Researchers have given attention to lower limb exoskeletons in recent years. Lower limb exoskeletons have been designed, prototype tested through experiments, and even produced. In general, lower limb exoskeletons have two different objectives: (1) rehabilitation and (2) assisting human work activities. Referring to these objectives, researchers have iteratively improved lower limb exoskeleton designs, especially in the location of actuators. Some of these devices use actuators, particularly on hips, ankles or knees of the users. Additionally, other devices employ a combination of actuators on multiple joints. In order to provide information about which actuator location is more suitable; a review study on the design of actuator locations is presented in this paper. The location of actuators is an important factor because it is related to the analysis of the design and the control system. This factor affects the entire lower limb exoskeleton’s performance and functionality. In addition, the disadvantages of several types of lower limb exoskeletons in terms of actuator locations and the challenges of the lower limb exoskeleton in the future are also presented in this paper.

An exoskeleton is a device used for walking rehabilitation. In order to develop a proper rehabilitation exoskeleton, a user’s walking intention needs to be captured as the initial step of work. Moreover, every human has a unique walking gait style. This work introduced a wearable sensor, which aimed to recognize the walking gait phase, as the fundamental step before applying it into the rehabilitation exoskeleton. The sensor used in this work was the IMU sensor, used to recognize the pitch angle generated from the knee joint while the user walks, as information about the walking gait cycle, before doing the investigation on how to identify the walking gait cycle. In order to identify the walking gait cycle, Neural Network has been proposed as a method. The gait cycle identification was generated to recognize the gait cycle on the knee joint. To verify the performance of the proposed method, experiments have been done in real-time application. The experiments were carried out with different processes such as walking on a flat floor, climbing up, and walking down stairs. Five subjects were trained and tested using the system. The experiments showed that the proposed method was able to recognize each gait cycle for all users as they wore the sensor on their knee joints. This study has the potential to be applied on an exoskeleton rehabilitation robot as a further research experiment.

The deep learning-based object detector accuracy has surpassed conventional detection methods. Although implementation is still limited to hardware capabilities, this problem can be overcome by combining edge devices with cloud computing. The recent study of cloud-based object detector architecture is generally based on representational state transfer (RESTful web services), which uses a pooling system method for data exchange. As a result, this system leads to a low detection speed and cannot support real-time data streaming. Therefore, this study aims to enhance the detection speed in cloud-based object recognition systems using gRPC and Protobuf to support real-time detection. The proposed architecture was deployed on the Virtual Machine Instance (VMI) equipped with a Graphics Processing Unit (GPU). The gRPC server and YOLOv3 deep learning object detector were executed on the cloud server to handle detection requests from edge devices. Furthermore, the captured images from the edge devices were encoded into Protobuf format to reduce the message size delivered to the cloud server. The results showed that the proposed architecture improved detection speed performance on the client-side in the range of 0.27 FPS to 1.72 FPS compared to the state-of-the-art method. It was also observed that it could support multiple edge devices connection with slight performance degradation in the range of 1.78 FPS to 1.83 FPS, depending on the network interface used.

A humanoid robot called BarelangFC was designed to take part in the Kontes Robot Indonesia (KRI) competition, in the robot coordination division. In this division, each robot is expected to recognize its opponents and to pass the ball towards a team member to establish coordination between the robots. In order to achieve this team coordination, a fast and accurate system is needed to detect and estimate the other robot’s position in real time. Moreover, each robot has to estimate its team members’ locations based on its camera reading, so that the ball can be passed without error. This research proposes a Tiny-YOLO deep learning method to detect the location of a team member robot and presents a real-time coordination system using a ZED camera. To establish the coordinate system, the distance between the robots was estimated using a trigonometric equation to ensure that the robot was able to pass the ball towards another robot. To verify our method, real-time experiments was carried out using an NVDIA Jetson NX Xavier, and the results showed that the robot could estimate the distance correctly before passing the ball toward another robot.

Due to the development of technology, it is easy to integrate this into many of applications such as automatic security door lock system. As presented in this paper, we proposed a voice recognition for the user authentication in order to open the door by using the word in Bahasa. In order to recognize the user voice, the convolution neural network (CNN) deep learning algorithm will be utilized as a proposed method. We also provide a simple web based which is allowed to recognize the user voice directly when the user spoke. Some experiments have been carried out to verify the performance of the system by using the work “Buka” and some words which has the similarity in pronunciation. Based on the experiment results, our system was able to recognize the trained voice accurately even the user used another word which was similar to say.

ABSTRAK Tujuan dari penelitian ini adalah untuk mendeteksi objek dan mengestimasi pose objek menggunakan kamera RGB-D. Dalam penelitian ini, kami mengusulkan pemrosesan data pada citra RGB dan citra depth saja, tanpa menggunakan point cloud, seperti pada umumnya. Metode yang diusulkan mendeteksi posisi dan orientasi objek menggunakan DRBox-v2 dari Region of Interest (ROI), yang sebelumnya diperoleh dari pendeteksian pada penanda ArUco. Hasil deteksi objek kemudian diskalakan dan digunakan pada citra depth untuk mendapatkan perkiraan posisi dan orientasi objek. Dari sisi pendeteksi objek, usulan metode memperoleh nilai Average Precision (AP) sebesar 0,740. Sedangkan untuk estimator pose, usulan metode menghasilkan kesalahan posisi rata-rata 13,36 mm dan kesalahan orientasi rata-rata 0,75 derajat. Metode yang diusulkan berpotensi menjadi alternatif sistem deteksi objek dan estimasi pose pada kamera RGB-D yang tidak memerlukan pemrosesan point cloud dan tidak memerlukan model referensi objek. Kata kunci: deteksi objek, estimasi pose, DRBox, ArUco, bin-picking ABSTRACT This study aims to detect objects and estimate the object's pose using an RGB-D camera. In this study, we proposed data processing on RGB images and depth images only, without using point clouds, as in general. The proposed method detected the object's position and orientation using the DRBox-v2 from the Region of Interest (ROI), which was previously obtained from detecting ArUco markers. The object detection results were then scaled and used in the depth image to get the object's approximate position and orientation. In object detection, the proposed method obtained an Average Precision (AP) value of 0.740. As for the pose estimator, our method generated an average position error of 13.36 mm and an average orientation error of 0.75 degrees. Therefore, this method can be an alternative object detection and pose estimation system on an RGB-D camera that does not require point cloud processing and an object reference model. Keywords: object detection, pose estimation, DRBox, ArUco, bin-picking