1Tbps optical fiber communication system
MERL engineers believe that with the popularity and depth of Internet applications, people need faster fiber-optic communication systems. The fastest optical fiber communication system available has a transmission rate of 100 Gbps, but through cooperation with the University of London, MERL has successfully implemented 1 Tbps data transmission on existing single-mode fibers. This multi-subcarrier technology can transmit 11 carrier signals simultaneously in a single-mode fiber. To prevent signal interference between carriers, a reference signal needs to be inserted before the signal is transmitted. Both the transmitting end and the receiving end know. After the reference number, therefore, the receiving device can make use of the change of the reference signal to compensate for the channel loss, thereby achieving compensation and equalization.
Channel separation technology
If we only have one microphone, but there are multiple people talking at the same time, is there a way to separate the speeches of multiple people separately? This is also a direction of MERL's current research, the purpose is to separate the mixed signals. MERL engineers said that when they did this research, they hoped that this method could be independent of the type of sound signal, so that it can be applied to different sound types and can distinguish any number of different sound sources. At present, the traditional optimization algorithm can achieve a signal-to-noise ratio improvement of about 3.1 dB, while MERL's end-to-end algorithm can achieve about 10 dB. In the optimization process of this algorithm, machine learning and deep learning methods are widely used.
For the research results displayed by MERL engineers, the entrepreneurs of the delegation generally went to the R&D gap between Chinese and Japanese companies, and also deeply realized the importance of grasping these forward-looking technologies that affect the future development direction, and also deeply felt a crisis. sense. Academician Li Peigen, who was inspected with the delegation, also mentioned the experience of taking the delegation to Wenzhou to inspect the development of local enterprises. He believed that Chinese enterprises generally only value the rapid transformation and production of research results, and the short-term return on income is long-term. Technical reserves and R&D are generally not of interest. If we do not change this situation, the lack of this belief will make the future Chinese manufacturing industry in crisis. We also hope that the government, enterprises and institutions can learn from the industrial development experience of industrialized countries such as Europe and the United States. The approach is to encourage companies to continue to increase research on cutting-edge technologies that affect future industrial development.