Ioannis Rekleitis | University of South Carolina (original) (raw)

Papers by Ioannis Rekleitis

Mapping and monitoring underwater environments are topics of progressively increasing importance,... more Mapping and monitoring underwater environments are topics of progressively increasing importance, but they also introduce several new challenges in robotics, due to the unique underwater conditions. Underwater robots operating close to underwater structures should be equipped with robust localization modules, robust navigation pipelines capable of safely navigating the underwater robot by sensing and avoiding obstacles, and collecting the necessary observations of the surroundings. Especially, tasks that require visual inspection executed by autonomous underwater robots are significantly challenging due to the visibility limitations of the underwater domain. We propose a new active perception framework for underwater robots utilizing arbitrary multi-camera configurations, which safely navigates the robot in close proximity to target structures. It also produces motions that encourage the robot to actively track multiple visual objectives, while dealing effectively with limited FOVs ...

2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2019

This paper presents a systematic approach on realtime reconstruction of an underwater environment... more This paper presents a systematic approach on realtime reconstruction of an underwater environment using Sonar, Visual, Inertial, and Depth data. In particular, low lighting conditions, or even complete absence of natural light inside caves, results in strong lighting variations, e.g., the cone of the artificial video light intersecting underwater structures, and the shadow contours. The proposed method utilizes the well defined edges between well lit areas and darkness to provide additional features, resulting into a denser 3D point cloud than the usual point clouds from a visual odometry system. Experimental results in an underwater cave at Ginnie Springs, FL, with a custom-made underwater sensor suite demonstrate the performance of our system. This will enable more robust navigation of autonomous underwater vehicles using the denser 3D point cloud to detect obstacles and achieve higher resolution reconstructions.

In this paper we present an analysis of the positioning uncertainty increase rate for a group of ... more In this paper we present an analysis of the positioning uncertainty increase rate for a group of mobile robots. The simplified version for a group of N robots moving along one dimension is considered. The one dimension restriction permits us to extract an exact expression for the accumulation of positioning uncertainty in a group of robots equipped with proprioceptive (odometric in this case) and exteroceptive (relative distance between robots) sensors. The solution obtained provides insight in the structure of the multirobot localization problem. In addition, it serves both as an approximation and a starting point for examining the more realistic case ofN robots moving on a plane. Our derivation is based on a Kalman filter estimator that combines all measurements from all robots in the group. Furthermore, we analyze the effect of initial uncertainty, number of robots (N ) and sensor noise on the rate of positioning uncertainty increase. The analytical results derived in this paper ...

2017 IEEE International Conference on Robotics and Automation (ICRA), 2017

The importance of communication in many multirobot information-gathering tasks requires the avail... more The importance of communication in many multirobot information-gathering tasks requires the availability of reliable communication maps. These provide estimates of the radio signal strength and can be used to predict the presence of communication links between different locations of the environment. In the problem we consider, a team of mobile robots has to build such maps autonomously in a robot-to-robot communication setting. The solution we propose models the signal's distribution with a Gaussian Process and exploits different online sensing strategies to coordinate and guide the robots during their data acquisition. Our methods show interesting operative insights both in simulations and on real TurtleBot 2 platforms.

The recent success of NASA’s Mars Exploration Rovers has demonstrated the important benefits that... more The recent success of NASA’s Mars Exploration Rovers has demonstrated the important benefits that mobility adds to landed planetary exploration missions. The Canadian Space Agency (CSA) has been conducting research in ground control and in autonomous robotics for several years already. One of the target applications is planetary exploration using mobile platforms. The emphasis of our research program is on reactive on-board autonomy software and long-range rover navigation. This paper describes recent activities of the CSA in this area. Key results are described in the areas of terrain modelling, path planning and rover guidance.

