due to its lack of false-positive matches.

Accurate sensor noise propagation is critical for many computer vision and robotic applications. Several probabilistic computer vision techniques require estimates of sensor noise after it has been propagated through one or many non-linear transformations. We investigate the unscented transform as an alternative to the standard linearisation technique for uncertainty propagation in a computer vision framework. An evaluation is performed using synthetic data for two common computer vision sensors, an RGB-D sensor and stereo camera pair. The unscented transform is shown to outperform linearisation when used to estimate distributions of reconstructed, 3D points from image features. Experimental results also indicate that the unscented transform is a viable replacement for linearisation when used in a probabilistic visual odometry framework.

Detection and tracking of moving objects (DATMO) is essential for autonomous navigation systems operating in general environments. Dynamic objects must be identified, localized, and their future positions predicted to assist in decision making regarding path planning and collision avoidance. To this end, we combine information from a short range frequency modulated continuous wave (FWCW) radar and stereo vision cameras, gathered from a moving vehicle. We extract measurements from both the radar and vision subsystems, using two-dimensional Fourier analysis and sparse feature detection respectively. A segmentation of moving objects is obtained by a hierarchical clustering process, on data composed of image feature tracks and Gaussian mixtures originating from radar-based state estimation. Segmented objects are ultimately tracked using a Gaussian inverse Wishart probability hypothesis density filter (GIW-PHD). Test results on real world data suggest the novel combination of radar and vision data within the PHD filtering framework to be a viable candidate for a DATMO system.

In this paper we propose a method of extrinsically calibrating two of the most commonly used sensors in autonomous robotic systems: a 2-D lidar sensor and camera. Using a tri-planar calibration target, our method extracts rich point and planar feature information from the measurements of both sensors. We approach the problem of estimating the extrinsic paramaters using optimisation, and by minimising the direct 3-D point-plane correspondences between these features, are able to accurately estimate the parameters. Our method is shown to be visually accurate and robust.

This paper presents a robust, sampling-based path-planning algorithm for commercial airliners that simultaneously performs collision avoidance both with intruder aircraft and terrain. The current resolution systems implemented on commercial airliners are effective and efficient, but have certain limitations; the algorithm proposed in this paper attempts to rectify some of these. Recent advances in automatic dependent surveillance-broadcast (ADS-B) and GPS technology provides the information required to resolve complex conflict scenarios that simultaneously involve multiple intruder aircraft and terrain. The proposed algorithm applies an incremental sampling-based technique to determine a safe path quickly and reliably. As the number of samples increases, the algorithm strives towards an optimal solution; this results in a safe, near-optimal path that avoids the conflict region. Simulation results show that the proposed algorithm is able to successfully resolve various conflict scenarios, including the generic two aircraft scenario, terrain only scenario and a two aircraft with terrain scenario. A statistical analysis of the simulation results shows that the algorithm finds near-optimal paths quickly and reliably.