The challenges facing society such as a declining birth rate and aging population, disaster response, labor shortages, and infection control are becoming increasingly complex. To address these diverse challenges, it is conceivable that robots with various functions could be networked to provide solutions. By integrating not only multifunctional robots, but also single-function robots, environmental sensors, and IoT devices into a system, a wider range of issues can be addressed. Social issues differ across regions, and a variety of applications can be developed by addressing these issues. Furthermore, if such a system is implemented into society, it could lead to the creation of new markets and contribute to social revitalization. Here, such an application of robotics is referred to as “community service robotics.” In fact, research on robotic collaboration and the development of platforms for this purpose are already underway. This article introduces trends in robotic collaboration networks and our efforts to tackle these societal challenges.

In order to research and develop flight system technologies for next-generation urban air mobility (UAM) such as drones, as well as to foster young human resources in related fields, the Center for Aerial Intelligent Vehicle (CAIV) was established on October 1, 2019 (https://caiv.chiba-u.jp/index.html).

The development of industrial drones is expected to expand in the future. In particular, applications such as delivery services are expected to drive demand for low-altitude flights. As drones become increasingly integrated into human environments, various challenges may arise, among which noise reduction represents a critical concern. To address this issue, a propeller design incorporating an enlarged blade area was investigated, wherein additional curved plates were attached to the trailing edges of the propellers to mitigate noise. The planar shape of the curved plates was designed using several parameters in a non-dimensional coordinate system. We introduced a relationship between the target noise reduction level and the second moment of area, subsequently used to scale the designed structure. The results of simultaneous force and noise measurements of a single propeller yielded significant noise reduction for all propellers with enlarged blade area. Strong correlations were observed between noise levels and the attachment length in the spanwise direction, as well as the inclination angle at the wing root side in the designed geometry. The in-flight test was performed using a six-rotor drone, achieving an overall noise reduction of 8.1 dBA compared to standard propellers, with significant reductions in the tonal noise specific to drones and in the high-frequency broadband noise, which causes discomfort to humans. These findings are expected to significantly inform future design guidelines for noise reduction in drones.

Multicopters for facility inspection and logistics applications are often required to fly very close to structures. However, it is known that the behavior of multicopters can be unstable in the vicinity of structures. In this study, the ground effect when there is an inclined ground plane below the quadcopter was investigated experimentally. The results show that imbalanced aerodynamic forces that can disturb the flight are generated by different mechanisms depending on the propeller configuration and flight altitude: when flying near an inclined plane with the thrust force of the four rotors maintained constant, the torque acting on the rotors changes depending on the inclination angle of the ground plane. The change in torque is influenced by the distance between the rotor and inclined plane and by the distance between neighboring rotors. It is suggested that an imbalance of yawing moments may occur in the quadcopter, which makes flight unstable, depending on the azimuth angle of the quadcopter with respect to the inclined plane. This study also examined the ground effect of ducted rotors above the inclined ground plane and observed different characteristics from the case without ducts.

In this study, we aim to develop a navigation and control system that enables drones to fly autonomously with high accuracy in indoor environments, including narrow aisles. First, we propose an EKF-based vision-aided inertial navigation system with altimeter (VINS-ALT), which combines monocular-SLAM results with data from the IMU and altimeter. In addition, a detection and correction system is designed to reduce altimeter errors caused by changes in ground surface characteristics, such as steps and slopes. Furthermore, a flight control system that achieves both trajectory tracking performance and robustness is developed. Finally, the effectiveness of the entire system is validated through autonomous flight control experiments in an indoor environment.

In robotics, realizing 6D tracking, that is, tracking the three-dimensional position and orientation of moving objects in real-time, is a significant challenge. In this study, we extend fast pixel-wise posterior 3D (Fast-PWP3D), a method for estimating the position and posture simultaneously. The extended method involves the segmentation of the target region from multiple camera images using a 3D target model. The energy function was modified to handle multiviewpoint posture estimation. Consequently, the accuracy of target estimation and robustness against occlusion were improved.

