AMATAMA is currently researching and developing a hierarchical computing architecture for controlling next-generation humanoid robots, and the two companies will work together to accelerate research and development using Fuku’s generative AI algorithm technology.
Since August 2024, AMATAMA has been working on “Development of an innovative computer system for robots using hierarchical computing technology” after receiving the Go-Tech grant from the Small and Medium Enterprise Agency. This is a research and development initiative to solve various issues with the computer system that controls the humanoid robot, which is expected to be released on the market in a few years. This is not like conventional control computer systems that control the entire body with a single or several types of computers, but rather features a new architecture (structure) that “hierarchically” connects small computers distributed throughout the body, which work together for each role to control the entire body. This research and development will make it possible to significantly increase the number of sensors installed, reduce the amount of power used for control, or allocate it to more advanced control processing. In recent years, with the development of generative AI, it is becoming possible for this advanced control processing to imitate the functions of human cognition and judgment, which are called foundation models.
こちらもお読みください： Amazon Music & Audible to Join AFEELA In-Car Service
In the hierarchical computing (hierarchical architecture) that AMATAMA is researching and developing, computers assigned roles at each layer, such as lower, middle, and upper, must exchange information with each other. Furthermore, this information must be passed on in a form that can be interpreted by the AI of the base model running on the higher-level computer. This time, with the assistance of Fuku’s advanced generative AI algorithms and their construction and design, we will conduct research and development so that each of these computers at each layer can use AI to exchange information with the AI of the base model. This will enable multiple control computers in the humanoid’s body to work together using AI and perform functions similar to those of the human nervous system.
ソース PR TIMES
