Intelligent mobile robots need self-localization, map generation, and the ability to explore unknown environments autonomously. Probabilistic processing can be applied to overcome the problems of movement uncertainties and measurement errors. Probabilistic robotics and simultaneous localization and mapping (SLAM) technologies are therefore strongly related, and they have been the focus of many studies. As more and more practical applications are found for intelligent mobile robots, such as for autonomous driving and cleaning, the applicability of these techniques has been increasing.

In this special issue, we provide a wide variety of very interesting papers ranging from studies and developments in applied SLAM technologies to fundamental theories for SLAM. There are five academic papers, one each on the following topics: first visit navigation, controls for following rescue clues, indoor localization using magnetic field maps, a new solution for self-localization using downhill simplex method, and object detection for long-term map management through image-based learning. In addition, in the next number, there will be a review paper by Tsukuba University’s Prof. Tsubouchi, who is famous for the Tsukuba Challenge and research related to mobile robotics.

We editors hope this special issue will help readers to develop mobile robots and use SLAM technologies and probabilistic approaches to produce successful applications.

In this study, we propose a navigation system that guides a robot at a location visited for the first time, without developing a map in advance. First, it estimates the position of a path that exists on the local map by matching the metric route information and the local map generated by simultaneous localization and mapping (SLAM); this is achieved by using a particle filter. Then, the robot travels to the destination along the estimated route. In this system, the geometric accuracy of the route information specified in advance and the accuracy of the map generated by SLAM are essential. Furthermore, it is necessary to recognize the traversable area. The experiment performed verifies the matching of the route information and local map. In the autonomous running experiment, we conduct a trial run on a course set up at the University of Tsukuba.

Carrying out firefighting activities at disaster sites is extremely difficult. Therefore, robots that support and enhance these operations are required. In this paper, a crawler robot that tracks the moving path of a firefighter is proposed. It is commonly believed that trained firefighters select the best route; thus, it was assumed that this route is the easiest for the crawler robot as well. Using two 3D light detection and ranging sensors, once the firefighter’s coordinates are detected, the coordinates are combined with 3D simultaneous localization and mapping results, then a target path is generated. The crawler robot follows the path using inverse optimal tracking control. The controller has a stability margin that guarantees robustness, which is an ideal property for disaster response robots used in severe conditions. The results of several experiments show that the proposed system is effective and practical for the crawler robot.

This study proposes an indoor self-localization for the estimation of the position and posture of an instrument using multiple magnetic sensors. First, a magnetic map for the localization is efficiently created using multiple sensors and a local positioning device made from an optical sensor and a gyroscope. For the localization estimating trajectories, the measurement error of the local positioning is corrected by matching it with the magnetic map. Our instrument is composed of six magnetic sensors, and the description of the self-localization details is based on the framework of a particle filter. The experimental results show better indoor path trajectories compared with a raw trajectory without map matching. The accuracy of the instrument using various numbers of magnetic sensors for the estimation is also investigated.

This paper describes a map-matching method which utilizes a downhill simplex method for self-localization estimation of a mobile robot for indoor and outdoor application. Although particle filter is widely established as a method of map-matching, it requires considerable time for recovery when the correct position is unidentifiable. One of the features of the downhill simplex method proposed in this paper is that the search point distribution is wide when it is challenging to determine a point as the correct position. However, it immediately shrinks when the correct position is identified. In this study, it is compared with particle filter and demonstrates the effectiveness of the proposed method through a discussion on the difference between the search methods.

Image change detection is a fundamental problem for robotic map maintenance and long-term map learning. Local feature-based image comparison is one of the most basic schemes for addressing this problem. However, the local-feature approach encounters difficulties when the query and reference images involve different domains (e.g., time of the day, weather, season). In this paper, we address the local-feature approach from the novel perspective of object-level region features. This study is inspired by the recent success of object-level region features in cross-domain visual place recognition (CD-VPR). Unlike the previous contributions of the CD-VPR task, in the cross-domain change detection (CD-CD) tasks, we consider matching a small part (i.e., the change) of the scene and not the entire image, which is considerably more demanding. To address this issue, we explore the use of two independent object proposal techniques: supervised object proposal (e.g., YOLO) and unsupervised object proposal (e.g., BING). We combine these techniques and compute appearance features of their arbitrarily shaped objects by aggregating local features from a deep convolutional neural network (DCN). Experiments using a publicly available cross-season NCLT dataset validate the efficacy of the proposed approach.