This paper presents a novel shared control algorithm based on quasi-linear parameter varying model predictive control (qLPV-MPC) for the purpose of preventing vehicle collisions during lane-changing maneuvers and maintaining the vehicle’s position within the lane at all times. The algorithm optimizes the control solution by transforming it into a quadratic programming problem, thus enhancing computational efficiency. By sharing control over both vehicle acceleration and steering wheel torque, the algorithm provides coupled lateral and longitudinal dynamic control, allowing simultaneous acceleration/deceleration adjustments and steering torque application for effective collision avoidance. Notably, the algorithm operates independently of a reference path, allowing the driver to retain primary control over vehicle movement, while the controller intervenes only in high-risk situations. Simulation scenarios, including rear-end, side-impact, and multi-vehicle collision cases, are employed to validate the algorithm’s effectiveness in various collision avoidance contexts.

Obtaining object-detection results through multi-viewpoint systems is a promising approach for gathering the surrounding information for mobile robots. In this study, we realized a low delay real-time multi-viewpoint object detection for an omnidirectional scanning shooting system by proposing a dual-stage detection approach with intermediate frames. This approach can synchronously search for valid viewpoints containing the desired objects, and perform low-update-delay real-time object detection on selected viewpoints with intermediate frames. We generated panoramic detection frames as real-time system output by projecting the individual viewpoint detection results. In an outdoor experimental scenario with 10 fps omnidirectional scanning shooting accommodating up to 40 viewpoints, our approach generated 20 fps real-time panoramic detection frames presenting the movement of the observed targets.

To realize the turning motion of the micro aerial vehicle, we developed a mechanism that reproduces the lead-lag motion of a butterfly. The effectiveness of this mechanism was verified through flight experiments and numerical simulations of a glider equipped with this mechanism. Butterflies turn using the lead-lag motion, which involves moving the left and right wings back and forth. In this study, we developed a mechanism to achieve this lead-lag motion using a shape memory alloy actuator based on our butterfly-type flapping robot. The mechanism was implemented in a butterfly-shaped glider, and flight experiments and numerical simulation analysis with computational fluid dynamics were conducted. The results of 3D motion analysis using multiple high-speed cameras revealed that the developed lead-lag mechanism allows the left and right wings to have different lead-lag angles during glide flight. This mechanism was confirmed to enable the right turn by increasing the lead-lag angle of the left wing as well as the left turn by increasing the lead-lag angle of the right wing. Numerical simulation analysis demonstrated that changes in the left and right lead-lag angles generated yawing and rolling moments, respectively, and that this produced turning motion with a change in the yaw angle.

Early detection of frailty is crucial for older adults’ health, particularly as lower limb muscle strength declines rapidly in the early stages of frailty. The five times sit-to-stand (FTSTS) test is commonly used to evaluate this strength. However, performing the FTSTS test at maximum effort poses risks, requiring assistance to prevent falls and increasing the likelihood of excessive strain on lower limb joints. Therefore, safer, less strenuous methods for assessing muscle strength are needed to minimize these risks. Therefore, to measure lower limb function in a safe environment, it is necessary to implement a version of the FTSTS test that does not require maximum effort speed, allowing the subject to perform at their own pace with minimal physical strain, and to develop a method to automate this measurement. This study proposes a method to assess frailty and fall risk by conducting the FTSTS test with minimal exertion and analyzing acceleration data. A system using a wearable accelerometer with wireless communication capabilities was developed to automatically measure the acceleration during the sit-to-stand movements. The J-CHS (Japanese version of the Cardiovascular Health Study) criteria for frailty diagnosis and the Kihon Checklist for fall risk assessment were conducted, and the results were used as a control for categorizing subjects based on frailty and fall risk. Clinical trials of the FTSTS test were conducted on older adults, and frailty and fall risk were assessed based on the acceleration data obtained to verify the effectiveness of the proposed method.

In this study, we consider a musculoskeletal system in which muscle tension is employed to control a linkage structure consisting of mechanical elements that correspond to a skeletal structure. Because muscles can only transmit force in the tensile direction, more muscles than joint degrees of freedom are needed to achieve control. The musculoskeletal potential method uses the potential field generated by the internal forces between muscles arising from this redundancy. In this method, constant muscle tension balanced at the desired posture is used as the step input to carry out feed forward control, in which sensory feedback and complex real-time computations are not required. However, the potential field generated by the muscle internal forces is highly dependent on muscular arrangement, some of which may fail to converge to the desired posture. For a musculoskeletal system with a specific structure consisting of two joints and six muscles, a previous study identified the conditions of the muscular arrangement required to achieve convergence to the desired posture. A further study analyzed the conditions of convergence for a more general multi-articular multimuscular system with several joints and muscles, although the joints were limited to a single degree of freedom and muscles to mono- and bi-articular ones, whose lengths were dependent on at most two adjacent joints. In this study, we consider cases consisting of joints with multiple degrees of freedom and muscles whose lengths depend on the angles of three or more joints. We mathematically analyze the conditions of convergence in the musculoskeletal potential method and verify the findings by simulation.

