Stanford University’s Andrea Goldsmith once remarked that, “It is generally not possible for radios to receive and transmit on the same frequency band because of the interference that results.”
The statement accepts the assumption that radios can only operate in half-duplex mode, which is something that researchers have, until recently, also assumed. The advent of full-duplex mode has the potential to render this assumption obsolete.
Below, we chart how full-duplex wireless systems could completely upend present half-duplex communications systems — from the way cell phones operate to how Wi-Fi networks offer connectivity — and pave the way for 5G.
The problem? Half duplex:
Half-duplex wireless devices are those that cannot transmit and receive signals simultaneously. Most wireless devices today are half duplex. This is because the signals a wireless device transmits are more powerful than the ones it receives. Owing in part to this, output signals in a half-duplex system are picked up by the device. This overwhelms the input signal and self-interference is created.
Consider the diagram below of a practical radio:
The receiving antenna (Rx) receives signals that the transmit antenna (Tx) produces. However, there is disparity in the signal strength when Rx receives it as compared to when Tx transmits it.
This disparity results in the receiver and the transmitter encountering signals on the same frequency, which is the antithesis of the half-duplex system the receiver and transmitter are operating in. As a result, the signal is looped back to the transmitter. This looping back results in self-interference, and the receiver’s performance is degraded.
The solution? Full duplex:
If a system is able to cancel any significant self-interference arising from a device’s own transmission, it is a full-duplex system. If, given minimal allowances, both the transmitted and the original baseband signals are maintained, the loop back can be cancelled. Hence, self-interference is countered, as the diagram below demonstrates:
Scientists at Stanford University are developing technologies to achieve precisely this. They have designed a single-channel full-duplex wireless transceiver using a combination of RF and baseband techniques to achieve full duplexing. The design primarily uses signal inversion cancellation, a technique that employs an RF signal inversion circuit and is shown below:
The system supports wideband and high-power systems, which is unprecedented in its field. In practice, “the signal inversion technique alone can cancel at least 45dB across a 40MHz bandwidth,” state the Stanford researchers. This means that, at least in this specific example of a 40MHz frequency, a considerable 45dB of interference can be cancelled.
The wireless landscape and the medium of full duplex:
Current 4G systems will eventually evolve into the fifth generation of cell network technology (5G), which will ostensibly offer faster data speeds and greater reliability. 5G, however, is still not fully deployable.
Various technologies have been proposed as mediums for completely achieving 5G. These include but are not limited to millimeter waves, beamforming and massive MIMO (multiple input, multiple output).
Full duplex, however, could be the most realistic. Current half-duplex cellphones and base stations are forced to operate on different frequencies if users require simultaneous transmittal and reception of information. Transmittal on the same frequencies means full-duplex systems could double the capacity of signal transfer and halve the conversation time.
Additionally, full duplex’s effects would also counter current problems, which include but are not limited to:
- Hidden terminals: When nodes in the system communicate with a wireless access point but not with other nodes. This leads to simultaneous data transfers from multiple nodes, which causes interference.
- High end-to-end latency: When the time taken for a data packet to be transmitted from source to destination is unusually high.
- Fairness: When certain users are receiving more than their fair share of system access and resources.
- Congestion: The reduced quality of service when the system has to bear more data than it can carry.
As the scientists at Stanford further state, the technology could also help realize a real-time in-band control channel. Full duplex, then, is not so much a replacement for 5G as it is a medium through which 5G can progress from probability to certainty.
Challenges and drawbacks:
Scientists are optimistic as to the strength of their designs. But it is also important to recognize the severity of the challenges ahead. Minimizing self-interference requires cancelling at various levels:
- The antennae: Two transmitter signals are offset by half a wavelength, which means they cancel each other out.
- The analog circuits: A noise cancellation circuit implements self-interference cancellation. The transmit signal is entered as noise reference in the circuit. This reference is then subtracted from the received signal.
- The digital baseband: This system uses the received digital samples after the analog-to-digital conversion that occurs between transmittal and reception. The received samples are then compared with the transmitted samples to determine where the transmitted packet begins.
Different levels make for drawbacks to full duplex. Residue from signal inversion may potentially affect the reception of signals; and while full duplex could halve the conversation time, it could also double the amount of interfering streams, leading to inter-cell interference.
Additionally, the use of the full-duplex technology is dependent on data traffic traveling in both uplink and downlink directions, which would cause a huge strain on already busy systems.
The future outlook of full-duplex wireless systems:
Researchers at institutions such as the University of Washington, New York University, and Columbia University, among others, are focusing their current and future efforts on better understanding limitations like those mentioned above. The primary challenge is perfecting the self-interference cancellation to improve the full-duplex design.
Scientists at Stanford University, in particular, are also exploring the potential of generalizing their full-duplex design in order to apply it to MIMO systems.
Meanwhile, researchers at Rice University have created a design that suppresses the self-interference signal by primarily using passive suppression, and this is particularly suitable for Wi-Fi systems. This is unprecedented, as no previous designs have been able to facilitate Wi-Fi networks. So these scientists are now determined to build on their conclusive demonstration that full-duplex Wi-Fi networks are practical. This demonstration proved theories such as the rate of a full-duplex link increasing with the transmit power of a communication device and the application of digital cancellation after analog cancellation sometimes having an adverse effect on interference.
Full-duplex technology is also being employed by militaries. Organizations such as TrellisWare Technologies, which was awarded a USD15.7 million contract by the United States military in 2017, are focusing their efforts on designing a circuit that can apply self-interference cancellation techniques. With full duplex, militaries aim to reduce the spectrum required for civilian and military communications by enabling the transmittal and reception of signals simultaneously.
Additionally, given the myriad of communication systems employed by different agencies, full-duplex systems would counter the need for extensive frequency planning — the organization of which agency’s system is allowed to transmit and receive signals at particular times.
Consequently, with extensive research being carried out, and applications ranging from telephone conversations and Wi-Fi networks to military technologies and upgrading entirely to 5G, it is not a question of if full-duplex systems are possible. It is a question of when.
No matter the particular application, however, full-duplex systems would do more than just simplify current communication processes. They would also decrease signal transfer time and increase the amount of data that is transferable.
