Jonathan Aitken | The University of Sheffield (original) (raw)
Papers by Jonathan Aitken
2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Sep 27, 2021
This is a repository copy of Robust top-down and bottom-up visual saliency for mobile robots usin... more This is a repository copy of Robust top-down and bottom-up visual saliency for mobile robots using bio-inspired design principles.
Frontiers in Robotics and AI, Mar 6, 2023
Buried sewer pipe networks present many challenges for robot localization systems, which require ... more Buried sewer pipe networks present many challenges for robot localization systems, which require non-standard solutions due to the unique nature of these environments: they cannot receive signals from global positioning systems (GPS) and can also lack visual features necessary for standard visual odometry algorithms. In this paper, we exploit the fact that pipe joints are equally spaced and develop a robot localization method based on pipe joint detection that operates in one degree-of-freedom along the pipe length. Pipe joints are detected in visual images from an on-board forward facing (electro-optical) camera using a bag-of-keypoints visual categorization algorithm, which is trained offline by unsupervised learning from images of sewer pipe joints. We augment the pipe joint detection algorithm with drift correction using vision-based manhole recognition. We evaluated the approach using real-world data recorded from three sewer pipes (of lengths 30, 50 and 90 m) and benchmarked against a standard method for visual odometry (ORB-SLAM3), which demonstrated that our proposed method operates more robustly and accurately in these feature-sparse pipes: ORB-SLAM3 completely failed on one tested pipe due to a lack of visual features and gave a mean absolute error in localization of approximately 12%–20% on the other pipes (and regularly lost track of features, having to re-initialize multiple times), whilst our method worked successfully on all tested pipes and gave a mean absolute error in localization of approximately 2%–4%. In summary, our results highlight an important trade-off between modern visual odometry algorithms that have potentially high precision and estimate full six degree-of-freedom pose but are potentially fragile in feature sparse pipes, versus simpler, approximate localization methods that operate in one degree-of-freedom along the pipe length that are more robust and can lead to substantial improvements in accuracy.
2021 20th International Conference on Advanced Robotics (ICAR), Dec 6, 2021
This is a repository copy of Assessing the feasibility of monocular visual simultaneous localizat... more This is a repository copy of Assessing the feasibility of monocular visual simultaneous localization and mapping for live sewer pipes : a field robotics study.
2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2021
This is a repository copy of Robust top-down and bottom-up visual saliency for mobile robots usin... more This is a repository copy of Robust top-down and bottom-up visual saliency for mobile robots using bio-inspired design principles.
In this paper we explore the factors and methodologies from a range of disciplines used to invest... more In this paper we explore the factors and methodologies from a range of disciplines used to investigate trust in human-robot interaction (HRI). Our investigation highlights a growing field, which recognises the importance of understanding the deployment of robots in real-world settings, but where a lack of common definitions and experimental clarity impedes the development of a comprehensive framework for investigation. As a result, we propose a bottom-up approach that emphasises context and user perspective as the foundation for future investigations into trust in HRI.
Lecture Notes in Computer Science, 2016
We report on experiences in the development of hybrid autonomous systems where high-level decisio... more We report on experiences in the development of hybrid autonomous systems where high-level decisions are made by a rational agent. This rational agent interacts with other subsystems via an abstraction engine. We describe three systems we have developed using the EASS BDI agent programming language and framework which supports this architecture. As a result of these experiences we recommend changes to the theoretical operational semantics that underpins the EASS framework and present a fourth implementation using the new semantics.
IEEE Robotics and Automation Letters
Software Engineering and Formal Methods
The verification community develops two kinds of verification tools: automatic verifiers and inte... more The verification community develops two kinds of verification tools: automatic verifiers and interactive verifiers. There are many such verifiers available, and there is steady progress in research. However, cooperation between the two kinds of verifiers was not yet addressed in a modular way. Yet, it is imperative for the community to leverage all possibilities, because our society heavily depends on software systems that work correctly. This paper contributes tools and a modular design to address the open problem of insufficient support for cooperation between verification tools. We identify invariants as information that needs to be exchanged in cooperation, and we support translation between two ‘containers’ for invariants: program annotations and correctness witnesses. Using our new building blocks, invariants computed by automatic verifiers can be given to interactive verifiers as annotations in the program, and annotations from the user or interactive verifier can be given to...
