Autonomous Mobile Systems Research Articles

On the Variability of the Circulation and Water Mass Properties in the Eastern Levantine Sea between September 2016–August 2017

The surface circulation and the thermohaline properties of the water masses of the eastern Levantine Sea (Mediterranean Sea) were monitored with mobile autonomous systems (surface drifters and gliders) during the period September 2016–August 2017. The drifters provided data for more than a year and revealed complex circulation features at scales ranging from the basin scale to the sub-mesoscale. Three drifters were captured in a semi-permanent gyre (Cyprus Eddy) allowing a quantitative study of its kinematics. During the experiment, three gliders were operated, in two different periods: September to December 2016 and February to March 2017. The autonomous instruments crossed the prevailing sub-basin structures several times. The collected in-situ observations were analyzed and interpreted in concert with remote sensing products (sea surface temperature and altimetry). The evolution of some of the prevailing features confirmed the complexity of the circulation of the basin. The Cyprus Eddy is the most persistent anticyclone, moving its geographical position and sometimes merging with the North Shikmona Eddy in a bigger structure. The gliders sampled this wide anticyclonic feature revealing its vertical structure in the two different periods. In fall, in stratified conditions, a high salinity core is evident below the thermocline. The isopycnals are characterized by an upward bending over the high salinity lens and a downward bending below it, typical of an anticyclonic modewater eddy. In winter, the core disappears following the vertical mixing that, homogenizes the upper Cyprus Eddy water down to 300 m.

A study on Web service supporting mobility of users using ICT-based autonomous feedback knowledge information

The 4th Industrial Revolution is influencing all the industries with the convergence and development of information communication technology. Particularly, the demand for “building a connected, more intelligent society” is amplified against the IoT (Internet of Things) and Web service area. The key of the 4th Industrial Revolution is to aim at optimized service by building the mutual communication system between the device and device and between the device and the person through the information and communication technology (ICT)-based IoT. However, since the existing IoT and Web service utilize only the limited form and amount of feedback knowledge information, it only could satisfy the individual service and could not provide integrated customized service demanded by the user. Therefore, in this study, the system that can support the customized mobile Web service utilizing autonomous feedback knowledge information with the existing service support methods is suggested. The information is generated in order to apply the Web service by collecting the data from the ICT environment devices, IoT devices, sensor, and wearable devices. The user can use the service suitable for diverse physical changes and spatial changes made during the movement using the information generated. The ICT-based autonomous feedback knowledge information suggested in this article is generated by collecting the changed data of the user. The generated information is identified through the analysis and the assessment as a stage to apply to the service level, and the identified service level gives feedback to the client by matching with the support service. If the autonomous feedback knowledge information mobile support Web service system suggested in this article is applied, the flexible service search is possible and the service by situation can be provided to the user. In addition, a customized mobile support service can be provided without increasing the complexity of the system design compared with the traditional service search method.

Special Issue on “Autonomous Mobile Systems” in Memory of Professor J. Karl Hedrick

Special Issue on “Autonomous Mobile Systems” in Memory of Professor J. Karl Hedrick

Efficient, Multifidelity Perceptual Representations via Hierarchical Gaussian Mixture Models

This paper presents a probabilistic environment representation that allows efficient high-fidelity modeling and inference toward enabling informed planning (active perception) on a computationally constrained mobile autonomous system. The proposed approach exploits the fact that real-world environments inherently possess structure that introduces dependencies between spatially distinct locations. Gaussian mixture models are employed to capture these structural dependencies and learn a semiparametric, arbitrary resolution spatial representation. A hierarchy of spatial models is proposed to enable a multifidelity representation with the variation in fidelity quantified via information-theoretic measures. Crucially for active perception, the proposed modeling approach enables a distribution over occupancy with an associated measure of uncertainty via incorporation of free space information. Evaluation of the proposed technique via a real-time graphics processing unit based implementation is presented on real-world data sets in diverse environments. The proposed approach is shown to perform favorably as compared to state-of-the-art occupancy mapping techniques in terms of memory footprint, prediction accuracy, and generalizability to structurally diverse environments.

