Unmanned systems based on the continuous iteration of big data and artificial intelligence technologies are becoming a major trend in the development of new strategic technologies worldwide. A large number of unmanned systems have been practically applied in many fields in recent years. On June 25, a reporter from the Global Times went to the Institute of Information Science (IIS) of CEC to learn more about the latest development of China's cutting-edge intelligent technology in related fields.
When it comes to unmanned systems, people are most likely to think of humanoid robots. The first thing the Global Times reporter saw during his visit to the Smart Institute was a dancing robot. This robot is a basic platform for the Intelligent Institute's usual research on intelligent unmanned technology. Its biggest feature is the integration of the "Sea Sparrow" cognitive processor developed independently by the Intelligent Institute in the front end of the robot. According to the traditional way, the robot needs to collect a large amount of information back to the back-end processing, once the network bandwidth is limited, the robot is likely to be unable to respond in a timely manner; after loading the "puffin", the robot can carry out perceptual reasoning and lightweight information intelligence pre-processing in the front-end, and then send the pre-processed data back, significantly reducing the pressure on network communications. The network communication pressure is significantly reduced.
According to experts, usually robots are positioned outdoors through GPS and Beidou systems, while in urban buildings, deep forests and caves, there are weak navigation signals, so the unmanned platform needs to have the ability to sense the environment independently. The Global Times reporter saw at the scene that another small robot could realize dynamic perception of the unknown surrounding environment and independent action through sensors such as vision cameras, and demonstrated the ability to build 3D maps in real time and independent positioning. At the same time, the 3D map can also be updated incrementally as the robot moves.