The solution of cooperative localization is of particular importance to teams of aerial or underw... more The solution of cooperative localization is of particular importance to teams of aerial or underwater robots operating in areas devoid of landmarks. The problem becomes harder if the localization system must be low-cost and lightweight enough that only consumer-grade cameras can be used. This paper presents an analytical solution to the six degrees of freedom cooperative localization problem using bearing only measurements. Given two mutually observing robots, each one equipped with a camera and two markers, and given that they each take a picture at the same moment, we can recover the coordinate transformation that expresses the pose of one robot in the frame of reference of the other. The novelty of our approach is the use of two pairs of bearing measurements for the pose estimation instead of using both bearing and range measurements. The accuracy of the results is verified in extensive simulations and in experiments with real hardware. At 6.5 m distance, position was estimated w...

This paper addresses the problem of robotic operations in the presence of adversarial forces. We ... more This paper addresses the problem of robotic operations in the presence of adversarial forces. We presents a complete framework for survey operations: waypoint generation, modelling of forces and tuning the control. In many applications of environmental monitoring, search and exploration, and bathymetric mapping, the vehicle has to traverse in straight lines parallel to each other, ensuring there are no gaps and no redundant coverage. During operations with an Autonomous Surface Vehicle (ASV) however, the presence of wind and/or currents produces external forces acting on the vehicle which quite often divert it from its intended path. Similar issues have been encountered during aerial or underwater operations. By measuring these phenomena, wind and current, and modelling their impact on the vessel, actions can be taken to alleviate their effect and ensure the correct trajectory is followed.

OCEANS 2018 MTS/IEEE Charleston, 2018

Autonomous exploration and rescue vehicles have been gaining wide interest over the past few year... more Autonomous exploration and rescue vehicles have been gaining wide interest over the past few years. Nowadays, demonstrations showed that those vehicles can fly, dive, surf, or drive while carrying out missions autonomously in some specific scenarios. Monitoring vehicles during missions is a crucial and challenging task to avoid the unnecessary cost of losing vehicles or potential accidents. In this paper, we present a cheap yet effective way for monitoring and communicating with autonomous vehicles over long distances by using off-the-shelf 900 MHz modems namely RFD 900+ and high gain antennas. Although the 900 MHz band has been around for over two decades, no complete analysis exists providing guidelines to use off the shelf modems for point-to-point and multi-point communications. Our main contribution is to provide experimental analysis of the communication capabilities in point-to-point and multi-point scenarios in both line of sight (LOS) and non line of sight (NLOS) using an a...

2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

OCEANS 2018 MTS/IEEE Charleston

This paper presents the design, development, and application of a sensor suite, made with the exp... more This paper presents the design, development, and application of a sensor suite, made with the explicit purpose of localizing and mapping in underwater environments. The design objectives of such an underwater sensor rig include simplicity of carrying, ease of operation in different modes, and data collection. The rig is equipped with stereo camera, inertial measurement unit (IMU), mechanical scanning sonar, and depth sensor. The electronics are enclosed in a water-proof PVC tube tested to sixty meters. The contribution of this paper is twofold: first, we open-source the design providing detailed instructions that are made available online; second, we discuss lessons learned as well as some successful applications where the presented sensor suite has been operated by divers.

2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Sep 1, 2017

International Conference on Robotics and Automation (ICRA), 2020

Coral species detection underwater is a challenging problem. There are many cases when even the e... more Coral species detection underwater is a challenging problem. There are many cases when even the experts (marine biologists) fail to recognize corals, hence limiting ground truth annotation for training a robust detection system. Identifying coral species is fundamental for enabling the monitoring of coral reefs, a task currently performed by humans, which can be automated with the use of underwater robots. By employing temporal cues using a tracker on a high confidence prediction by a convolutional neural network-based object detector, we augment the collected dataset for the retraining of the object detector. However, using trackers to extract examples also introduces hard or mislabelled samples, which is counterproductive and will deteriorate the performance of the detector. In this work, we show that employing a simple deep neural network to filter out hard or mislabelled samples can help regulate sample extraction. We empirically evaluate our approach in a coral object dataset, ...