Frontiers in Robotics and AI, 2021
Digital twins offer a unique opportunity to design, test, deploy, monitor, and control real-world... more Digital twins offer a unique opportunity to design, test, deploy, monitor, and control real-world robotic processes. In this paper we present a novel, modular digital twinning framework developed for the investigation of safety within collaborative robotic manufacturing processes. The modular architecture supports scalable representations of user-defined cyber-physical environments, and tools for safety analysis and control. This versatile research tool facilitates the creation of mixed environments of Digital Models, Digital Shadows, and Digital Twins, whilst standardising communication and physical system representation across different hardware platforms. The framework is demonstrated as applied to an industrial case-study focused on the safety assurance of a collaborative robotic manufacturing process. We describe the creation of a digital twin scenario, consisting of individual digital twins of entities in the manufacturing case study, and the application of a synthesised safet...
Sensors
Image tracking and retrieval strategies are of vital importance in visual Simultaneous Localizati... more Image tracking and retrieval strategies are of vital importance in visual Simultaneous Localization and Mapping (SLAM) systems. For most state-of-the-art systems, hand-crafted features and bag-of-words (BoW) algorithms are the common solutions. Recent research reports the vulnerability of these traditional algorithms in complex environments. To replace these methods, this work proposes HFNet-SLAM, an accurate and real-time monocular SLAM system built on the ORB-SLAM3 framework incorporated with deep convolutional neural networks (CNNs). This work provides a pipeline of feature extraction, keypoint matching, and loop detection fully based on features from CNNs. The performance of this system has been validated on public datasets against other state-of-the-art algorithms. The results reveal that the HFNet-SLAM achieves the lowest errors among systems available in the literature. Notably, the HFNet-SLAM obtains an average accuracy of 2.8 cm in EuRoC dataset in pure visual configuration...
Abstract. Typical theorem-proving workloads on the Poly/ML runtime may execute for several hours,... more Abstract. Typical theorem-proving workloads on the Poly/ML runtime may execute for several hours, occupying multi-gigabyte heaps. The run-time heap size may be fixed at execution startup time, or it may be al-lowed to vary dynamically. To date, runtime heap size growth has been implemented using simple hard-coded heuristics. In this position paper, we argue that a mathematically rigorous approach to heap sizing, based on control theory, is more appropriate.
Lecture Notes in Computer Science, 2016
OpenCV is a commonly used computer vision library containing a wide variety of algorithms for the... more OpenCV is a commonly used computer vision library containing a wide variety of algorithms for the AI community. This paper uses deep parameter optimisation to investigate improvements to face detection using the Viola-Jones algorithm in OpenCV, allowing a tradeoff between execution time and classification accuracy. Our results show that execution time can be decreased by 48% if a 1.80% classification inaccuracy is permitted (compared to 1.04% classification inaccuracy of the original, unmodified algorithm). Further execution time savings are possible depending on the degree of inaccuracy deemed acceptable by the user.
IFAC Proceedings Volumes, 2014
This paper lays down the foundations of developing a reconfigurable control system within the Rob... more This paper lays down the foundations of developing a reconfigurable control system within the Robot Operating System (ROS) for autonomous robots. The essential components of robots are programmed under a ROS system. A formal model is defined as a tripartite graph to represent the robots functional architecture. ROS systems are then generalised to component libraries for any ROS architecture and an abstract model as a "system graph" is introduced. Orthogonality of a library and a system graph is defined and redundancy levels of robot components are studied for maintaining full functionality of the robot by automated reconfiguration in face of hardware malfunction. This allows AI planning tools, such as Planning Domain Definition Language (PDDL), to compute permissible reconfigurations. We present an example of a pair of robotic arms which requires reconfiguration of the underlying control system in order to retain the capability to carry out a task.
