**4. Humanoid robot**

A humanoid robot is a robot having two legs, two arms, the shape of a human body, a trunk, and a head. Usually, it is associated as the robot with the appearance of a full human body and has the ability to walk, for example, Honda ASIMO, HRP, and HUBO robot. These three examples have similarity with its appearance as well as in mechanical design. The research in humanoid robot was initiated around the 1970s in Japan after the development of "Honda P3" by Honda Motor Co., Ltd. The purpose of this development was to build humanoid robots that can walk stably and mimic how humans walk. Since then, many research groups in humanoid robot pursued developing practical humanoid robot. At that time, Japan and South Korea are probably the leading countries in the research of humanoid robot. However, in literature, it seems the research problems in humanoid robot become broader topics and diverse from control walking, grasping, visual recognition, social interaction, virtual simulation, intelligent robot, and so on, including human-robot collaboration and interaction. Here, the author discusses and highlights humanoid robot toward industry 4.0, in particular human-robot collaboration and human-robot interaction. Other aspects of intelligent robot and virtual robot simulation will be briefly presented.

At present, commonly there are two types of actuators used in humanoid robots: First, motor or servo type with harmonic drives, for example, as in Honda ASIMO [64], DRC-HUBO+ [65], and Valkyrie NASA [66], and second, hydraulic type as in Atlas robot [67] and PETMAN [68]. The significance of the two actuators is very different in the sense that the first group of robots has slower response than the second group. The hydraulic actuator has a greater torque relative to the same size of electric motor, and thus, the robots that use hydraulic system are comparatively free from insufficient joint torque problem, while the robots that use electric motor have some problem with insufficient joint torque [65]. Despite of various actuators being used in humanoid robot, the trend of humanoid robot development is the robot that is lightweight with slim body. This design would make it easier for the robot to maneuver and perform certain tasks.

In industrial application, a semi-humanoid type robot such as Baxter robot is well known as a collaborative robot, and it is used in various industry applications to perform certain tasks. Although this robot is not a full body of a humanoid robot, most of it appears as a human except the fact that the robot has no legs to walk. Indeed, the advantage of full-body humanoid robot is that it can maneuver more easily in a complex terrain. However, in industry application or in other manufacturing industries, the terrain is usually simple with flat terrain, but the obstacles are more complex and dynamic. The general comparison of different robots including humanoid robots is presented in **Table 3**. It is given in **Table 3** that the humanoid robots are commonly used for research, technology demonstrator for specific tasks, or for human-robot interaction. This shows that the humanoid robots so far are not intended for competing with other robots in industrial applications. The use of semi-humanoid robot for collaborative robots in industrial application is more practical than that of a full body of humanoid robots. In fact, the full-body humanoid type of robot is commonly used for research only so far in order to solve practical engineering problems. For example, DCR-HUBO+ in [69] is used to solve the challenging problems of simple tasks such as debris removal, door opening, and wall breaking in the event of the DARPA competition.

The success of humanoid robot in real-world environments is largely dependent on the ability to interact with both humans and its environments [70] in which the humanoid robot has some form of awareness to the real-world context. Hence, the robot's perception is a key issue for performing high-level tasks such as understanding and learning human-robot interaction. This perception can be detected from the high-level features of human facial expression and body gestures [71]. The perception systems are proposed in [71], but the variety of robotic software architecture and hardware platforms would make the customized solutions hardly interchangeable and adaptable for different human-robot interaction contexts. Another aspect of learning (in control point of view) for the humanoid robot is, for example, in the situation when the robot is falling to the ground. At this circumstance, the robot must be able

*Computer Simulation of Human-Robot Collaboration in the Context of Industry Revolution 4.0 DOI: http://dx.doi.org/10.5772/intechopen.88335*


#### **Table 3.**

*General comparison of various robots and its applications.*

**Figure 11.**

*Different types of humanoid robots: (a) NAO humanoid, (b) pepper humanoid robot, (c) Sanbot robot, (d) Toyota partner robot, (e) Honda ASIMO, and (f) HUBO humanoid robot.*

to get up immediately with certain self-learning process or automatic learning system. Some of the existing humanoid robots are given in **Figure 11**. They are commonly used for research as in human-robot interaction, control methods of bipedal walks, indoor localization, and navigation. Moreover, some of them are also used for education [72], entertainment [73], and for home service.

The Cloud technology is one of the key components in the new industrial paradigm of industry 4.0. In relation to humanoid robot, one of the potential applications of Cloud technology is to provide collective robot learning, that is, robot sharing trajectories, control policies, and outcome [74]. One good practical example of Cloud technology application in humanoid robot is in the development of simulation of humanoid robot to complete certain tasks [75] which is part of the DARPA virtual robotic challenge. In [75], the existing Cloud technology was combined with Gazebo simulator for simulating humanoid robot. This developed robot simulation is not only applied to virtual humanoid robot in action but also to other specific challenging environments that must be handled by the humanoid robot. Another example of robotic simulation in the context of humanoid robots is given in [76] where V-REP robotic simulation was used as a testbed to observe how virtual robots would behave in completing certain tasks from given commands by humans. This research was related to teleoperation method based on human-robot interaction by mimicking human's movement visually for which Baxter robot would learn the movements. Again, robotic simulation would

*Computer Simulation of Human-Robot Collaboration in the Context of Industry Revolution 4.0 DOI: http://dx.doi.org/10.5772/intechopen.88335*

be a key component in robotic research and development, in particular in the field of humanoid robot. Moreover, this opportunity is due to the new emerging of Cloud computing that can be incorporated with robotic simulation.
