Publications

Grasp force estimation based on surface electromyography (sEMG) is essential for the dexterous control of a prosthetic hand. Nowadays, although increasing the number of sEMG measurement positions and extracting more features are common methods to increase the accuracy of grasp force estimation, it will increase the computational burden. In this paper, an approach based on analysis of variance (ANOVA) and generalized regression neural network (GRNN) for optimal measurement positions and features is proposed, with the purpose of using fewer measurement positions or features to achieve higher estimation accuracy. Firstly, we captured six channels of sEMG from subjects’ forearm and grasp force synchronously. Then, four kinds of features in time domain are extracted from each channel of sEMG. By combining different measurement position sets (MPSs) and feature set (FSs), we construct 945 data sets. These data sets are fed to GRNN to realize grasp force estimation. Normalized root mean square error (NRMS), normalized mean of absolute error (NMAE), and correlation coefficient (CC) between estimated grasp force and actual force are introduced to evaluate the performance of grasp force estimation. Finally, ANOVA and Tukey HSD testing are introduced to analyze grasp force estimation results so as to obtain the optimal measurement positions and features. We obtain the optimal MPSs for grasp force estimation when different FSs are employed, and the optimal FSs when different MPSs are utilized.

Aerial manipulators (AMs) have evolved to fulfill various interaction tasks, such as push-and-slide tasks. However, a push-and-slide operation on a moving object is still challenging for a regular AM because its flying platform is usually a rotorcraft unmanned aerial vehicle (UAV)-a floating base with complex aerodynamics, under-actuation property, body vibration, etc. This paper proposes an easy-to-implement approach-an interaction control framework based on a remodeling idea of a closed-loop system, allowing an AM to steadily transition from a free flight to a contact flight and slide along a moving object. Then, a decoupled control idea is applied to implement this framework on an AM system. Besides, a contact model and a trajectory planning issue in application scenarios are also studied. Finally, several push-and-slide operation experiments evaluate the proposed interaction control framework. The experimental process and results validate its effectiveness.

Automatic melanoma diagnosis based on image processing can give more objective results. To facilitate examination for patients at home, we propose a new automatic melanoma diagnosis framework based on common images. Firstly, we use illumination assessment based on variational framework for Retinex (VFR) to filter the images with illumination problem caused by variation of capturing cameras viewpoint and ambient light. Secondly, the GrabCut algorithm based on colour difference is used to segment lesion area. It can complete segmentation automatically and efficiently. Thirdly, we use convolutional neural network (CNN) to extract high-level features and choose support vector machine (SVM) classifier to complete melanoma classification. Compared to hand-craft features, CNN can acquire deep information of images. Because of the lack of medical images, the SVM classifier is better than other classifiers. Finally, we validated our approach from different perspectives and the accuracy is increased by about 5% over other methods.

The controlling of robotic arms based on brain–computer interface (BCI) can revolutionize the quality of life and living conditions for individuals with physical disabilities. Invasive electroencephalography (EEG)-based BCI has been able to control multiple degrees of freedom (DOFs) robotic arms in three dimensions. However, it is still hard to control a multi-DOF robotic arm to reach and grasp the desired target accurately in complex three-dimensional (3D) space by a noninvasive system mainly due to the limitation of EEG decoding performance. In this study, we propose a noninvasive EEG-based BCI for a robotic arm control system that enables users to complete multitarget reach and grasp tasks and avoid obstacles by hybrid control. The results obtained from seven subjects demonstrated that motor imagery (MI) training could modulate brain rhythms, and six of them completed the online tasks using the hybrid-control-based robotic arm system. The proposed system shows effective performance due to the combination of MI-based EEG, computer vision, gaze detection, and partially autonomous guidance, which drastically improve the accuracy of online tasks and reduce the brain burden caused by long-term mental activities.

Aerial manipulator (AM) systems are significantly more effective than both conventional manipulators and flying robots, especially in disaster rescue settings, because they can perform human-like interaction tasks in flight. However, controlling the contact force of an AM is difficult due to the coupling between its flying platform and manipulator. This paper proposes a contact force control framework to address this problem. First, the UAV/AM position response under an external force is studied, and it is theoretically shown that a closed-loop UAV/AM behaves as a spring-mass-damper system. Second, a contact force controller is designed using an inverse-dynamics method. Third, an attitude feed-forward approach is employed to improve the force tracking performance. Then, the real-time contact position is introduced into the control system to achieve interaction with an actively moving environment. Finally, an innovative AM is developed and subjected to flight experiments, validating the proposed framework. An AM’s characteristics in interaction operations are summarized, and general conclusions are drawn. This study is novel in that 1) the contact force control is implemented without relying on force sensors, 2) the whole framework is applicable for controlling a constant/variable contact force with high closed-loop performance, and 3) it can also perform reliable interaction with a dynamic and unknown environment. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">Note to Practitioners</i> —This study is motivated by the contact force control problem of an aerial manipulator (AM) during interaction operations. Previous approaches have explored the feasibility of controlling contact forces to some extent, but they lack universality and assume a static environment. Ensuring sufficient safety and providing reliable solutions in real-world applications remain challenging. This paper aims to investigate the position response of a closed-loop aircraft system under external forces, without disrupting its existing steady flight. The proposed method transforms contact force control into position control, eliminating the need for a force sensor. A series of aerial experiments were conducted to validate the effectiveness and applicability of the method in various scenarios. Our ongoing work will focus on migrating such an AM system from a laboratory scenario to the real world.

Adjustable joint compliance is essential for ensuring safe and natural human-robot interaction in rehabilitation scenarios. This paper presents an elbow rehabilitation robot driven by a dual-motor load sharing variable stiffness actuator and coordinated via a hierarchical control framework. The mechanical design integrates a power transmission mechanism with a variable stiffness mechanism, enabling task-adaptive compliance without increasing sensor complexity. At the control level, a highlevel control layer generates assistive joint torques based on taskspace impedance, while a low-level control layer translates torque commands into precise deflection angle for robust execution. Experimental evaluations, including bench tests and human-robot interaction trials, validate the system's capabilities in accurate torque rendering, high transparency under low stiffness, strong assistance under high stiffness, and safe, comfortable interaction across diverse rehabilitation tasks.

This paper presents the design and implementation of a wireless sensor and actuator network (WSAN) system for calling home robots in indoor environments. The proposed system consists of a network of wireless pyroelectric infrared (PIR) sensor nodes and a mobile actuator node. The PIR sensor nodes are deployed on the ceiling of a room to form a grid topology. The location of a person within the detection area can be calculated by a localization algorithm based on multi-sensor information fusion. Depending on the location information, the mobile node serves as a home robot that can navigate to the person asking for help. A prototype system of the proposed WSAN has been implemented. Experimental results show that both the localization accuracy and the navigation accuracy are good enough for the mobile node to put itself within easy reach. This service plays a very important role in assisting elderly and disabled people to maintain their independence and improve their quality of life in their homes.