ACM SIGPLAN Notices, 2013
We propose a new, principled approach to adaptive heap sizing based on control theory. We review ... more We propose a new, principled approach to adaptive heap sizing based on control theory. We review current state-of-the-art heap sizing mechanisms, as deployed in Jikes RVM and HotSpot. We then formulate heap sizing as a control problem, apply and tune a standard controller algorithm, and evaluate its performance on a set of well-known benchmarks. We find our controller adapts the heap size more responsively than existing mechanisms. This responsiveness allows tighter virtual machine memory footprints while preserving target application throughput, which is ideal for both embedded and utility computing domains. In short, we argue that formal, systematic approaches to memory management should be replacing ad-hoc heuristics as the discipline matures. Control-theoretic heap sizing is one such systematic approach.
Lecture Notes in Computer Science, 2014
The use of general descriptive names, registered names, trademarks, service marks, etc. in this p... more The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. While the advice and information in this book are believed to be true and accurate at the date of publication, neither the authors nor the editors nor the publisher can accept any legal responsibility for any errors or omissions that may be made. The publisher makes no warranty, express or implied, with respect to the material contained herein.
Empirical Software Engineering, 2014
Software systems that rely on ad-hoc networks are becoming increasingly complex and increasingly ... more Software systems that rely on ad-hoc networks are becoming increasingly complex and increasingly prevalent. Some of these systems provide vital functionality to military operations, emergency services and disaster relief; such systems may have significant impact on the safety of people involved in those operations. It is therefore important that those networks support critical software requirements, including those for latency of packet transfer. If a network ceases to meet the software's requirements (e.g. due to a link failure) then engineers must be able to understand it well enough to reconfigure the network and restore it to a requirement-satisfying state. Given a complex network, it is difficult for a human to do this under time pressure. In this paper we present a search-based tool which takes a network defined using the Network Description Language (NDL), annotated with a set of network-hosted applications and a set of latency requirements between each. We then evolve variants of the network configuration which meet the requirements and are robust to single link failures. We use network calculus tools to get a fast, conservative evaluation of whether a given network meets its requirements. We demonstrate that this approach is viable, designs networks much faster than a human engineer could, and is superior to a random generate-and-test approach.
Towards Autonomous Robotic Systems, 2021
Visual odometry can be used to estimate the pose of a robot from current and recent video frames.... more Visual odometry can be used to estimate the pose of a robot from current and recent video frames. A problem with these methods is that they drift over time due to the accumulation of estimation errors at each time-step. In this short paper we propose and briefly demonstrate the potential benefit of using prior 2D, top-down map information combined with multiple hypothesis particle filtering to correct visual odometry estimates. The results demonstrate a substantial improvement in robustness and accuracy over the sole use of visual odometry.
UK-RAS Conference: Robots Working For and Among Us Proceedings
This is a repository copy of Graphical Signage Decreases Negative Attitudes towards Robots and Ro... more This is a repository copy of Graphical Signage Decreases Negative Attitudes towards Robots and Robot Anxiety in Human-Robot Co-working.
2015 European Conference on Mobile Robots (ECMR), 2015
Robotic assistants operating in multi-floor buildings are required to use lifts to transition bet... more Robotic assistants operating in multi-floor buildings are required to use lifts to transition between floors. To reduce the need for environments to be tailored to suit robots, and to make robot assistants more applicable, it is desirable that they should make use of existing navigational cues and interfaces designed for human users. In this paper, we examine the scenario whereby a guide robot uses a lift to transition between floors in a building. We describe an experiment into combining multiple data sources, available to a typical robot with simple sensors, to determine which floor of the building it is on. We show the robustness of this approach to realistic scenarios in a busy working environment.
