The COVID-19 pandemic has severely affected public health, healthcare systems, and daily life, especially amid resource shortages and limited workers. This crisis has underscored the urgent need for automation in hospital environments, particularly disinfection, which is crucial to controlling virus transmission and improving the safety of healthcare personnel and patients. Ultraviolet (UV) light disinfection, known for its high efficiency, has been widely adopted in hospital settings. However, most existing research focuses on maximizing UV coverage while paying little attention to the impact of human activity on virus distribution. To address this issue, we propose a mobile robotic system for UV disinfection focusing on the virus hotspot. The system prioritizes disinfection in high-risk areas and employs an approach for optimized UV dosage to ensure that all surfaces receive an adequate level of UV exposure while significantly reducing disinfection time. It not only improves disinfection efficiency, but also minimizes unnecessary exposure in low-risk areas. In two representative hospital scenarios, our method achieves desired disinfection effectiveness while reducing the time spent evidently. The video of the experiment is available at: https://youtu.be/wHcWzOcoMPM.

As robot surgery becomes more popular these days, it consequently solicits a high demand on its training system. Currently, most of the systems are intended for personal training, while it may be more effective if an experienced expert can also be involved. It thus motivates us to propose a dual-user robotic surgical training system that can closely link the mentor and trainee together. The system is developed in a mixed-reality environment that integrates the information from both the virtual and real worlds. The haptic feedback and virtual fixtures are employed for assistance during skill transfer. The proposed system can assess the proficiency level of the trainee for the mentor to provide proper assistance during training. The experimental results demonstrate that the proposed system is capable of enhancing the trainee's ability for surgery. Especially, similar effectiveness is observed during comparative studies with direct hands-on guidance frequently adopted for skill transfer during sport training.

Following its applications on reception, space exploration, health care, and others, the dual-arm robot is employed for industrial automation more intensively these days. With a human-like two-arm structure of high degrees of freedom, it should be a strong competitor for industrial tasks of high complexity. However, up to date, its deployment is still far behind the traditional single-arm robot manipulator. It may be partially due to its higher price and larger installation space. Meanwhile, the lack of proper methods or devices to deal with task planning and teaching that usually involves robot motions of more than 10 DOF is also influential. Motivated by it, in this paper, we propose a novel manipulation system for the industrial dual-arm robot manipulator using virtual reality (VR). Especially, we propose taking advantage of the similarity between the dual-arm robot and human arm and come up with a human-like path planner, with an intention to incorporate that of human into robot path planning. In addition, we also furnish the system with several assistive tools, including that for physical behavior emulation and self-collision detection. Simulations based on realistic environments are conducted to demonstrate the effectiveness of the proposed manipulation system, along with the questionnaires to evaluate user's responses.

Along with the progress in automation and Industry 4.0, the use of industrial robots becomes much more popular. Meanwhile, the conventional manipulative devices, such as teaching pendant and joystick, are not intuitive enough and very time-consuming for task teaching. In facing more complicated tasks these days, it then induces high cost and demands experts with proper training. This paper thus aims to develop an intuitive manipulatory interface for assistance in simplifying the teaching process. We propose a wireless 3D manipulation system based on augmented reality (AR) with the tablet PC as the platform. The system allows the user to govern the movement of the real robot in a 3D space via a virtual one generated through the AR. We also develop several intelligent tools to assist the manipulation, so that accurate and collision-free path can be generated for the robot to follow for task execution. In addition, the system can also be used for offline training, which yields salient training effect for the operator. Experiments are executed to demonstrate the effectiveness of the proposed system, and questionnaires conducted to investigate user's response.

With the arrival of aging society, the passive robot walking helper is introduced to provide safe mobility for the elders, which features continuous energy dissipation from the system and is thus intrinsically safe. As both static and moving obstacles may be present in daily activities, obstacle avoidance during guidance is very imperative for its practical use, which motivates us to propose such a scheme based on receding horizon control. With a desired goal to reach, the proposed scheme first plans a smooth path for the walking helper to follow. A novel strategy, combined with receding horizon control, is then proposed to avoid the incoming obstacles detected by the equipped sensing system. Experiments based on our developed passive robot walking helper, i-Go, are conducted in real environments to demonstrate the effectiveness of the proposed scheme.

As the elderly population is rapidly growing in our society, robot walking helpers are receiving more attention these days. To be practical and safe for use in daily life, an important issue is whether these robot walking helpers can be effectively maneuvered and provide guidance to the users. These concerns become even more important for a passive robot walking helper, as it relies mainly on the user's applied forces to move. In this paper, we propose a motion guidance system for a passive robot walking helper. More specifically, we develop a system to identify user's intentions from applied grip forces. This system consists of a pair of force-sensing grips attached to the handles of the robot walking helper and a learning scheme for the mapping between the measured grip forces and the driving force/torque imposed on the helper. The learning scheme, combined with an assistive strategy extended from our previous work, provides motion guidance to the elders during walking. The feasibility of the proposed system is demonstrated via a series of experiments involving motion assistance with i-Go, a passive robot walking helper developed in our laboratory.

