Researchers at the University of Michigan have recently developed a new method that allows different types of wireless networks to operate in the same space. In practice, their approach is remarkably simple: one network essentially says “excuse me, please yield” to the others.
Currently, the frequency band used by wireless local area networks (Wi-Fi) is shared with popular systems like Bluetooth and ZigBee. These devices often congregate in the same locations. However, these three technologies cannot effectively coordinate the use of signal channels, making interference a common and hard-to-prevent problem. Furthermore, because Wi-Fi itself uses radio channels of varying standard widths, even different generations of Wi-Fi devices sometimes fail to coordinate signal exchanges effectively. Both issues can lead to reduced transmission speeds and connection drops during network use.
As early as 2011, Professor Kang Shin of the University of Michigan’s Department of Computer Science and then-graduate student Xinyu Zhang (now an assistant professor at the University of Wisconsin) had already begun tackling this problem. Last July, they developed a software called GapSense, which uses specially transmitted energy pulses to regulate and coordinate data transmission for Wi-Fi, Bluetooth, and ZigBee. Professor Shin stated that GapSense is ready to be installed in devices and access points, and the work can begin immediately if a standards-setting organization or mainstream vendor is willing.
For countless phones, tablets, and PCs used in homes, offices, and public spaces, Wi-Fi is a data lifeline. Bluetooth, while using a slower wireless protocol, consumes far less power, making it more suitable for connecting peripherals. ZigBee, an even lower-power system, is often found in home automation, healthcare, and other specialized devices.
Professor Shin points out that although the three wireless protocols inside a device all have mechanisms to coordinate their time on the air, the trouble lies in the fact that each one’s specific approach is entirely different from the others.
Professor Shin explains: “This means they cannot speak a common language to communicate and understand each other.”
He elaborated further. They all use Carrier Sense Multiple Access (CSMA) technology, which commands a device to pause transmission if it finds the radio channel in use. But this still cannot prevent all interference.
The core of the problem is that Wi-Fi, Bluetooth, and ZigBee are out of sync with one another. These issues occur sporadically because one type of network starts up faster than the others. For instance, a Wi-Fi device using CSMA might not realize there’s a risk of collision with another transmission, even if a nearby ZigBee device is ready to start sending data. According to Professor Shin, the root cause of these conflicts is that ZigBee’s startup speed is 16 times slower than the time it takes Wi-Fi to leave idle mode and transmit data.
Professor Shin notes that making ZigBee boost its performance to keep up with nearby Wi-Fi speeds would defeat its original purpose of transmitting and receiving small amounts of data with extremely low power to ensure long battery life.
As for Wi-Fi devices, they currently struggle to communicate effectively even among themselves. To achieve faster speeds, each new generation of Wi-Fi standards continues to expand the spectrum blocks it uses. Professor Shin says the hidden danger of this approach is that if an 802.11b device tries to tell the rest of the Wi-Fi network that it needs to send a data packet, an 802.11n device using a 40MHz channel might not receive the signal at all because the 802.11b device uses a channel width of only 10MHz. In other words, the 802.11b device becomes a “hidden terminal” on the network. The inevitable result is collisions between data packets from the two devices.
To distinguish the frequency bands used by all these different types of devices, Professor Shin and Zhang developed a new communication model. GapSense uses a series of energy pulses separated by gaps, relying on the varying lengths between pulses to differentiate message types. Only after confirming a channel is free to complete the task will a device begin data transmission. Moreover, this signal can be sent not only before a transmission begins but also slipped in between data packets.
It appears that GapSense can significantly improve the performance of Wi-Fi, Bluetooth, and ZigBee. Collisions between different networks can slow network speeds and even cause connection drops or call interruptions. To determine the changes GapSense could bring to wireless networks, Professor Shin and Zhang created a simulated office environment with moderate wireless traffic for testing. Results showed that the initial collision rate between ZigBee and Wi-Fi was 45%, but after using GapSense, this figure dropped sharply to 8%. According to a just-released press release, tests targeting the “hidden terminal” problem showed that using GapSense caused the original collision rate of up to 40% to plummet to nearly zero.
Additionally, GapSense has another potential use: lowering the standby power consumption of Wi-Fi devices. Currently, in the working mode used by Wi-Fi devices, the idle receiver must be ready to accept traffic whenever it hears a signal from the access point. Professor Shin states that with GapSense, a wireless access point can send a series of repeating pulses and gaps that a receiver can recognize even while operating at a very low clock frequency. This means the receiver can confirm an access point’s repeated attempts to send data without fully leaving idle mode. Professor Shin claims this feature alone can reduce the existing power consumption of Wi-Fi devices by at least 44%.
Installing GapSense would require firmware and driver updates for all hardware, including devices and wireless access points. Professor Shin estimates that because most manufacturers will not choose to do this for devices already in use, the technology’s real-world adoption may have to wait for a hardware refresh cycle.
A patent