A study of Autonomous Mobile System in Outdoor Environment: Part 59

A study of Autonomous Mobile System in Outdoor Environment: Part 59

Proposal of New Information Processing Model for Implementation of Autonomous Mobile System

Proposal of New Information Processing Model for Implementation of Autonomous Mobile System

Distributed Software Architecture for AGVs for Separation of Dynamic and Static Obstacles

The current trend towards mass customization confronts established mass production systems with specific challenges. In future production layouts will have to provide a high degree of flexibility and changeability in order to be able to respond to the demand for individualized mass products. This leads to a production network in which the work schedule of each product and thus the sequence of the individual production resources are flexibly arranged. A challenge in this scenario is the implementation of material transport between the machines. The use of mobile transport robots to realize a flexible connection of all material sources and sinks proves to be effective. However, the current design forms, as a monolithic AGV system or the use of flexible autonomous units, show specific weaknesses. The paper describes the research approach to establish an intralogistics service-based software architecture as an intralogistics execution system, which contains the elementary functions of a cross-platform use of mobile autonomous systems. One key application is to digitize the workspace of the mobile systems with a distributed system of infrastructural sensors. An embedded unit separates the scenario into static and dynamic obstacles. The method discussed also includes the aggregation of the images with static obstacles to a world model as a basis for global cross-platform routing. Thus, it is possible to include the static obstacles in an early phase of path planning. The researched architecture integrates a routing service, which offers a cross-platform map for each type of transport robots in use. The dynamic obstacles of the respective scenery are tracked directly by the embedded decentralized ceiling sensors. By combining the Hungarian method with a Kalman filter, the trajectory of the detected dynamic obstacles can be determined. Within the discussed method, this information is transmitted directly from the sensor units to the autonomous systems and allows the local path planner to avoid collisions. Within the implementation of both methods in an industrial demonstrator, the functionality of both services could be validated. The necessary real-time capability for a valid collision avoidance was not achieved due to the computing capacities in use.

Vehicle Positioning in GNSS-Deprived Urban Areas by Stereo Visual-Inertial Odometry

Accurate and continuous positioning in global navigation satellite system (GNSS) deprived urban areas is crucial for autonomous vehicles and mobile mapping systems. To achieve this goal, we propose a stereo visual-inertial odometry approach using a multistate constraint Kalman filter (MSCKF). In contrast with the conventional MSCKF, in which an inertial navigation system (INS) propagates the vehicle motion, and then the propagation is corrected by measurements of salient features extracted from images of a single camera, we update the propagation with observations extracted from images of a stereo pair of cameras. This way, additional constraints across the stereo pairs of images are exploited to improve the pose estimation. Experimental results on several KITTI datasets show that the stereo MSCKF outperforms the mono one achieving an average positioning error of 0.9% of the trajectory length compared to 1.3% for the mono approach and 2.5% for INS-only integration. These results show that visual-inertial odometry has a promising potential for vehicle positioning in short periods of GNSS signal outage.

Common Platform for Autonomous Low-Speed Mobility System

Development for safety autonomous mobility navigation system that deal with low speed vehicles, such as wheel chair, guidance robot and transportation robot is introduced. Our R&D policy based on Ethical, Legal, Social and Economic issues for the mobility system is also introduced.

Avoiding Collisions in Cucker--Smale Flocking Models under Group-Hierarchical Multileadership

The Cucker--Smale (C-S) flocking model can be applied to design several mobile autonomous agent systems such as unmanned aerial vehicles and robot systems. In these applications, the problem of collision is usually a nonnegligible issue that should be addressed. In the literature, extra bonding/repulsive forces are added to solve the problem of collision. However, all the known results are limited to the symmetric coupling topology. In this paper, we address this problem under a group-hierarchical multileadership consisting of symmetric coupling topologies and the hierarchical leadership. We obtain the following three main results. First, we show that the augmented C-S model achieves asymptotic flocking provided, roughly, that the ratio between the rate of bounding force and communication rate is sufficiently large. Second, we prove that, for a given augmented C-S model, there exists a set of initial states, which can be explicitly described, such that their corresponding solutions maintain some minimum distance between them. Finally, we show that, for a given set of initial states, we can control the system so that its corresponding solution is collision free by suitably choosing bounding forces, which are independent of given initial states. This in turn suggests that the corresponding system can be collision free by having tolerance for error in their initial states.

A study of Autonomous Mobile System in Outdoor Environment: Part 58

A study of Autonomous Mobile System in Outdoor Environment: Part 58

Guest Editorial

Guest Editorial

MTRNNを用いたEnd to End Learningによる移動行動学習システムの開発

This paper presents an autonomous mobile system using E2E learning. We examined the autonomous mobile system which could move in real environment for service robot and autonomous mobile vehicles. As a result, we proposed a system using MTRNN for E2E learning method of the autonomous mobile system. In addition, we developed the mobile robot system for the experiments. We suggested the machine learning simulator which could collect enormous learning data at low cost and developed it. By the above, we developed learning system of the autonomous mobile system by E2E learning.