IEEE International Conference on Robotics and Biomimetics (ROBIO), 2018

Monitoring coral reef populations as part of environmental assessment is essential. Recently, man... more Monitoring coral reef populations as part of environmental assessment is essential. Recently, many marine science researchers are employing low-cost and power efficient Autonomous Underwater Vehicles (AUV) to survey coral reefs. While the counting problem, in general, has rich literature, little work has focused on estimating the density of coral population using AUVs. This paper proposes a novel approach to identify, count, and estimate coral populations. A Convolutional Neural Network (CNN) is utilized to detect and identify the different corals, and a tracking mechanism provides a total count for each coral species per transect. Experimental results from an Aqua2 underwater robot and a stereo hand-held camera validated the proposed approach for different image qualities.

OCEANS 2018 MTS/IEEE Charleston, 2018

Classifying coral species from visual data is a challenging task due to significant intra-species... more Classifying coral species from visual data is a challenging task due to significant intra-species variation, high inter-species similarity, inconsistent underwater image clarity, and high dataset imbalance. In addition, point annotation, the labeling method used for coral reef images by marine biologists, is prone to mislabeling. Point annotation also makes existing datasets incompatible with state-of-the-art classification methods which use the bounding box annotation technique. In this paper, we present a novel end-to-end Convolutional Neural Network (CNN) architecture, Multi-Patch Dense Network (MDNet) that can learn to classify coral species from point annotated visual data. The proposed approach utilizes patches of different scale centered on point annotated objects. Furthermore, MDNet utilizes dense connectivity among layers to reduce over-fitting on imbalanced datasets. Experimental results on the Moorea Labeled Coral (MLC) benchmark dataset are presented. The proposed MDNet ...

OCEANS 2016 MTS/IEEE Monterey, 2016

Historical shipwrecks are important for many reasons, including historical, touristic, and enviro... more Historical shipwrecks are important for many reasons, including historical, touristic, and environmental. Currently, limited efforts for constructing accurate models are performed by divers that need to take measurements manually using a grid and measuring tape, or using handheld sensors. A commercial product, Google Street View1, contains underwater panoramas from select locations around the planet including a few shipwrecks, such as the SS Antilla in Aruba and the Yongala at the Great Barrier Reef. However, these panoramas contain no geometric information and thus there are no 3D representations available of these wrecks. This paper provides, first, an evaluation of visual features quality in datasets that span from indoor to underwater ones. Second, by testing some open-source vision-based state estimation packages on different shipwreck datasets, insights on open challenges for shipwrecks mapping are shown. Some good practices for replicable results are also discussed.

2017 International Conference on Unmanned Aircraft Systems (ICUAS), Jun 1, 2017

2018 15th Conference on Computer and Robot Vision (CRV)

Deep Neural Networks (DNN) have gained tremendous popularity over the last years for several comp... more Deep Neural Networks (DNN) have gained tremendous popularity over the last years for several computer vision tasks, including classification and object detection. Such techniques have been able to achieve human-level performance in many tasks and have produced results of unprecedented accuracy. As DNNs have intense computational requirements in the majority of applications, they utilize a cluster of computers or a cutting edge Graphical Processing Unit (GPU), often having excessive power consumption and generating a lot of heat. In many robotics applications, the above requirements prove to be a challenge, as there is limited power on-board and heat dissipation is always a problem. In particular in underwater robotics with limited space, the above two requirements have been proven prohibitive. As first of this kind, this paper aims at analyzing and comparing the performance of several state-of-the-art DNNs on different platforms. With a focus on the underwater domain, the capabilities of the Jetson TX2 from NVIDIA and the Neural Compute Stick from Intel are of particular interest. Experiments on standard datasets show how different platforms are usable on an actual robotic system, providing insights on the current state-of-the-art embedded systems. Based on such results, we propose some guidelines in choosing the appropriate platform and network architecture for a robotic system.

