Merge of networking robot and artificial intelligenceRobot will be at the core of IoT society in the near future

by Masaru Nagura, Senior Fellow

 The time has come again to pay attention to robot. The extension of the concept of the robot with networking functions and the merger with AI establishes robot as the leading core of IoT society. “Network Robots”, which will network with other robots, systems or devices, will renew the conventional image of robots and will have function as a cradle for developing and enhancing AI.

 Network Robots are capable of acting more widely and flexibly compared with the single or non-networked robot by receiving a variety of information to take in their own AI. For example, the basis of IoT such as wireless senor networks equipped into a smart city in the future will collaborate and complement with Network Robots.

 More specifically, sensors such as cameras installed or equipped in the city streets will be communicated with “robot-cars” driven full automatically by AI and make traffic running smooth and securely. It can be argued that the robot-car alone can be judged as ‘single robot’ as a function but also the total robot-cars in combination with sensors installed in the city can be judged as “ total network robots”.

 We will see that the Network Robot acts as one unit collaborating with each other and keeping its own autonomy as well as keeping away from concentrated controls by a cloud system.

 It may be difficult to understand the appropriate image of Network Robot if you are only staying with the image of conventional standalone robot such as a humanoid or an animal type robot, because this new type of Network Robots can decide by themselves how to think and how to act freely.

 Networking of robots is quite useful for AI (the brain). The performance of AI depends on how much good quality of data for its development can be collected particularly in the field of deep learning which has recently become main-stream. It is quite sensible that you think it a good idea to collect and integrate this learning dataset for AI by using autonomous robots. This means that the utilization of data-collecting Network Robots will be the key to efficient development of AI.

 It will be related with the remark that from the cognitive science viewpoints growth of AI needs embodiment. As human beings have acquired knowledge making full use of five senses, it is considered indispensable that AI should also do trial and error with the feedback loop in order to acquire knowledge and concepts. It is nothing more or less than robot to give the physical functions necessary for active data-gathering activities for AI.

 Computer networks and AI will merge with robots. In IoT society in the near future, there remains great opportunity yet for Japan to play an active role as Japan still keeps advantages in the robot engineering.