Human action recognition is an important research area in the field of computer vision that can be applied in surveillance, assisted living, and robotic systems interacting with people. Although various approaches have been widely used, recent studies have mainly focused on deep-learning networks using Kinect camera that can easily generate data on skeleton joints using depth data, and have achieved satisfactory performances. However, their models are deep and complex to achieve a higher recognition score; therefore, they cannot be applied to a mobile robot platform using a Kinect camera. To overcome these limitations, we suggest a method to classify human actions in real-time using a single RGB camera, which can be applied to the mobile robot platform as well. We integrated two open-source libraries, i.e., OpenPose and 3D-baseline, to extract skeleton joints on RGB images, and classified the actions using convolutional neural networks. Finally, we set up the mobile robot platform including an NVIDIA JETSON XAVIER embedded board and tracking algorithm to monitor a person continuously. We achieved an accuracy of 70% on the NTU-RGBD training dataset, and the whole process was performed on an average of 15 frames per second (FPS) on an embedded board system.

As the number of patients with amputations increases, research on assistive devices such as prosthetic limbs is actively being conducted. However, the development of assistive devices for patients with partial amputations is insufficient. In this study, we developed a finger prosthesis for patients with partial amputations. The design and mathematical modeling of the prosthesis are briefly presented. A pneumatic actuator, based on the McKibben muscle design, was employed to drive the finger prosthesis. We characterized the relationship between the actuator’s force and axial length changes with varying pressure. An empirical model derived from conventional mathematical modeling of force and axis length changes was proposed and compared with experimental data, and the error was measured to be between about 3% and 13%. In order to control the actuator using an electromyography (EMG) signal, an electrode was attached to the user’s finger flexors. The EMG signal was measured in relation to the actual gripping force and was provided with visual feedback, and the magnitude of the signal was evaluated using root mean square (RMS). Depending on the evaluated EMG signal magnitude, the pressure of the actuator was continuously adjusted. The pneumatic pressure was adjusted between 100 kPa and 250 kPa, and the gripping force of the finger prosthesis ranged from about 0.7 N to 6.5 N. The stiffness of the prosthesis can be varied using the SMA spring. The SMA spring is switched to a fully austenite state at 50 °C through PID control, and when the finger prosthesis is bent to a 90° angle, it can provide approximately 1.2 N of assistance force. Finally, the functional evaluation of the finger prosthesis was performed through a pinch grip test of eight movements.

Stroke causes neurological pathologies, including gait pathologies, which are diagnosed by gait analysis. However, existing gait analysis devices are difficult to use in situ or are disrupted by external conditions. To overcome these drawbacks, a flexible capacitance sensor was developed in this study. To date, a performance comparison of flexible sensors with different dimensions has not been carried out. The aim of this study was to provide optimized sensor dimension information for gait analysis. To accomplish this, sensors with seven different dimensions were fabricated. The dimensions of the sensors were based on the average body size and movement range of 20- to 59-year-old adults. The sensors were characterized by 100 oscillations. The minimum hysteresis error was 8%. After that, four subjects were equipped with the sensor and walked on a treadmill at a speed of 3.6 km/h. All walking processes were filmed at 50 fps and analyzed in Kinovea. The RMS error was calculated using the same frame rate of the video and the sampling rate of the signal from the sensor. The smallest RMS error between the sensor data and the ankle angle was 3.13° using the 49 × 8 mm sensor. In this study, we confirm the dimensions of the sensor with the highest gait analysis accuracy; therefore, the results can be used to make decisions regarding sensor dimensions.

Telemanipulation systems have been widely utilized in industrial, surgical, educational, and even military fields. One of the important issues is that when a robot interacts with environment or objects, it can damage the robot itself or the objects due to hard contact. To address this problem, we propose a novel compliance control of a slave robot using the electromyography (EMG) signal, which is measured by the sensors attached onto the master operator’s arm. The EMG signal is used since it is easy to process and it provides humans with an intuitive capability to perform the operational work. Furthermore, it has been proved that the EMG signal is useful in the control of the stiffness of the slave robot. This research identifies the muscle that is the best suitable to a precision-grip operation, and a series of experiments were performed. A compliance control algorithm with a variable stiffness of a slave robot is proposed, where the stiffness is changed based on the EMG signal, and it is confirmed by a series of experiments using a two-channel force/position teleoperation architecture.