OCEANS 2018 MTS/IEEE Charleston

Distributed Autonomous Robotic Systems

Mapping and monitoring underwater environments are topics of progressively increasing importance,... more Mapping and monitoring underwater environments are topics of progressively increasing importance, but they also introduce several new challenges in robotics, due to the unique underwater conditions. Underwater robots operating close to underwater structures should be equipped with robust localization modules, robust navigation pipelines capable of safely navigating the underwater robot by sensing and avoiding obstacles, and collecting the necessary observations of the surroundings. Especially, tasks that require visual inspection executed by autonomous underwater robots are significantly challenging due to the visibility limitations of the underwater domain. We propose a new active perception framework for underwater robots utilizing arbitrary multi-camera configurations, which safely navigates the robot in close proximity to target structures. It also produces motions that encourage the robot to actively track multiple visual objectives, while dealing effectively with limited FOVs ...

2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2019

This paper presents a systematic approach on realtime reconstruction of an underwater environment... more This paper presents a systematic approach on realtime reconstruction of an underwater environment using Sonar, Visual, Inertial, and Depth data. In particular, low lighting conditions, or even complete absence of natural light inside caves, results in strong lighting variations, e.g., the cone of the artificial video light intersecting underwater structures, and the shadow contours. The proposed method utilizes the well defined edges between well lit areas and darkness to provide additional features, resulting into a denser 3D point cloud than the usual point clouds from a visual odometry system. Experimental results in an underwater cave at Ginnie Springs, FL, with a custom-made underwater sensor suite demonstrate the performance of our system. This will enable more robust navigation of autonomous underwater vehicles using the denser 3D point cloud to detect obstacles and achieve higher resolution reconstructions.

In this paper we present an analysis of the positioning uncertainty increase rate for a group of ... more In this paper we present an analysis of the positioning uncertainty increase rate for a group of mobile robots. The simplified version for a group of N robots moving along one dimension is considered. The one dimension restriction permits us to extract an exact expression for the accumulation of positioning uncertainty in a group of robots equipped with proprioceptive (odometric in this case) and exteroceptive (relative distance between robots) sensors. The solution obtained provides insight in the structure of the multirobot localization problem. In addition, it serves both as an approximation and a starting point for examining the more realistic case ofN robots moving on a plane. Our derivation is based on a Kalman filter estimator that combines all measurements from all robots in the group. Furthermore, we analyze the effect of initial uncertainty, number of robots (N ) and sensor noise on the rate of positioning uncertainty increase. The analytical results derived in this paper ...

2017 IEEE International Conference on Robotics and Automation (ICRA), 2017

The importance of communication in many multirobot information-gathering tasks requires the avail... more The importance of communication in many multirobot information-gathering tasks requires the availability of reliable communication maps. These provide estimates of the radio signal strength and can be used to predict the presence of communication links between different locations of the environment. In the problem we consider, a team of mobile robots has to build such maps autonomously in a robot-to-robot communication setting. The solution we propose models the signal's distribution with a Gaussian Process and exploits different online sensing strategies to coordinate and guide the robots during their data acquisition. Our methods show interesting operative insights both in simulations and on real TurtleBot 2 platforms.

The recent success of NASA’s Mars Exploration Rovers has demonstrated the important benefits that... more The recent success of NASA’s Mars Exploration Rovers has demonstrated the important benefits that mobility adds to landed planetary exploration missions. The Canadian Space Agency (CSA) has been conducting research in ground control and in autonomous robotics for several years already. One of the target applications is planetary exploration using mobile platforms. The emphasis of our research program is on reactive on-board autonomy software and long-range rover navigation. This paper describes recent activities of the CSA in this area. Key results are described in the areas of terrain modelling, path planning and rover guidance.

