This year we have witnessed several new humanoid robots that were introduced by some companies. One of the biggest trends we see in the media is Figure; an advanced humanoid robot presented by OpenAI. Figure is equipped with Generative AI that gives the robot the ability to recognize the surrounding environment and respond to human requests. Tesla also presented a new version of their humanoid robot Optimus, which the company is planning to use in their car factories in the coming years. Moreover, there are some companies such as Boston Dynamics that have a long history in humanoid robot development. Boston Dynamics has introduced a new version of Atlas that they are planning to release to the market within the coming years.
In the beginning, I want to go back in history and check similar trends in the past about inventions in robotics that were expected to make big revolutions in the industry. More than a decade ago Rethink Robotics, a company founded by Prof. Rodney Brooks from MIT; introduced its robot, Baxter. Baxter was a two-arm robot that could learn tasks simply by moving its arms. Rethink Robotics was expecting that this robot would allow small shops and factories to implement robot technology. However, in just seven years, the company went bankrupt because there were not enough sales of this robot. Similar to the case of Rethink Robotics, SoftBank Robotics introduced its humanoid robot Pepper 3 years after the release of Baxter. It was also expected to revolutionize retail businesses. The robot can interact with people which is ideal for customer services. However, three years ago, Softbank stopped the production of this robot because there were not enough sales. So will the current trend be a similar event that happened in the past?
I want to argue that several new factors will make the current situation much different from previous cases. First I want to analyze the failure of the two previous big hits; Baxter and Pepper. Looking at the design of Baxter, we will find its arms are big and harder to move. This was not attracted to small shop owners, who were the targeted customers of the robot; to use the user in the businesses. Maybe Rethink Robotics discovered this issue that had pushed them to redesign their robot system. They develop a single-arm robot, Sawyer; which is much smaller than Baxter. But it seemed that the move was too late. The development of hardware is very costly which was cause for the company’s financial difficulties.
SoftBank Robotics did not go bankrupt. It is still in business selling several other robots such as cleaning robots. If we look at the design of Pepper, we will find that it’s a little bit complex. The robot has two arms with a five-finger hand like a human hand. Even though the hands are fully functional hands that can grasp and manipulate objects, they were not used to perform any specific task or job. It rather seems that the hands were just designed to make some demonstrations and act similar to humans. This makes a robot over-engineered because there are a lot of components in the robot which makes it very complex and very expensive. Moreover, putting the robot in public areas made it very difficult to predict the scenarios. Sometimes people might talk to the robot in a way that the robot cannot interpret. Such limitations made Pepper not attractive to retail businesses since it might not work efficiently lowering the return on investment for implementing such a robot.
Looking at the recent development of humanoid robots. OpenAI is planning to use its robot Figure in the automotive industry. The company had agreed with BMW to deploy a flat of robots in its factories. A similar move by Tesla, in which they are planning to use their humanoid robot Optimus in their car factory. Agility Robotics is also introducing its simple humanoid robot to work in warehouses and distribution centers. Looking at all these robotic companies, we can find that they are planning to use their robots in a constructed and controlled environment like factories and warehouses. In such environments, it will be much easier to put the robot under control since the scenarios that the robot is expected to face will be easily predicted. It is different from the previous experience of Rethink Robotics and Softbank which had introduced their robot to small businesses and public areas. Such situations made it difficult to predict the scenarios that the robot might face. Moreover, introducing the robot to big clients such as BMW and Tesla will give the chance for the developers to test the robot's performance on a large scale. The result of the new humanoid performance will encourage other companies to buy and implement these robots into their businesses. It is similar to the introduction of the robot manipulator more than 60 years ago. The first robot was installed in an automotive factory in New Jersey, which was a controlled and structured environment. The success of the robotic manipulator led the path toward the expansion of robot installation in the automotive industry during the 1960s and 1970s. So we might be in a similar situation that will open the era of the huge development of Humanoid robot implementation.
Comments