1. Trends in a society where robots serve
Looking ahead to the after-corona period, expectations for robots are increasing from the perspectives of non-contact, remote work, and further labor savings. Human-machine interactions are changing remarkably within such a trend. Readers feel that the changes occurring from such a trend are not temporary changes but can permanently revolutionize our worldview of the future.
The need to work from home and secure social distancing in various industries is a good opportunity to review the work that has been done manually. To this day, we have often heard that manual tasks would be better because they are more flexible when performed manually. However, while reviewing individual tasks today, the approach has changed to a consideration of whether the task could not be performed by a machine and how to make such changes possible. Such an approach is expanding to various sites, such as medical treatment, food processing, office operation, and research and development, in addition to manufacturing sites. Robots are attracting attention because they have the potential to perform a variety of different tasks so far done manually, unlike specially engineered machines to replace humans in specific tasks.
On the other hand, when trying to meet these expectations, replacing a specific task will become a significant challenge.
The first challenge is the task that requires nimble on-site judgment. People use their brains to adjust their actions according to the situation, even if it is a straightforward task. Also, if a person commits mistakes in a part of the operations, the person will consider how to recover and will adjust the operational procedure. Since it is difficult to replace these persons with machines, there are also ideas where a human assists the operation from a remote location to combine human flexibility with machines.
The second challenge is the task where humans and machines work together. For instance, persons share the work of cell manufacturing lines in factories, serving lines in food processing, and crop sorting lines in the agriculture. In order to leave some work to machines, some cases require safety and health considerations because the operating range of the persons and machines overlap. However, it is not easy for machines to precisely detect and avoid persons moving around. Furthermore, when considering human-machine collaborative operations, harmonization of actions between humans and machines is required, but support and intervention by machines understanding the conditions of persons are difficult.
The third challenge is the adjustment of the machine and auxiliary work, which remain even if the operation itself can be automated. Most machines require delicate settings and adjustments for reliable run, but the problem at many sites is that such operation depends on specific skilled persons. Since these adjustments and auxiliary work require engineering expertise, it is not easy to realize the processes combining multiple machines. Moreover, even with the progress in automation, there are some sites where handling simple tasks, such as feeding and exchanging components, raise issues between automation.
Looking back on the various issues that have emerged, we can see that a new collaborative relationship between humans and machines becomes necessary. So, a platform that not only “substitutes” human work but also “collaborates” between humans and machines and “harmonizes” humans and machines is required to integrate human intelligence and machine efficiency. OMRON believes so and has been undertaking research and development to realize such a platform, but we are entering an era in which we should accelerate the research and development.
2. Current state and issues surrounding robots
So far, we have discussed without distinguishing robots from other machines. Needless to say, robots are a subcategory of machines, but it also implies the nuance of machines used for general purposes or multiple purposes when we dare say “robot.” This creates a gap between user expectations and the current level of technology. There are a variety of expectations that robot users have. In addition to reliability and safety as a machine, users expect robots to perform different tasks by a single one, teaching various operations easily, etc. Depending on the cases, it is also expected that robots can easily perform the intended task for the user even when the target, the work content itself, or the environment changes.
However, current robot technology looks far behind such user expectations in certain aspects. First, so that the robot can perform the task correctly, robot hands and the operating environment must be properly designed for the respective part and work content. Such a design requires an expert engineer and spoils versatility, an essential attribute of a robot because it consists of techniques and know-how tailored to the specific part and work content. Moreover, the operation of the robot must be implemented by repeating tests. Since the fight against errors is endless, the tremendous cost of equipment settings and adjustments is necessary to control the divergence caused by the part and environment to a simple and practical level. A jig that suppresses errors with the mechanical approach is useful to avoid this quickly, but it comes one step closer to a special-purpose machine and loses versatility because it is highly specialized to the target. As explained above, the reality is that today’s robot exists as one of the general-purpose components used in a special-purpose machine rather than a general-purpose machine. In other words, for effective use of a robot, it must be converted into a special-purpose machine after all, and special expertise, as well as tremendous time and costs, is required for design and adjustment. So, the user feels uncomfortable with the phrase “general-purpose machine.”
In order to overcome obstacles, there are several steps that must be taken.
The first step is to make the existing engineering chain move more smoothly. For example, make the fabrication of a jig easier, make the selection of component that forms the robot hand easier, and make an implementation of the robot movement on a simulator easier.
The second step is to realize the task by minimum implementation so that various parts can be handled. For example, it may be possible to generalize the robot to various objects and tasks by building intelligence that can make situational judgments like humans while using general-purpose robots and robot hands. Because of recent AI technology, these steps are becoming feasible even in actual applications with sophisticated modeling that reflects human expertise.
The issues that arise from the changes in the relationship between humans and machines mentioned at the beginning and the evolution of robots on the axis of versatility will become the mutual stimulation of science, technology, and society discussed in SINIC Theory.1,2) In this way, we believe that human-machine and human-robot relations in the post-COVID-19 era will form. We hope that this special issue will show readers some of the challenges for that.