The performance of humanoid robots is improving, owing in part to their participation robot games such as the DARPA Robotics Challenge. Along with 2018 Winter Olympics Pyeongchang, a Skiing Robot Competition was held which participated autonomously giant slalom alpine skiing competition. were required transit through many red or blue gates on ski slope reach finish line. course relatively short at 100 m long and had an intermediate-level rating. A 1.23 tall robot, 'DIANA', developed for this...

On modern construction sites, guidance and automation systems are increasingly applied to excavators. Recently, studies have been actively conducted compare the estimation results of bucket tip with motion measurement method boom, stick, sensor selection. This study selected measuring cylinder length bucket, directly arm, which commonly used in systems. A low-cost that can be attached detached excavator modular form was apply above methods commercial excavator. After selection, hardware...

Many solutions about the teleoperation of unmanned robotic excavators have been studied continuously. However, if that are remotely controlled were employed, probability an accident would be higher. For this reason, there many ways proposed to attempt reduce accidents in these dangerous situations. In paper, a novel methodology will introduced with focus on multiple themselves. The method details how one excavator can generated or expressed as obstacle Configuration-Space (C-space) respect...

To improve the stability of bipedal walking humanoid robots, we developed a method setting trajectory parameters using reinforcement learning on treadmill like testbed in real-world environment. A deep deterministic policy gradient (DDPG) was used as algorithm. By improving reward zero moment point (ZMP), optimum value and speed determined. The robot designed to measure ZMP mount weights upper body. In addition, manufactured operate at same robot. Reinforcement divided into unweighted cases...

Besides the difficulties in control and gait generation, designing a full-sized (taller than 1.3m) bipedal humanoid robot that can walk with two legs is very challenging task, mainly due to large torque requirements at joints combined need for actuators’ size weight be small. Most of handful successful robots this class exist today utilize harmonic drives gear reduction gain high compact package. However, makes cost such too thus puts it out reach most those who want use general research,...

Humans play sports to improve their athletic ability. The robot, especially humanoid is also able its performances, such as reaction speed and balancing, through robot sports. Therefore, robots have been developed performing various events soccer, marathon, fight so on. In this reason, Ski Robot Challenge was held in South Korea commemoration of the PyeongChang 2018 Winter Olympic Games. event an Alpine slalom skiing competition almost same rules human's but on a relatively short course...

The preview control walking method, which is commonly used in bipedal walking, requires jerk and ZMP errors as cost functions to generate body trajectory. Since the two inputs are dependent, optimization form trajectory performed simultaneously with weight factors. Therefore, it often seen that resulting rapidly changes on velocity according This eventually a torque actuator order perform such action. In overcome this problem, we apply method quadruped robot. Since, only targets minimize...

The purpose of this study is to increase the energy efficiency a biped-walking humanoid robot. A traditional robot walks with its knees bent, which places an unnecessarily large load on knees. Therefore, in study, toe joints developed. This allows it walk naturally by moving center gravity forward, without bending In addition, as acting knee joint reduced, four-bar linkage mechanism applied avoid excessive joint. structure distributes transmitted each axis 52.2-cm was developed above toes...

This paper discusses the development of electric linear actuators to be used in link structure robots. Linkage is parallel mechanism robots, which can solve disadvantage serial mechanisms that cannot share load with other and have low accuracy. In order for robots (especially walking humanoids), high speed required, unlike conventional focused on forces. addition, actuator should lightweight reduce moment inertia robot. will required torque motor power consumption Additionally, measure...

As the importance of emotional interaction between humans and robots continues to gain attention, numerous studies have been conducted identify characteristics effects HRI (Human-Robot Interaction) elements applied robots. However, no study has yet combined various into a single robot large-scale user experiments determine which element users prefer most. This selected four that facilitate attachment bonding animals: grooming, transfer, imprinting, cooperative hunting (play). These were...

The development of a mobile robot entails the utilization diverse components such as motors, sensors, controllers, and communication systems. Integrating all these can be time-consuming labor-intensive task. Typically, engineers need to search for motor drivers, controllers individually integrate them develop robot. However, most drivers on market are not specifically designed robots, other platforms specialized this purpose either. Therefore, paper aims facilitate robots by providing...

The robot's decision-making is an essential component of the autonomous robot. To fulfill this component, accurate position estimation robot a fundamental prerequisite. Localization determines relative within map environment, and existing mobile robots have been extensively studied. However, localization still challenging for humanoids because bipeds' movement not as stable robots, camera view frames oscillate from side to side. Developing with high accuracy under above-limited constraints...

In this paper, we propose an effective localization method with only a stereo camera that has obstacles using particle filter. When flow planning rather than robot scanned map, the error of increases when there is obstacle. To solve problem, First, two types obstacle recognition method: "Image Split Obstacle" and "Obstacle Image" through image processing Opencv contour function. Afterwards, problems caused by filter weight calculation process new sensing model interval angle. addition,...

The process of learning good features for machine applications can be very computationally expensive and may prove difficult in cases where little data is available.A prototypical example this the one-shot setting, which we must correctly make predictions given only a single each new class.In paper, explore method siamese neural networks employ unique structure to naturally rank similarity between inputs.Once network has been tuned, then capitalize on powerful discriminative generalize...