2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Sep 27, 2021
This is a repository copy of Robust top-down and bottom-up visual saliency for mobile robots usin... more This is a repository copy of Robust top-down and bottom-up visual saliency for mobile robots using bio-inspired design principles.
Frontiers in Robotics and AI, Mar 6, 2023
Buried sewer pipe networks present many challenges for robot localization systems, which require ... more Buried sewer pipe networks present many challenges for robot localization systems, which require non-standard solutions due to the unique nature of these environments: they cannot receive signals from global positioning systems (GPS) and can also lack visual features necessary for standard visual odometry algorithms. In this paper, we exploit the fact that pipe joints are equally spaced and develop a robot localization method based on pipe joint detection that operates in one degree-of-freedom along the pipe length. Pipe joints are detected in visual images from an on-board forward facing (electro-optical) camera using a bag-of-keypoints visual categorization algorithm, which is trained offline by unsupervised learning from images of sewer pipe joints. We augment the pipe joint detection algorithm with drift correction using vision-based manhole recognition. We evaluated the approach using real-world data recorded from three sewer pipes (of lengths 30, 50 and 90 m) and benchmarked against a standard method for visual odometry (ORB-SLAM3), which demonstrated that our proposed method operates more robustly and accurately in these feature-sparse pipes: ORB-SLAM3 completely failed on one tested pipe due to a lack of visual features and gave a mean absolute error in localization of approximately 12%–20% on the other pipes (and regularly lost track of features, having to re-initialize multiple times), whilst our method worked successfully on all tested pipes and gave a mean absolute error in localization of approximately 2%–4%. In summary, our results highlight an important trade-off between modern visual odometry algorithms that have potentially high precision and estimate full six degree-of-freedom pose but are potentially fragile in feature sparse pipes, versus simpler, approximate localization methods that operate in one degree-of-freedom along the pipe length that are more robust and can lead to substantial improvements in accuracy.
2021 20th International Conference on Advanced Robotics (ICAR), Dec 6, 2021
This is a repository copy of Assessing the feasibility of monocular visual simultaneous localizat... more This is a repository copy of Assessing the feasibility of monocular visual simultaneous localization and mapping for live sewer pipes : a field robotics study.
2021 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2021
This is a repository copy of Robust top-down and bottom-up visual saliency for mobile robots usin... more This is a repository copy of Robust top-down and bottom-up visual saliency for mobile robots using bio-inspired design principles.
In this paper we explore the factors and methodologies from a range of disciplines used to invest... more In this paper we explore the factors and methodologies from a range of disciplines used to investigate trust in human-robot interaction (HRI). Our investigation highlights a growing field, which recognises the importance of understanding the deployment of robots in real-world settings, but where a lack of common definitions and experimental clarity impedes the development of a comprehensive framework for investigation. As a result, we propose a bottom-up approach that emphasises context and user perspective as the foundation for future investigations into trust in HRI.
Lecture Notes in Computer Science, 2016
We report on experiences in the development of hybrid autonomous systems where high-level decisio... more We report on experiences in the development of hybrid autonomous systems where high-level decisions are made by a rational agent. This rational agent interacts with other subsystems via an abstraction engine. We describe three systems we have developed using the EASS BDI agent programming language and framework which supports this architecture. As a result of these experiences we recommend changes to the theoretical operational semantics that underpins the EASS framework and present a fourth implementation using the new semantics.
IEEE Robotics and Automation Letters
Software Engineering and Formal Methods
The verification community develops two kinds of verification tools: automatic verifiers and inte... more The verification community develops two kinds of verification tools: automatic verifiers and interactive verifiers. There are many such verifiers available, and there is steady progress in research. However, cooperation between the two kinds of verifiers was not yet addressed in a modular way. Yet, it is imperative for the community to leverage all possibilities, because our society heavily depends on software systems that work correctly. This paper contributes tools and a modular design to address the open problem of insufficient support for cooperation between verification tools. We identify invariants as information that needs to be exchanged in cooperation, and we support translation between two ‘containers’ for invariants: program annotations and correctness witnesses. Using our new building blocks, invariants computed by automatic verifiers can be given to interactive verifiers as annotations in the program, and annotations from the user or interactive verifier can be given to...
