Researchers from Virginia Tech used brain-inspired machine learning approach to increase the energy efficiency of wireless receivers
A team of researchers from Virginia Tech used brain-inspired machine learning techniques to enhance the energy efficiency of wireless receivers. The team’s findings published as, ‘Realizing Green Symbol Detection Via Reservoir Computing: An Energy-Efficiency Perspective,’ was awarded the Best Paper Award from the IEEE Transmission, Access, and Optical Systems Technical Committee. The research was conducted in collaboration with the U.S. Air Force Research Laboratory (AFRL).
The novel approach enables signals to travel from transmitter to receiver with the help of multiple paths at the same time. According to the team, their technology offers advantages as minimal interference and avoiding multipath fading over simpler. To minimize the inefficiency, the system uses artificial neural networks, which are computing systems inspired by the inner workings of the brain. Virginia Tech researcher Yang (Cindy) Yi stated that the receiver conducts channel estimation before detecting the transmitted signals and that artificial neural networks can be used to create a completely new framework through detection of transmitted signals directly at the receiver.
According to Matyjas, the technical advisor of AFRL’s Computing and Communications Division and an Air Force Research Laboratory Fellow, the approach can enhance system performance in situations where it is challenging to model the channel or to establish a straightforward relation between the input and output. The team suggested use of a special architecture called Echo State Network (ESN) as a method to train the artificial neural network to operate more efficiently on a transmitter-receiver pair. ESN as a recurrent neural network is capable of combining high performance with low energy. According to Wysocki, the chief engineer of the Air Force Research Laboratory Information Directorate, the approach enables to create a model that defines propagation of a specific signal from a transmitter to a receiver and makes it possible to establish a straightforward relationship between the input and the output of the system.