Japan’s construction industry has a shortage of workers and skilled operators, and the number of operators of hydraulic excavators has actually been decreasing. In addition, the operation of hydraulic excavators is complicated and non-intuitive, so learning and maintaining the skill requires a considerable amount of time and experience. One technology that could solve this problem is the tele-operated hydraulic excavator, but the most common type of tele-operated system is difficult for ordinary hydraulic excavator operators to use. The introduction of operation simulators would be effective, but simulators have to be large and expensive if the interface is reproduced in life size. In this paper, we propose a cross-platform system for operating not only a tele-operated hydraulic excavator but also a hydraulic excavator operation simulator. The cross-platform system uses a tele-operated cockpit with feedback and an interface that gives the operator the sensation of sitting in an actual hydraulic excavator. The cockpit consists of a three-dimensional visual information system using stereo videos and a head mounted display, a motion simulator seat that reproduces the vibration and tilting of the operator’s seat of a hydraulic excavator, and a lever unit used in actual hydraulic excavators. With the cross-platform system, an actual hydraulic excavator can be remotely controlled, and the same cockpit can also be used as a training simulator. One can therefore use the system as an inexpensive means of acquiring and maintaining operation skills. We have successfully produced this proposed system and checked its functions, confirming that it works practically.

Many transport units for large production devices now incorporate large-sized gantry type linear motor sliders comprising two parallel linear sliders linked by a joint table. This type of linear motor slider develops a unique mechanical distortion, generating a repulsive force between the two axes that can raise the motor output forces higher than their rated limit. A previous study proposed a method to suppress the repulsive force. However, as feedback gains are set high, force references oscillate and the control system becomes unstable. In past study, yawing vibration suppression methods have been proposed. But, we consider that this vibration is not yawing vibration because the force references include same phase vibration with high gains. Therefore, the modal analysis is performed to analyze this vibration. As a result, it was found that the pitching vibration of the slider was greatly affected. This paper considers this vibration phenomenon, and suppression of the vibration by control method which is similar to impedance control is presented. Hence, it is shown that considering multi-degree of freedom vibration which means yawing vibration and pitching vibration included is important in order to control the large-sized gantry type linear motor sliders.

This paper presents an effective mobile three-dimensional (3D) measurement system that can obtain measurements from the inside of large structures such as railway vehicles, elevators, and escalators. In the proposed method, images are acquired by moving measurement equipment composed of a ring laser and two cameras. From the acquired images, accurate cross-sectional shapes, which are obtained via a light-section method by each camera, are integrated into a unified coordinate system using pose estimation based on bundle adjustment. We focus on the method of separately extracting the information necessary for the two processes – the light-section method and pose estimation – from the acquired images. The laser areas used for the light-section method are detected by a bandpass color filter. Further, a new block matching technique is introduced to eliminate the influence of the laser light, which causes incorrect detection of corresponding points. Through an experiment, we confirm the validity of the proposed 3D measurement method.

In this study, passive dynamic control of a manipulator is designed and realized. According to the control strategy, the shift in the position and orientation of an end effector attributable to an external force is regarded as deformation of the robot. The Zener model, known as a standard linear solid model, is used to generate the deformable behavior, which describes the combination of plastic and elastic deformation. Based on the relation analysis between the Zener model and two other deformable models, two types of control methods are proposed in terms of the model’s expression. Physical simulations with a robotic arm are executed to validate the proposed control laws.

This paper proposes a method for prompting drivers’ spatial attention by presenting visual cue in their peripheral visual field. Computer-generated images of forward-facing driving scenes were projected on a screen 6 m wide and 1.8 m high, with a 140° viewing angle. The gaze movement of subjects was measured when hazardous events were presented, such as cardboard boxes collapsing onto the road or a child running out into the road. The task defined for the subjects was to detect visual cue presented in their central visual field while observing the driving scene in front of them. A preceding visual cue was presented in the right and left visual fields, at a visual angle of 10° to 40°, for 1–5 s in advance of the visual cue presented in the center of the visual field. The detection time for the visual cue in the central visual field was then measured. The results of the experiments conducted with six subjects revealed two types of gaze movement patterns with respect to a hazardous event. In one type, the subjects broadly captured the overall scene without shifting their gaze markedly; in the other type, the subjects sequentially scanned the scene and fixed their gaze on the hazardous event when it occurred. The former type tended to be seen in subjects with long driving experience. It was also found that presenting visual cue in the peripheral visual field quickened recognition of the visual cue in the central visual field. By varying the viewing angle at which the preceding cue was presented in the peripheral visual field and the time interval between the presentation of the preceding cue and the detection cue in the central visual field, conditions were found for assisting prompt detection of the latter visual cue.

For warehouse automation, stocking items in storage shelves (stowing stage) and picking them off the shelves (picking stage) should be automated. To achieve this, we propose a Suction Pinching Hand, an unstable-to-stable transformation (UST) strategy, and an entrance passing (EP) strategy. The Suction Pinching Hand has two underactuated fingers and one extendable and foldable suction finger whose fingertip has a suction cup. The suction finger is useful during the picking stage to enter and manipulate various objects in a cluttered narrow space, with such spaces being typical in efficient warehouses. As this manipulation is sometimes unstable, the underactuated fingers are used simultaneously to stabilize the grasp after the suction finger moves the object to a relatively open space. The UST strategy realizes the transition from unstable object manipulation in a narrow space to stabilization in a relatively open space. In the stowing stage, proximity sensors in the underactuated fingers are used to sense when the object is touched, for securing a direct pinching motion executable in the stowing stage. Finally, the EP strategy ensures that the picked object is able to pass through the entrance of the shelf bin and be successfully stowed. Experiments demonstrate both the ability of our robot hand to grasp various objects stably and safely, and the effectiveness of the UST and EP strategies during the picking and stowing stages.

