In the near future, the exponentially increasing demand for higher data rates will exceed the capacity of current coherent optical systems employing wavelength-division multiplexing (WDM). Once the physical dimensions of polarization, amplitude, phase, and frequency are exhausted, exploring the space dimensionemerges as an attractive solution to overcome this capacity limit. Space-division multiplexing (SDM) refers to the simultaneous transmission of different signals through independent spaces by using either single-mode fibers bundles, multicore fibers or multimode fibers. Among the challenges faced by SDM, the new technologies should maintain the trend of decreasing costs per bit while coping with the new channel impairments. As both multicore fibers in the coupled-core regime and multimode fibers introduce nonnegligible crosstalk between multiple parallel transmission paths, multiple-input multiple-output (MIMO) signal processing techniques are required. Besides mode coupling, the multiplicity of orthogonal modes in multimode fibers also induces modal dispersion and mode dependent loss, impairments that increase the complexity of the MIMO equalizer and limit the system capacity. In the scope of elastic systems,channel estimation based on digital signal processing (DSP) allows adapting remotely the operating conditions from the knowledge of the optical channel state. In the same way, future SDM systems must flexibly adapt the modulation format and the digital signal processing algorithms to improve the system performance. In this scenario, this project aims to investigate techniques for channel estimation in systemswith spatial multiplexing subject to several channel impairments. The project will resort to classic and machine-learning based techniques applied to experimental data captured in the host university.
News published in Agência FAPESP Newsletter about the scholarship: