Human Robot Interaction

        It has been stated by Liu (2007) that technically, it is quite challenging to create co-robot behaviour that works in a deterministic and structured environment, co-robots have to operate in stochastic and unstructured environments. The basic research question for being addressed in this study was to make sure that co-robots function safely and efficiently in uncertain environments. The concentration of the study has actually been several things including setting up an analytical framework for robot-human systems and establishing a methodology for designing a robot behaviour for addressing the basic issue. A multi-agent system was introduced by the author for modelling the robot-human systems. For addressing uncertainties during robot-human interactions, a unique control and planning architecture was introduced by the author, which a module of cognition for human motion and behavior estimation, a long term planner for ensuring robot efficiency, and a short-term planner for ensuring the safety of real-time under different uncertainties which are present in the system (Liu, 2007).

           The cognition module was further discussed by Liu (2007), which involves online adaption and offline classification of different human behaviours. Optimization problems’ optimal control for long-term and short-term robot planning is covered in the study. Optimization issues, in a cluttered environment, are non-convex and nonlinear, thus they are quite tough to solve in real-time, which might delay the response of a robot in emergency situations. Online rapid algorithms are created for handling problem: CFS or algorithm of convex feasible set for optimization at long-term and SSA or safe set algorithm for optimization at short-term. Particularly, non-convex issue of optimization is transformed by the CFS algorithm into a set of issues of convex optimization that can be resolved online, which converges in a few iterations and also seems to run faster than conventional solvers of non-convex optimization (Liu, 2007).

            HRI’s current status was reviewed by Sheridan (2016) together with present issues of research for the community of human factors are explai9ned. It was concluded by the author that HR is an evolving field. And specialized robots under teleportation of human have proven effective in medical application and hazardous environments, as have specialized robots under the supervisory control of human for space and industrial tasks. Study in the areas of autonomous cars, intimate collaboration with machines, human control over different robots, and social interaction with them is still primate. Human robots’ efficacy still is not proven. Now, HRI is applied in almost all tasks of robots including, military operations, policing, package fetch, education, agriculture, surgery, undersea, aviation, space, and manufacturing tasks (Sheridan, 2016).

            In accordance with a study of Lasota et al (2017), making sure that the safety of humans is a significant consideration within the field of HRI. It doesn’t just include preventing collisions among robots and humans operating in individual space, all possible ways in which a person could be harmed should also be considered, ranging to adverse psychological factors to physical contact which results from dangerous and unpleasant interaction. This work collection was classified by authors into four categories including psychological factors, prediction, motion planning, and control safety. Recent work was discussed by authors in each of the category, several sub-categories were identified, and discussed how these methods can be utilized in improving the safety of HRI. Gaps in the present studies were discussed by Lasota et al (2017) and future directions were suggested for future work. By developing an organization field categorization, it was hoped by authors for supporting future development and research of new technologies for a safe interaction, along with facilitating the utilization of these methods by authors in the community of HRI (Lasota et al, 2017).

        In accordance with the research of Yu et al (2015), different rehabilitation robots have direct interaction with all humans. Actually, the control design is seemingly based on the model of the actuator with consideration of the dynamics of interaction. It includes compensation of human interaction, friction compensation, and is seemingly enhanced with an observed of disturbance. In fact, such a scheme is capable of enabling the robot in achieving impedance of low input while working in a mode of human-in-charge and achieving accurate tracking of force when working in a model of force control. Because of interaction with humans, the design of the controller must also meet the requirement of stability. And the outcomes can be extended to other assistive and rehabilitation robots which are driven with compliant actuators without many issues (Yu et al, 2015).