The solution of cooperative localization is of particular importance to teams of aerial or underw... more The solution of cooperative localization is of particular importance to teams of aerial or underwater robots operating in areas devoid of landmarks. The problem becomes harder if the localization system must be low-cost and lightweight enough that only consumer-grade cameras can be used. This paper presents an analytical solution to the six degrees of freedom cooperative localization problem using bearing only measurements. Given two mutually observing robots, each one equipped with a camera and two markers, and given that they each take a picture at the same moment, we can recover the coordinate transformation that expresses the pose of one robot in the frame of reference of the other. The novelty of our approach is the use of two pairs of bearing measurements for the pose estimation instead of using both bearing and range measurements. The accuracy of the results is verified in extensive simulations and in experiments with real hardware. At 6.5 m distance, position was estimated w...

This paper addresses the problem of robotic operations in the presence of adversarial forces. We ... more This paper addresses the problem of robotic operations in the presence of adversarial forces. We presents a complete framework for survey operations: waypoint generation, modelling of forces and tuning the control. In many applications of environmental monitoring, search and exploration, and bathymetric mapping, the vehicle has to traverse in straight lines parallel to each other, ensuring there are no gaps and no redundant coverage. During operations with an Autonomous Surface Vehicle (ASV) however, the presence of wind and/or currents produces external forces acting on the vehicle which quite often divert it from its intended path. Similar issues have been encountered during aerial or underwater operations. By measuring these phenomena, wind and current, and modelling their impact on the vessel, actions can be taken to alleviate their effect and ensure the correct trajectory is followed.

OCEANS 2018 MTS/IEEE Charleston, 2018

Autonomous exploration and rescue vehicles have been gaining wide interest over the past few year... more Autonomous exploration and rescue vehicles have been gaining wide interest over the past few years. Nowadays, demonstrations showed that those vehicles can fly, dive, surf, or drive while carrying out missions autonomously in some specific scenarios. Monitoring vehicles during missions is a crucial and challenging task to avoid the unnecessary cost of losing vehicles or potential accidents. In this paper, we present a cheap yet effective way for monitoring and communicating with autonomous vehicles over long distances by using off-the-shelf 900 MHz modems namely RFD 900+ and high gain antennas. Although the 900 MHz band has been around for over two decades, no complete analysis exists providing guidelines to use off the shelf modems for point-to-point and multi-point communications. Our main contribution is to provide experimental analysis of the communication capabilities in point-to-point and multi-point scenarios in both line of sight (LOS) and non line of sight (NLOS) using an a...

2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

OCEANS 2018 MTS/IEEE Charleston

This paper presents the design, development, and application of a sensor suite, made with the exp... more This paper presents the design, development, and application of a sensor suite, made with the explicit purpose of localizing and mapping in underwater environments. The design objectives of such an underwater sensor rig include simplicity of carrying, ease of operation in different modes, and data collection. The rig is equipped with stereo camera, inertial measurement unit (IMU), mechanical scanning sonar, and depth sensor. The electronics are enclosed in a water-proof PVC tube tested to sixty meters. The contribution of this paper is twofold: first, we open-source the design providing detailed instructions that are made available online; second, we discuss lessons learned as well as some successful applications where the presented sensor suite has been operated by divers.

2019 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

2017 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), Sep 1, 2017

International Conference on Robotics and Automation (ICRA), 2020

Coral species detection underwater is a challenging problem. There are many cases when even the e... more Coral species detection underwater is a challenging problem. There are many cases when even the experts (marine biologists) fail to recognize corals, hence limiting ground truth annotation for training a robust detection system. Identifying coral species is fundamental for enabling the monitoring of coral reefs, a task currently performed by humans, which can be automated with the use of underwater robots. By employing temporal cues using a tracker on a high confidence prediction by a convolutional neural network-based object detector, we augment the collected dataset for the retraining of the object detector. However, using trackers to extract examples also introduces hard or mislabelled samples, which is counterproductive and will deteriorate the performance of the detector. In this work, we show that employing a simple deep neural network to filter out hard or mislabelled samples can help regulate sample extraction. We empirically evaluate our approach in a coral object dataset, ...

