Seniors who live alone at home are at risk of falling and injuring themselves and, thus, may need a mobile robot that monitors and recognizes their poses automatically. Even though deep learning methods are actively evolving in this area, they have limitations in estimating poses that are absent or rare in training datasets. For a lightweight approach, an off-the-shelf 2D pose estimation method, a more sophisticated humanoid model, and a fast optimization method are combined to estimate joint angles for 3D pose estimation. As a novel idea, the depth ambiguity problem of 3D pose estimation is solved by adding a loss function deviation of the center of mass from the center of the supporting feet and penalty functions concerning appropriate joint angle rotation range. To verify the proposed pose estimation method, six daily poses were estimated with a mean joint coordinate difference of 0.097 m and an average angle difference per joint of 10.017 degrees. In addition, to confirm practicality, videos of exercise activities and a scene of a person falling were filmed, and the joint angle trajectories were produced as the 3D estimation results. The optimized execution time per frame was measured at 0.033 s on a single-board computer (SBC) without GPU, showing the feasibility of the proposed method as a real-time system.

In this study, we propose a modified particle swarm optimization (PSO) algorithm, which is an improved version of the conventional PSO algorithm. To improve the performance of the conventional PSO, a novel method is applied to intelligently control the number of particles. The novel method compares the cost value of the global best (gbest) in the current iteration to that of the gbest in the previous iteration. If there is a difference between the two cost values, the proposed algorithm operates in the exploration stage, maintaining the number of particles. However, when the difference in the cost values is smaller than the tolerance values assigned by the user, the proposed algorithm operates in the exploitation stage, reducing the number of particles. In addition, the algorithm eliminates the particle that is nearest to the best particle to ensure its randomness in terms of the Euclidean distance. The proposed algorithm is validated using five numerical test functions, whose number of function calls is reduced to some extent in comparison to conventional PSO. After the algorithm is validated, it is applied to the optimal design of an interior permanent magnet synchronous motor (IPMSM), aiming at minimizing the total harmonic distortion (THD) of the back electromotive force (back EMF). Considering the performance constraint, an optimal design is attained, which reduces back EMF THD and satisfies the back EMF amplitude. Finally, we build and test an experimental model. To validate the performance of the optimal design and optimization algorithm, a no-load test is conducted. Based on the experimental result, the effectiveness of the proposed algorithm on optimal design of an electric machine is validated.

During the process of plant growth, such as during the flowering stages and fruit development, the plants need to be provided with the various minerals and nutrients to grow. Nutrient deficiency is the cause of serious diseases in plant growth, affecting crop yield. In this article, we employed artificial neural network models to recognize, classify, and predict the nutritional deficiencies occurring in tomato plants (Solanum lycopersicum L.). To classify and predict the different macronutrient deficiencies in the cropping process, this paper handles the captured images of the macronutrient deficiency. This deficiency during the fruiting and leafing phases of tomato plant are based on a deep convolutional neural network (CNN). A total of 571 images were captured with tomato leaves and fruits containing the crop species at the growth stage. Among all images, 80% (461 captured images) were used for the training dataset and 20% (110 captured images) were applied for the validation dataset. In this study, we provide an analysis of two different model architectures based on convolutional neural network for classifying and predicting the nutrient deficiency symptoms. For instance, Inception-ResNet v2 and Autoencoder with the captured images of tomato plant growth under the greenhouse conditions. Moreover, a major type of statistical structure, namely Ensemble Averaging, was applied with two aforementioned predictive models to increase the accuracy of predictive validation. Three mineral nutrients, i.e., Calcium/Ca²⁺, Potassium/K⁺, and Nitrogen/N, are considered for use in evaluating the nutrient status in the development of tomato plant with these models. The aim of this study is to predict the nutrient deficiency accurately in order to increase crop production and prevent the emergence of tomato pathology caused by lack of nutrients. The predictive performance of the three models in this paper are validated, with the accuracy rates of 87.273% and 79.091% for Inception-ResNet v2 and Autoencoder, respectively, and with 91% validity using Ensemble Averaging.

Many studies have explored emotional and mental services that robots can provide for older adults, such as offering them daily conversation, news, music, or health information. However, the ethical issues raised by using sensors for frail older adults to monitor their daily movements or their medication intake, for instance, are still being discussed. In this study, we develop an older adult-guided, caregiver-monitored robot, Dori, which can detect and recognize movement by sensing human poses in accordance with two factors from the human-centered artificial intelligence (HCAI) framework. To design the care robot’s services based on sensing movement during daily activities, we conducted focus group interviews with two groups—caregivers and medical staff—on the topic of care robot services not for patients but for prefrail and frail elderly individuals living at home. Based on their responses, we derived the focal service areas of cognitive support, emotional support, physical activity support, medication management, and caregiver management. We also found the two groups differed in their ethical judgments in the areas of dignity, autonomy, controllability, and privacy for services utilizing sensing by care robots. Therefore, the pose recognition technology adopted in the present work uses only joint coordinate information extracted from camera images and thus is advantageous for protecting human dignity and personal information.

As human-robot collaboration increases tremendously in real-world applications, a fully autonomous and reliable mobile robot for the collaboration has been a central research topic and investigated extensively in a large number of studies. One of the most pressing issues in such topic is the collision-free navigation that has a moving goal and unknown obstacles under the unstructured environment. In this paper, a cognitive robotic system is proposed for the robot to navigate itself to the moving target person without obstacle collision. This cognitive robotic system consists of a cognitive agent, which is created based on Soar cognitive architecture to reason its current situation and make action decision for the robot to avoid obstacles and reach the target position, and a speed planning module, which is based on dynamic window approach to generate appropriate linear and angular velocities for driving the robot’s motors. For the implementation of the proposed system, we use a differential drive wheel robot equipped with two ultra-wideband sensors and a color depth camera as the experimental platform. Finally, to evaluate the performance of our system in actual operating conditions, we conduct experiments with a scenario that includes main tasks: avoiding consecutive unknown obstacles and turning at corner while the robot follows continuously human user along the corridor.

People who have difficulty moving owing to problems in walking spend their lives assisted by wheelchairs. In the past, research has been conducted regarding the application of various technologies to electric wheelchairs for user convenience. In this study, we evaluated a method of applying an autonomous driving function and developed an autonomous driving function using ROS. An electric wheelchair with a control unit designed to enable autonomous driving was used to test the basic performance of autonomous driving. The effectiveness of the technology was confirmed by comparing the results of autonomous driving with those of manual driving on the same route. It is expected that the evaluation and improvement of the usability and ride quality as well as additional studies will help improve the mobility convenience of physically disabled persons.