Frontiers in Robotics and AI, 2021
Digital twins offer a unique opportunity to design, test, deploy, monitor, and control real-world... more Digital twins offer a unique opportunity to design, test, deploy, monitor, and control real-world robotic processes. In this paper we present a novel, modular digital twinning framework developed for the investigation of safety within collaborative robotic manufacturing processes. The modular architecture supports scalable representations of user-defined cyber-physical environments, and tools for safety analysis and control. This versatile research tool facilitates the creation of mixed environments of Digital Models, Digital Shadows, and Digital Twins, whilst standardising communication and physical system representation across different hardware platforms. The framework is demonstrated as applied to an industrial case-study focused on the safety assurance of a collaborative robotic manufacturing process. We describe the creation of a digital twin scenario, consisting of individual digital twins of entities in the manufacturing case study, and the application of a synthesised safet...
Sensors
Image tracking and retrieval strategies are of vital importance in visual Simultaneous Localizati... more Image tracking and retrieval strategies are of vital importance in visual Simultaneous Localization and Mapping (SLAM) systems. For most state-of-the-art systems, hand-crafted features and bag-of-words (BoW) algorithms are the common solutions. Recent research reports the vulnerability of these traditional algorithms in complex environments. To replace these methods, this work proposes HFNet-SLAM, an accurate and real-time monocular SLAM system built on the ORB-SLAM3 framework incorporated with deep convolutional neural networks (CNNs). This work provides a pipeline of feature extraction, keypoint matching, and loop detection fully based on features from CNNs. The performance of this system has been validated on public datasets against other state-of-the-art algorithms. The results reveal that the HFNet-SLAM achieves the lowest errors among systems available in the literature. Notably, the HFNet-SLAM obtains an average accuracy of 2.8 cm in EuRoC dataset in pure visual configuration...
Abstract. Typical theorem-proving workloads on the Poly/ML runtime may execute for several hours,... more Abstract. Typical theorem-proving workloads on the Poly/ML runtime may execute for several hours, occupying multi-gigabyte heaps. The run-time heap size may be fixed at execution startup time, or it may be al-lowed to vary dynamically. To date, runtime heap size growth has been implemented using simple hard-coded heuristics. In this position paper, we argue that a mathematically rigorous approach to heap sizing, based on control theory, is more appropriate.
Lecture Notes in Computer Science, 2016
OpenCV is a commonly used computer vision library containing a wide variety of algorithms for the... more OpenCV is a commonly used computer vision library containing a wide variety of algorithms for the AI community. This paper uses deep parameter optimisation to investigate improvements to face detection using the Viola-Jones algorithm in OpenCV, allowing a tradeoff between execution time and classification accuracy. Our results show that execution time can be decreased by 48% if a 1.80% classification inaccuracy is permitted (compared to 1.04% classification inaccuracy of the original, unmodified algorithm). Further execution time savings are possible depending on the degree of inaccuracy deemed acceptable by the user.
IFAC Proceedings Volumes, 2014
This paper lays down the foundations of developing a reconfigurable control system within the Rob... more This paper lays down the foundations of developing a reconfigurable control system within the Robot Operating System (ROS) for autonomous robots. The essential components of robots are programmed under a ROS system. A formal model is defined as a tripartite graph to represent the robots functional architecture. ROS systems are then generalised to component libraries for any ROS architecture and an abstract model as a "system graph" is introduced. Orthogonality of a library and a system graph is defined and redundancy levels of robot components are studied for maintaining full functionality of the robot by automated reconfiguration in face of hardware malfunction. This allows AI planning tools, such as Planning Domain Definition Language (PDDL), to compute permissible reconfigurations. We present an example of a pair of robotic arms which requires reconfiguration of the underlying control system in order to retain the capability to carry out a task.