Force sensors require different measurement ranges and sensitivities depending on the operating environment. Therefore, we developed a force sensor with both variable sensitivity and measurement range based on structural modification induced by simple bending. In our previous study, we changed the distance between the force application point and the detection area. However, it was necessary to change the position and direction of the measured force. Therefore, in the present study, we changed the cross-sectional shape of the sensor. First, we introduce the theoretical basis for the proposed force sensor. We then describe prototype sensors and the experimental methods used to determine their performance. We fabricated the prototypes by attaching a strain gauge to one side of shape-memory polymer plates, which can be deformed at temperatures above the glass transition temperature (T_g) by applying a small load and maintain their rigid shape after they have been cooled below T_g. We designed two types of sensor with different cross-sectional shapes (i.e., hollow cylinder and channel beam) and experimentally evaluated them. The results indicated that the sensitivity and measurement range of the sensor can be changed by modifying the cross-sectional shape. For example, for hollow cylinder and channel beam types of sensor, the maximum change in sensitivity before and after bending was 6.2 and 2.7 times, respectively. Although these values were smaller than that (12 times) obtained in our previous study, it was possible to change the sensitivity without changing the position or direction of the measured force.

Welfare assistive devices and robotic devices have been developed to assist the elderly and people/individuals with disabilities in walking. For the elderly and people with disabilities to achieve stable walking, the development of wearable robots to support the shift of the center of mass to the left and right during walking is important. Therefore, in this study, we developed a wearable robot to improve walking function through a haptic presentation. The wearable robot is equipped with a pneumatic actuator that enables haptic presentation and an inertial measurement unit (IMU) for estimating walking events. Walking support in real-time is hypothesized to be possible through this wearable robot. We verified the effectiveness of haptic presentation and developed a walking-event estimation algorithm using the IMU. From the experimental results, left and right center of pressure (CoP) changes were confirmed to be possible through haptic presentation; moreover, walking events can be estimated by the IMU attached to the waist.

High-speed robotic hands are capable of agile movements and short-cycle control and can perform a variety of actions that are not possible with ordinary robotic hands. In this study, this characteristic is focused on and verified for the construction of an algorithm to detect screw loosening from force sensor data acquired when an object is vibrated at high-speeds. First, a simple model of the system was developed by applying vibrations to a part connected by a screw to a robotic hand. Based on the constructed model, simulations were performed to confirm weather the presence or absence of screw loosening influenced force data. Subsequently, experiments were conducted using a robotic hand and models to verify the differences caused by screw loosening. These results confirmed the differences in the force data available for detecting screw loosening. Furthermore, to confirm the reliability of the constructed model and consider the inspection method, the grasping position of the inspected object was added as a condition for verification. This additional experiment confirmed that there were significant changes in the force data depending on the grasping position. Additionally, the optimal grasping conditions for the inspection method were discussed, and the validity of the method was confirmed based on simple discrimination results. In the future, simulations using models with extended degrees of freedom, along with further experimental verification, will be conducted to develop more accurate inspection methods.

Patrolling with multiple robots enables efficient surveillance for detecting and managing undesirable situations. This necessitates improved patrol efficiency and operator situational awareness at base stations. An enhanced situational awareness enables operators to predict robot behavior, support recognition and decision-making, and execute emergency interventions. This paper presents the local reactive and partition (LR-PT) algorithm. It is a novel multi-robot patrolling approach. In the simulations, LR-PT outperformed the existing methods by ensuring frequent patrols of all the locations of interest and enhancing the situational awareness of the base station. The robots independently select patrol targets based on locally available information. It integrates patrol requirements and the urgency of reporting mission progress to base stations into a unified utility function. This locality also contributes to the robustness against communication constraints and robot failures, as demonstrated in this study. The algorithm further autonomously performs emergent area partitioning. This can prevent falling into local optima and realize a comprehensive patrol over the entire mission area. The simulation results demonstrate the superior performance of LR-PT for multi-robot patrolling. It utilizes the advantages of swarm robotics and addresses real-world operational challenges.