A topological navigation system for indoor environments based on perception events

The aim of the work presented in this article is to develop a navigation system that allows a mobile robot to move autonomously in an indoor environment using perceptions of multiple events. A topological navigation system based on events that imitates human navigation using sensorimotor abilities and sensorial events is presented. The increasing interest in building autonomous mobile systems makes the detection and recognition of perceptions a crucial task. The system proposed can be considered a perceptive navigation system as the navigation process is based on perception and recognition of natural and artificial landmarks, among others. The innovation of this work resides in the use of an integration interface to handle multiple events concurrently, leading to a more complete and advanced navigation system. The developed architecture enhances the integration of new elements due to its modularity and the decoupling between modules. Finally, experiments have been carried out in several mobile robots, and their results show the feasibility of the navigation system proposed and the effectiveness of the sensorial data integration managed as events.

Gyro-based odometry associated with steering characteristics for wheeled mobile robot in rough terrain

Mobile robot used for planetary exploration performs several scientific missions over long distance travel and needs to have a high degree of autonomous mobility system because the communication delay from the Earth impedes its direct teleoperation. Localization of a mobile robot is of particular importance on the autonomous mobility. Classical localization methods such as wheel/visual odometry have been widely investigated and demonstrated, but they possess a well-known trade-off between computational cost and localization accuracy. This paper proposes an accurate gyro-based odometry method for a wheeled mobile robot in rough terrain. The robot in rough terrain is often subject to large wheel slip or vehicle sideslip related with its steering maneuver, and those slips degrade the localization accuracy. The basic approach of the proposed method is to exploit odometry data for the robot distance traveled as well as gyroscope data for the robot heading calculation; however each data-set is weighted in accordance with steering characteristics of a robot in rough terrain. The usefulness of the proposed method is examined through field experiments using a wheeled mobile robot testbed in Martian analog site. The experimental result confirms that the proposed method accurately estimates the robot trajectory.

Object Detection Applied to Indoor Environments for Mobile Robot Navigation.

To move around the environment, human beings depend on sight more than their other senses, because it provides information about the size, shape, color and position of an object. The increasing interest in building autonomous mobile systems makes the detection and recognition of objects in indoor environments a very important and challenging task. In this work, a vision system to detect objects considering usual human environments, able to work on a real mobile robot, is developed. In the proposed system, the classification method used is Support Vector Machine (SVM) and as input to this system, RGB and depth images are used. Different segmentation techniques have been applied to each kind of object. Similarly, two alternatives to extract features of the objects are explored, based on geometric shape descriptors and bag of words. The experimental results have demonstrated the usefulness of the system for the detection and location of the objects in indoor environments. Furthermore, through the comparison of two proposed methods for extracting features, it has been determined which alternative offers better performance. The final results have been obtained taking into account the proposed problem and that the environment has not been changed, that is to say, the environment has not been altered to perform the tests.

Autonomous mobile lock-in thermography system for detecting and quantifying voids in liquefied natural gas cargo tank second barrier

In this study, an autonomous mobile inspection system that can detect and quantify hidden voids in a secondary barrier (triplex layer) of a liquefied natural gas cargo tank is developed using lock-in thermography. The triplex layer is the secondary barrier to prevent gas leaks from a Mark III–type membrane liquefied natural gas carrier cargo tank and consists of three sub-layers: a flexible secondary barrier, a bonding layer, and a rigid secondary barrier. The proposed mobile inspection system consists of a lock-in thermography measurement unit, a mobile inspection unit, and image processing algorithms. First, thermal images are obtained using the lock-in thermography unit as the mobile inspection system maneuvers over triplex layers. Second, the raw thermal images are processed by several image processing techniques to compensate for non-uniform heating, eliminate noise components, and disregard trivial voids in accordance with the current inspection guideline. Third, the void size is more precisely quantified using an empirical mapping function that relates the void size estimated in the previous step to that measured by an independent X-ray test. The contributions of this study include the following: (1) an autonomous mobile inspection system is developed for real-time inspection of the triplex during its installation, significantly saving the inspection cost and time; (2) a suite of image processing techniques is developed, overcoming shortcomings of the existing thermography non-destructive testing techniques; and (3) the sizes as well as the locations of the hidden voids are quantified with high accuracy, reliability, and fast inspection speed.