With the increase of aging population, robot walking helpers have been developed to assist the elderly in their daily life. Based on our previous work, a passive robot walking helper (named i-Go) equipped with a guidance strategy, in this paper, we further propose a navigation scheme for obstacle avoidance. The proposed scheme first locates the waypoints in the collision-free areas, and then guides the i-Go to reach the desired target with the specified orientation. Simulations and experiments are performed to verify the effectiveness of the proposed scheme. The results demonstrate that the scheme can guide the user to avoid the obstacles and successfully approach the target with the specified orientation.

This paper presents a simple and effective approach to govern robot arm motion in real time using upper limb EMG signals. Considering the non-stationary and nonlinear characteristics of the EMG signals, in the design for feature extraction, we introduce the empirical mode decomposition (EMD) to decompose the EMG signals into intrinsic mode functions (IMFs). Each IMF represents different physical characteristic, so that the muscular movement can be recognized. We then integrate it with a so-called initial point detection method previously proposed to establish the mapping between the upper limb EMG signals and corresponding robot arm movements in real time. In addition, for each individual user, we adopt a fuzzy approach to select proper system parameters for motion classification. The experimental results show the feasibility of the proposed approach with accurate motion recognition.

Simulation systems nowadays are applied for various tasks, and thus demands a versatile manipulative system for the user to interact with the corresponding simulated environments. To make a single manipulative device applicable for different tasks, the concept of virtual mechanism has been previously proposed, in which virtual motion constraints are constructed via the software to constrain the manipulative device to move within a limited workspace that corresponds to task requirements. Motivated by the idea, in this paper, we propose a systematic approach to design and implement the virtual motion constraints for a multi-functional virtual manipulation system. The motion constraints are generated from sets of virtual walls, built of different shapes and physical properties, to deal with the compliance task, which involves both position and force management. An algorithm on how to properly assemble the walls and a graphics-based method for smooth force rendering between them are proposed for given compliance tasks. For demonstration, in experiments, we implement a virtual omni-directional wrench based on using a 2-DOF force-reflection joystick.

The self-organizing map (SOM), as a kind of unsupervised neural network, has been used for both static data management and dynamic data analysis. To further exploit its search abilities, in this paper we propose an SOM-based algorithm (SOMS) for optimization problems involving both static and dynamic functions. Furthermore, a new SOM weight updating rule is proposed to enhance the learning efficiency; this may dynamically adjust the neighborhood function for the SOM in learning system parameters. As a demonstration, the proposed SOMS is applied to function optimization and also dynamic trajectory prediction, and its performance compared with that of the genetic algorithm (GA) due to the similar ways both methods conduct searches.

Teleoperation techniques unite the human operator as the supervisor and the machine as the manipulator. Because both the human operator and machine are involved in the control loop and they are connected via the network instead of a direct link, the development of a telerobotic system poses challenges different from systems involving machines alone. And these challenges become more severe when the telerobotic system is used for compliance task, in which simultaneous control of both position and force are demanded and inevitable contact with the environment is encountered. In this paper, we discuss the challenges we may face in designing a telerobotic system. We then describe the networked VR-based telerobotic system developed in our laboratory and report the results by using the developed telerobotic system to execute several different kinds of compliance tasks.

It has been observed that human limb motions are not very accurate, leading to the hypothesis that the human motor control system may have simplified motion commands at the expense of motion accuracy. Inspired by this hypothesis, we propose learning schemes that trade motion accuracy for motion command simplification. When the original complex motion commands capable of tracking motion accurately are reduced to simple forms, the simplified motion commands can then be stored and manipulated by using learning mechanisms with simple structures and scanty memory resources, and they can be executed quickly and smoothly. We also propose learning schemes that can perform motion command scaling, so that simplified motion commands can be provided for a number of similar motions of different movement distances and velocities without recalculating system dynamics. Simulations based on human motions are reported that demonstrate the effectiveness of the proposed learning schemes in implementing this accuracy-simplification tradeoff.

