Tesla humanoid robot evolves again! Commercialization expectations are greatly enhanced!

Recently, Tesla released a video showing new progress in the control and execution capabilities of Tesla's humanoid robot "Optimus". The robot is now able to classify objects relying solely on vision. You can even complete yoga moves. In this regard, Musk could not hide his excitement and said bluntly in the comment area: Progress.

Institutional analysis shows that Tesla humanoid robots have two major highlights in this evolution:

1. Visual perception ability upgrade: 1) Able to accurately realize limb position calibration and motion control, realize the classification and grabbing of color-difference objects; 2) With the ability to eliminate interference, Automatic error correction ability (not affected by external intervention, automatically righting inverted objects); 3) Motion control is trained by inputting video, is similar to the human learning mechanism, and is different from the Programming control (requires pre-programming, takes up memory); 4) is similar to the training mode of FSD V12, which means that the development experience on self-driving cars can be transferred to humanoid robots.

2. Upgrading of motion control ability: 1) In the process of grabbing and placing objects, the ability to control strength is strong, the entire movement shows delicacy and flexibility close to that of a human hand; 2) To achieve yoga stretching movements and maintain balance, this requires extremely high center of gravity distribution, movement positioning, and adjustment capabilities.

Overall, this latest upgrade of the humanoid robot sends a clear signal. Optimus can complete a series of complex actions through independent learning. Further, through the input and training of a large number of task materials in specific scenarios, Optimus can perform in general scenarios. It is only a matter of time to realize more complex working conditions and operations.

From the first generation of Optimus on AI Day in 2022 to the high-level operation and force control supported by the end-to-end algorithm seen in the current video, the progress of Tesla robots has been very obvious in just one year. The agency predicts that the mass production of Optimus may be the most certain industry trend in 2024. From an investment perspective, the industrialization of humanoid robots has generated demand for motion modules. Humanoid robots have as many as 25-50 joints, a higher degree of freedom, many types of parts and a large amount of parts, and the cost accounts for up to 50% of the whole machine. Taking Tesla Optimus as an example, it has a total of 40 actuators in the whole body, including 28 actuators in the trunk and limbs, and 12 actuators in the hands. It can achieve more than 200 degrees of freedom in the whole body and 11 degrees of freedom in the hands. The core components involved include reducers, servo motors, coreless motors, ball screws, sensors, etc.

Regarding the detailed investment logic of the robot sector, we have produced a special report "AI Empowers Robots to Open a Trillion Track".

It sorted out the latest progress of industrial robots and humanoid robots, how AI empowers robots, which links in the industrial chain are expected to benefit from the development of humanoid robots, and the list of core beneficiary companies, etc.

. Friends in need are welcome to request it for free.

The above is the detailed content of Tesla humanoid robot evolves again! Commercialization expectations are greatly enhanced!. For more information, please follow other related articles on the PHP Chinese website!