Road network extraction: a neural-dynamic framework based on deep learning and a finite state machine

Extracting road networks from very-high-resolution (VHR) aerial and satellite imagery has been a long-standing problem. In this article, a neural-dynamic tracking framework is proposed to extract road networks based on deep convolutional neural networks (DNN) and a finite state machine (FSM). Inspired by autonomous mobile systems, the authors train a DNN to recognize the pattern of input data, which is an image patch extracted in a detection window centred at the current location of the tracker. The pattern is predefined according to the environment and associated with the states in the FSM. A vector-guided sampling method is proposed to generate the training data set for the DNN, which extracts massive image-direction pairs from the imagery and existing vector road maps. In the tracking procedure, the size of the detection window is determined by a fusion strategy and the extracted image patches represent the orientation features of the road (local environment) that can be recognized by the trained DNN. The reactive unit in FSM associates states with behaviours of the tracker while continually modifying the orientation to follow the road and generating a sequence of states and locations. In this way, our framework combines the DNN and FSM. DNN acts as a key component to recognize patterns from a complex and changing environment; FSM translates the recognized patterns to states and controls the behaviour of the tracker. The results illustrate that our approach is more accurate and efficient than the traditional ones.

Fabrication and Characterization of All-Solid-state3D Li-Ion Microbatteries Based on TiO2 Nanotubes

The spectacular development of autonomous mobile systems has enhanced the race for the miniaturization of micropower sources. However, the storage performances of devices are seriously compromised when decreasing the size of the components down to the micrometer scale. This is particularly true for the development of all-solid-state microbatteries which are rapidly reaching performance limit in terms of energy and power densities. To overcome these issues, 3D microbatteries have attracted interest while remaining a technological challenge since it requires the marriage between materials science, electrochemistry, and microfabrication processes. In 2009, we proposed to use self-organized TiO2 nanotubes (ntTiO2) as a 3D negative electrode (Figure) for Li-ion microbatteries (Chem. Mater., 2009). This particular geometry is achieved by a simple electrochemical process (anodization) which improves significantly the electrochemical performance of the system. In this study will be proposed an all-solid-state 3D microbattery composed of a polymer electrolyte and a cathode material (LiFePO4) successively deposited at the surface of the nanotubes. While the nanotubes and polymer layers are obtained by electrodeposition process, the cathode material is achieved by cathodic sputtering. We will present in this work, the influence of morphology, structure and chemical composition of the cathode material on the electrochemical performance of the microbatteries in terms of areal capacity, rate capability, and capacity fading. References : [1] B. Ellis, P. Knauth, and T. Djenizian, “Three-dimensional Self-supported Metal Oxides for Advanced Energy Storage”, Adv. Mater., DOI 10.1002/adma.201306126(2014). In press (IF : 14,829) [2] N. Plylahan, M. Letiche, M. Barr, and T. Djenizian, “All-solid-state lithium-ion batteries based on self-supported titania nanotubes”, Electrochem. Commun., 43, 121, 2014 (IF : 4,425) [3] N. Plylahan, N. A. Kyeremateng, M. Eyraud, F. Dumur, H. Martinez, L. Santinacci, P. Knauth, and T. Djenizian, “Highly conformal electrodeposition of copolymer electrolytes into titania nanotubes for 3D Li-ion batteries”, Nanoscale Res. Lett., 7, 349 (2012). (IF : 2,726) [4] F. Vacandio, M. Eyraud, P. Knauth, and T. Djenizian, “Tunable electrical properties of self-organized zirconia nanotubes”, Electrochem. Commun., 13, 1060 (2011). (IF : 4,859) [5] G. Ortiz, I. Hanzu, T. Djenizian, P. Lavela, J. L. Tirado, and P. Knauth, “Alternative Li-ion battery electrode based on self-organized titania nanotubes”, Chem. Mater., 21, 63 (2009).(IF : 7,286) Figure 1

1A1-J10 機能統合による事前計測を必要としない汎用自律ナビゲーション

This paper describes autonomous mobile robots navigation system and automatic large-scale 3D-map construction method. In previous works, most of them are using detailed map, which is created by operator in advance. Prior detailed map is a very effective tool to improve working efficiency, stability of the robot, but increasing the workload of operator is undesirable to generalize the robot in the real world. Furthermore, autonomous mobile robots have to function appropriately even in areas where they have never been before. Thus pre-driving details are not needed in our system. LIDAR is mainly used in our system, in creating local map, avoiding the moving object (Motion Planning), and recognizing own position (Localization) and environment around the robot (Environment Recognition). Moreover, we construct the detailed map automatically by using the autonomous robot driving data to improve accuracy of the navigation system without increasing the workload of operator. The availability of proposed method is confirmed by the experiments in Meiji University.