The present study set out to devise a means of designing a spring-balance mechanism (hereinafter, a SBM) based on a general solution. The goal was to apply a general solution as a guideline for identifying the optimum design solution. A SBM is used to reduce the actuator output in the design of a mechanical device. As a result, the effect of gravity can be offset and energy-saving in operation can be realized. To date, however, SBMs have been designed on a trial and error basis, with a reliance on the designer’s knowledge or experience. Thus, in some instances, it may not be possible to identify a design solution that should, in theory, be achievable. To overcome this issue, this study examined the application of a general solution to the design of a SBM. Such a general solution would identify a solution space including viable design solutions. Subsequently, once a design that satisfies the general solution had been identified, a search could be made of all the design solutions within the solution space. First, the sequence whereby the general solution is used to identify a SBM design that satisfies the general solution, thus providing a guideline for the search action, is determined. Herein, the method used to identify this initial SBM design is presented. The devised method is based on the description of an energy-conservation formula that does not involve any trial and error. The proposed method incorporates a conventional method into the design solution, which can be derived when the conventional method cannot identify a design solution. As a result, a design method using a general solution for the design of an SBM is realized.

Most bridges, tunnels, dams, and other infrastructure facilities were constructed in the 1950s to 1960s. At present, aging in these facilities is a social problem. Most of the maintenance check-ups of these infrastructure facilities are manually conducted by close visual inspection, which has limitations. As a result, there have been many proposals on inspection systems for unmanned infrastructure facilities, such as drones and cable robots. However, there are also problems with the use of cable robots and drones, such as restrictions in places that they can be used and effects from the environment. Thus, this study proposes a robotics system for infrastructure inspection that solves such problems. Two robots move on parallel ropes stretching in an H shape, and one robot with a three-dimensional gimbal and camera moves on a rope stretching between those robots. At this time, the three robots are capable of efficient, highly accurate wide-ranging observation by carrying out synchronous position control using wireless communication. The control system has optimum synchronous position capabilities and uses a disturbance observer. Regarding this infrastructure inspection robotics system, this study discusses the development of three mobile robotics systems and the synchronous position control of two robots.

Landslide hazard prediction is a difficult, time-consuming process when traditional methods are used. This paper presents a method that uses machine learning to predict landslide hazard levels automatically. Due to difficulties in obtaining and effectively processing rainfall in landslide hazard prediction, and to the existing limitation in dealing with large-scale data sets in the M-chameleon algorithm, a new method based on an uncertain DM-chameleon algorithm (developed M-chameleon) is proposed to assess the landslide susceptibility model. First, this method designs a new two-phase clustering algorithm based on M-chameleon, which effectively processes large-scale data sets. Second, the new E-H distance formula is designed by combining the Euclidean and Hausdorff distances, and this enables the new method to manage uncertain data effectively. The uncertain data model is presented at the same time to effectively quantify triggering factors. Finally, the model for predicting landslide hazards is constructed and verified using the data from the Baota district of the city of Yan’an, China. The experimental results show that the uncertain DM-chameleon algorithm of machine learning can effectively improve the accuracy of landslide prediction and has high feasibility. Furthermore, the relationships between hazard factors and landslide hazard levels can be extracted based on clustering results.

In clustering problems based on fuzzy c-means (FCM) for uncertain interval data, points within the interval are usually assumed to have uniform distribution, resulting in the difficulty of accurately describing the interval. Furthermore, the clustering results are considerably affected by the initial clustering centers, and the update speed of the membership degree is slow. To address these problems, a new clustering algorithm called uncertain FCM for interval data (EFCM-ID) is presented. On the basis of a quartile, a median quartile-spacing distance measurement for generally distributed interval data based on machine learning is designed to precisely determine these data. Simultaneously, we sample the whole dataset and consider the density centers as the initial clustering centers to increase accuracy. We call this method samplingbased density-center selection (SDCS). To reduce the running time, a new measurement based on competitive-learning theory to update the membership is developed. It accelerates the update speed by different degrees according to value of the membership degree. Experiments conducted on synthetic interval datasets show the feasibility of EFCM-ID.

This paper proposes a design of a multi-master communication platform for mobile distributed systems, which can be applied to distributed systems without concern on the formation. Experiments were conducted with the testing modules, and the result shows that synchronization of test modules can be effectively achieved with the aid of the platform.