Driven by artificial intelligence technology, industrial production is accelerating its transformation from traditional "automation" to "intelligence". The industrial AI intelligent robot control system developed by Beijing based technology enterprise Dongtu Technology has recently made its public debut. It can convert human natural language instructions into robot motion commands through the ability of artificial intelligence big models, and control robot work through human language, greatly reducing the threshold for robot application in industrial production lines. This is also the first intelligent controller in China to achieve industrial AI.

Rotate the red phone case 90 degrees and place it in the box. "At the demonstration site, the staff gave voice commands to an industrial robot in front of them that looked like a robotic arm. After a brief pause, the robot that "understood" the command lifted its robotic arm and picked up the red phone case on the stage, rotating it while placing it into the box next to it.

The system operator only needs to say one or two simple voice commands, and the system can quickly complete semantic analysis in the "brain", convert human language into "language" that the machine can understand, that is, executable code, and send it to the robotic arm to perform the corresponding work.

Industrial controllers are like the 'brain' of industrial robots and automated production lines. In the past, in the development and design process of traditional industrial controllers and robots, in order to mobilize the robot's work through the "brain", complex code writing, motion control, robot workflow design, work path planning and other development processes required developers to spend a lot of time.

Nowadays, with the introduction of technologies such as artificial intelligence big models, it used to take professional engineers about 3 days to develop robots, but now it only takes 5 to 10 minutes to complete, greatly reducing the difficulty and complexity of development. Engineers on the front line can free themselves from the tedious programming, development, and design work, concentrate on effectively interacting with robots, and optimize production efficiency.

The biggest challenge of this system is to handle complex AI tasks while performing real-time motion control to complete them, "said the technician. This industrial controller system is embedded with the Intewell industrial operating system and industrial software, which is 100% independently developed by Chinese enterprises. The control cycle of Intewell industrial operating system is only one millisecond, and it integrates an AI large model inference engine. Based on the enterprise's self-developed semantic algorithm, the robot completes corresponding action instructions by analyzing and parsing a series of parameters and indices. In order to improve the communication efficiency of information between multiple devices, the system has upgraded from traditional communication mode to shared memory communication mode, ensuring the information transmission efficiency of robots in completing complex tasks such as "understanding instructions" and "real-time control", and ensuring system stability and reliability in harsh industrial environments.

It is reported that industrial grade embodied robots based on Dongtu Industrial AI controllers have entered industrial enterprises for testing and optimization, and will soon be put into commercial use.