ACM SIGPLAN Notices, 2013
We propose a new, principled approach to adaptive heap sizing based on control theory. We review ... more We propose a new, principled approach to adaptive heap sizing based on control theory. We review current state-of-the-art heap sizing mechanisms, as deployed in Jikes RVM and HotSpot. We then formulate heap sizing as a control problem, apply and tune a standard controller algorithm, and evaluate its performance on a set of well-known benchmarks. We find our controller adapts the heap size more responsively than existing mechanisms. This responsiveness allows tighter virtual machine memory footprints while preserving target application throughput, which is ideal for both embedded and utility computing domains. In short, we argue that formal, systematic approaches to memory management should be replacing ad-hoc heuristics as the discipline matures. Control-theoretic heap sizing is one such systematic approach.
Lecture Notes in Computer Science, 2014
The use of general descriptive names, registered names, trademarks, service marks, etc. in this p... more The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. While the advice and information in this book are believed to be true and accurate at the date of publication, neither the authors nor the editors nor the publisher can accept any legal responsibility for any errors or omissions that may be made. The publisher makes no warranty, express or implied, with respect to the material contained herein.
Empirical Software Engineering, 2014
Software systems that rely on ad-hoc networks are becoming increasingly complex and increasingly ... more Software systems that rely on ad-hoc networks are becoming increasingly complex and increasingly prevalent. Some of these systems provide vital functionality to military operations, emergency services and disaster relief; such systems may have significant impact on the safety of people involved in those operations. It is therefore important that those networks support critical software requirements, including those for latency of packet transfer. If a network ceases to meet the software's requirements (e.g. due to a link failure) then engineers must be able to understand it well enough to reconfigure the network and restore it to a requirement-satisfying state. Given a complex network, it is difficult for a human to do this under time pressure. In this paper we present a search-based tool which takes a network defined using the Network Description Language (NDL), annotated with a set of network-hosted applications and a set of latency requirements between each. We then evolve variants of the network configuration which meet the requirements and are robust to single link failures. We use network calculus tools to get a fast, conservative evaluation of whether a given network meets its requirements. We demonstrate that this approach is viable, designs networks much faster than a human engineer could, and is superior to a random generate-and-test approach.
Towards Autonomous Robotic Systems, 2021
Visual odometry can be used to estimate the pose of a robot from current and recent video frames.... more Visual odometry can be used to estimate the pose of a robot from current and recent video frames. A problem with these methods is that they drift over time due to the accumulation of estimation errors at each time-step. In this short paper we propose and briefly demonstrate the potential benefit of using prior 2D, top-down map information combined with multiple hypothesis particle filtering to correct visual odometry estimates. The results demonstrate a substantial improvement in robustness and accuracy over the sole use of visual odometry.
UK-RAS Conference: Robots Working For and Among Us Proceedings
This is a repository copy of Graphical Signage Decreases Negative Attitudes towards Robots and Ro... more This is a repository copy of Graphical Signage Decreases Negative Attitudes towards Robots and Robot Anxiety in Human-Robot Co-working.
2015 European Conference on Mobile Robots (ECMR), 2015
Robotic assistants operating in multi-floor buildings are required to use lifts to transition bet... more Robotic assistants operating in multi-floor buildings are required to use lifts to transition between floors. To reduce the need for environments to be tailored to suit robots, and to make robot assistants more applicable, it is desirable that they should make use of existing navigational cues and interfaces designed for human users. In this paper, we examine the scenario whereby a guide robot uses a lift to transition between floors in a building. We describe an experiment into combining multiple data sources, available to a typical robot with simple sensors, to determine which floor of the building it is on. We show the robustness of this approach to realistic scenarios in a busy working environment.