In this research, the problems of how to teach a robot to execute skilled operations are studied. Human workers usually accumulate his experience after executing the same task repetitively. In the process of training, the worker must find ways of adjusting his/her execution. In our system, the parameters for the impedance control scheme are used as the targets for adjustment. After mass amount of training, the worker is supposed to be able to execute deburring tasks successfully. This is because the worker might have gotten some knowledge about tuning the parameters required in the impedance control scheme. Thus, the rules for adjusting the parameters in impedance control are the operational skills to be identified. In this research, a training scheme, called the evolution-based virtual training scheme, is proposed in extracting knowledge for robotic deburring tasks. In this approach, an evolution strategy is employed for searching for the best set of fuzzy rules. This learning scheme has been successfully applied in adjusting the parameters of impedance controllers required in deburring operations. In general, the results of deburring are much more satisfactory when compared with those in previous research. When executing a deburring task, the robot simulator can find its optimal adjusting rules for parameters after several generations of evolution.

In robot learning control, the learning space for executing general motions of multijoint robot manipulators is quite large. Consequently, for most learning schemes, the learning controllers are used as subordinates to conventional controllers or the learning process needs to be repeated each time a new trajectory is encountered, although learning controllers are considered to be capable of generalization. In this paper, we propose an approach for larger learning space coverage in robot learning control. In this approach, a new structure for learning control is proposed to organize information storage via effective memory management. The proposed structure is motivated by the concept of human motor program and consists mainly of a fuzzy system and a cerebellar model articulation controller (CMAC)-type neural network. The fuzzy system is used for governing a number of sampled motions in a class of motions. The CMAC-type neural network is used to generalize the parameters of the fuzzy system, which are appropriate for the governing of the sampled motions, to deal with the whole class of motions. Under this design, in some sense the qualitative fuzzy rules in the fuzzy system are generalized by the CMAC-type neural network and then a larger learning space can be covered. Therefore, the learning effort is dramatically reduced in dealing with a wide range of robot motions, while the learning process is performed only once. Simulations emulating ball carrying under various conditions are presented to demonstrate the effectiveness of the proposed approach.

In this paper, we propose a novel scheme for governing similar robot motions by using learning mechanisms. Most learning schemes need to repeat the learning process each time a new trajectory is encountered. The main reason for this deficiency is that the learning space for executing general motions of multi-joint robot manipulators is too large. To reduce the complexity in learning, we first classify robot motions according to their similarity. A new learning structure, which is motivated by the concept of a motor program, is then used to learn a class of motions. The proposed structure consists mainly of a fuzzy system and a CMAC-type neural network. The fuzzy system is used for learning of the samples in a class of motions. The CMAC-type neural network is used to generalize the parameters of the fuzzy system, which are appropriate for the control of the sampled motions, to deal with the whole class of motions. The learning process is performed only once and the learning effort is dramatically reduced for a wide range of robot motions.

This study focuses on the characterization and control of voluntary movements by analyzing motor command patterns, ensuing analysis from involuntary limb movements. As a result, a neuromuscular-like model with two muscle-reflex models is proposed, and motor command patterns are approximated identified from the experimental data. Based on the identified motor command patterns, a control strategy, muscle modulation control, is proposed for possible applications in robotic control using neuromuscular-like models.>

The muscle-reflex mechanisms of private limbs are studied and modeled so that robotic controls may benefit from the findings. An extensive body of experimental evidence indictates that velocity-dependent force responses of the neuromuscular system have a nonlinear damping effect proportional to a fractional power of velocity. This highly nonlinear viscosity may help limbs adapt to different loads and bring movements to graceful terminations. To explore the characteristics of this nonlinear damping property, a theoretical study using the phase-plane approach is presented. The effects of different loads, damping constants, and stiffnesses are analyzed and simulated. From the results of this phase-plane analysis, a muscle-reflex model is developed and proposed for robotic compliance control.>

An efficient approach is presented for solving the inverse Jacobian problem for wrist-partitioned robots; specifically, the differential inverse kinematics problems. By selecting the wrist coordinate frame as the reference coordinate frame, a simplified Jacobian relation can be obtained for deriving a set of orthogonal bases for the screw system. From these orthogonal screws, the instantaneous Cartesian motion can be separated into translational and rotational parts for solving the differential joint variable rates. As a result, a set of concise, closed-form equations for solving the above inverse problem can be easily generated. For illustration, the PUMA 560 robot is selected as an example. On the basis of the simplified equations, the exact form of the inverse Jacobian matrix can also be obtained.>

A design is presented of a robot accuracy compensator after calibration, which can efficiently calculate and compensate the calibrated kinematic errors for robot manipulators. This compensator consists of a direct robot accuracy calculator (DRAC) that efficiently calculates the accurate robot Cartesian positions using a recursive formulation of the T/sub /i matrix, and an inverse robot accuracy compensator (IRAC), that calculates the equivalent changes in robot joint positions caused by the Cartesian errors obtained from the DRAC. By subtracting these equivalent joint position changes to the current existing kinematic joint solutions, the robot accuracy caused by the kinematic errors can be compensated. As an example, an efficient accuracy compensator designed for PUMA 560 robots is given.>