With the aging population, personal vehicles (PVs) have become widely used and are important mobility devices for the elderly. Driving a PV for the elderly involves risks such as collisions. Therefore, driving assistance and a training system utilizing mixed reality (MR) have been developed. However, as these systems are controlled manually, safe operation cannot be guaranteed, particularly for users who are not good at operating them. In this study, an autonomous driving system for a PV using a brain-computer interface is proposed. Measurement objects generating flicker stimuli are projected onto the MR device, and steady-state visual evoked potentials in the visual cortex are detected. Using the automatic driving system, safe operation is realized without stopping, as the intention of the traveling direction of the PV is estimated. The effectiveness of the proposed system is demonstrated through experiments, and the driving workloads are evaluated using NASA-TLX.

Recently, outdoor robotics applications have increasingly adopted the “cloud robotics” approach, offloading processing to computationally rich locations via cloud communication due to growing task complexity. However, network conditions in outdoor environments are often volatile, leading to significant performance degradation or unstable behavior in robots owing to poor cloud communication. To address this issue, this study proposes combining minimal self-localization capabilities on the robot side with advanced self-localization processing on the cloud side by integrating and interpolating the two. This paper presents the implementation of 2D self-localization on a robot and 3D self-localization on the cloud, clarifies their characteristics, and proposes a fusion method that combines both self-localization techniques to enhance accuracy and robustness against communication failures. Adaptive Monte Carlo localization (AMCL), a standard algorithm for 2D self-localization, was used on the robot side. Two fusion methods—time-varying weighted average (TVWA) and unscented Kalman filter (UKF)—were implemented. It was shown that the root mean squared error (RMSE) could be reduced to 0.096 m for TVWA and 0.094 m for UKF on closed paths within a 3 m × 4 m rectangle, compared to 0.124 m for AMCL. Furthermore, even with random data loss in the self-localization estimation results from Fast-LIO, the RMSE decreased to 0.102 m for TVWA and 0.097 m for UKF. In this case, the coordinate change before and after the time step due to data loss was reduced from 0.408 to 0.263 m for TVWA and 0.108 m for UKF, indicating that the proposed method reduces the sudden coordinate shifts caused by data loss.

In nuclear fuel fabrication facilities, gloveboxes are typically dismantled manually. The integration of remotely controlled equipment, comprising a robot arm and a size reduction tool, can enhance work efficiency and mitigate radiation exposure risks in dismantling operations. The hydraulic cutter is regarded as a highly effective tool for reducing the size of steel frame structures, which are commonly composed of gloveboxes. However, when an object is severed by a hydraulic cutter fixed to a robot arm, the resultant reaction force may compromise the integrity of the robot arm or nearby structures. Consequently, in this study, we designed and manufactured a buffer device that can loosely hold the cutter to automatically align the object and absorb the reaction force. Furthermore, a visual support system was developed to assist the operator in performing remote dismantling operations. This system utilized a 3D viewer to project the robot arm, the buffer device, and the working environment. The functionality of the buffer device and 3D viewer was evaluated for the glovebox test bed. The experimental results satisfactorily confirmed the functionality of the buffer device to self-align the object and absorb sudden movements of the hydraulic cutter. Moreover, the 3D viewer provided the robot arm operator with an unobstructed perspective of the work environment, thereby confirming the efficacy of the visual support system in facilitating remote dismantling operations.

The advancement in automation technology for excavators signifies a shift from individual excavation tasks to collaborative multi-machine operations, with the aim of enhancing efficiency and safety in extensive operations. This study presents a concise overview of multi-intelligent excavator collaboration systems (MECS), introducing a framework that includes networked communication, task analysis, and motion planning. Networked communication is foundational, bolstered by the widespread use of Ethernet and the industrialization of 5G technology. Task analysis, which is the core of system, is bifurcated into single-agent intelligence and multi-machine collaboration, considering the task efficiency and collaborative completeness in complex environments. Motion planning, inherently linked to task analysis, is divided into operational and mobility aspects. Finally, this paper concludes by summarizing and projecting key technologies within the framework of collaborative systems.