IEEE International Conference on Robotics and Biomimetics (ROBIO), 2018

Monitoring coral reef populations as part of environmental assessment is essential. Recently, man... more Monitoring coral reef populations as part of environmental assessment is essential. Recently, many marine science researchers are employing low-cost and power efficient Autonomous Underwater Vehicles (AUV) to survey coral reefs. While the counting problem, in general, has rich literature, little work has focused on estimating the density of coral population using AUVs. This paper proposes a novel approach to identify, count, and estimate coral populations. A Convolutional Neural Network (CNN) is utilized to detect and identify the different corals, and a tracking mechanism provides a total count for each coral species per transect. Experimental results from an Aqua2 underwater robot and a stereo hand-held camera validated the proposed approach for different image qualities.

OCEANS 2018 MTS/IEEE Charleston, 2018

Classifying coral species from visual data is a challenging task due to significant intra-species... more Classifying coral species from visual data is a challenging task due to significant intra-species variation, high inter-species similarity, inconsistent underwater image clarity, and high dataset imbalance. In addition, point annotation, the labeling method used for coral reef images by marine biologists, is prone to mislabeling. Point annotation also makes existing datasets incompatible with state-of-the-art classification methods which use the bounding box annotation technique. In this paper, we present a novel end-to-end Convolutional Neural Network (CNN) architecture, Multi-Patch Dense Network (MDNet) that can learn to classify coral species from point annotated visual data. The proposed approach utilizes patches of different scale centered on point annotated objects. Furthermore, MDNet utilizes dense connectivity among layers to reduce over-fitting on imbalanced datasets. Experimental results on the Moorea Labeled Coral (MLC) benchmark dataset are presented. The proposed MDNet ...

OCEANS 2016 MTS/IEEE Monterey, 2016

Historical shipwrecks are important for many reasons, including historical, touristic, and enviro... more Historical shipwrecks are important for many reasons, including historical, touristic, and environmental. Currently, limited efforts for constructing accurate models are performed by divers that need to take measurements manually using a grid and measuring tape, or using handheld sensors. A commercial product, Google Street View1, contains underwater panoramas from select locations around the planet including a few shipwrecks, such as the SS Antilla in Aruba and the Yongala at the Great Barrier Reef. However, these panoramas contain no geometric information and thus there are no 3D representations available of these wrecks. This paper provides, first, an evaluation of visual features quality in datasets that span from indoor to underwater ones. Second, by testing some open-source vision-based state estimation packages on different shipwreck datasets, insights on open challenges for shipwrecks mapping are shown. Some good practices for replicable results are also discussed.

2017 International Conference on Unmanned Aircraft Systems (ICUAS), Jun 1, 2017

2018 15th Conference on Computer and Robot Vision (CRV)

Deep Neural Networks (DNN) have gained tremendous popularity over the last years for several comp... more Deep Neural Networks (DNN) have gained tremendous popularity over the last years for several computer vision tasks, including classification and object detection. Such techniques have been able to achieve human-level performance in many tasks and have produced results of unprecedented accuracy. As DNNs have intense computational requirements in the majority of applications, they utilize a cluster of computers or a cutting edge Graphical Processing Unit (GPU), often having excessive power consumption and generating a lot of heat. In many robotics applications, the above requirements prove to be a challenge, as there is limited power on-board and heat dissipation is always a problem. In particular in underwater robotics with limited space, the above two requirements have been proven prohibitive. As first of this kind, this paper aims at analyzing and comparing the performance of several state-of-the-art DNNs on different platforms. With a focus on the underwater domain, the capabilities of the Jetson TX2 from NVIDIA and the Neural Compute Stick from Intel are of particular interest. Experiments on standard datasets show how different platforms are usable on an actual robotic system, providing insights on the current state-of-the-art embedded systems. Based on such results, we propose some guidelines in choosing the appropriate platform and network architecture for a robotic system.

OCEANS 2018 MTS/IEEE Charleston

Distributed Autonomous Robotic Systems