Interference mitigation still confusing
The best-known example today of wireless interference concerns Sprint Nextel’s problems with public-safety users, which spawned a massive reconfiguration of 800 MHz airwaves. But interference is a growing problem both inside and outside the network for all operators, as systems compete for a finite amount of spectrum.
Everything from background noise — such as microwave ovens — to the electronic noise caused by base stations themselves contributes to poor voice signals, slower data speeds and latency problems. Network operators have used filters in base stations or constructed more towers to address interference. But receivers have become broader and cover more spectrum than ever, making it more difficult for filters to keep interference out, and disgruntled local governments are curbing the buildout of new communications towers.
Interference becomes more complex in CDMA-based technologies — the basis of both CDMA2000 and wideband-CDMA/high-speed data packet access (W-CDMA/HSDPA), 3G technologies that may very well be deployed by public safety in the 700 MHz band — as they are inherently vulnerable to fading and multipath. This vulnerability becomes more pronounced as CDMA cells are loaded with signals.
Network interference leads to lower data transfer rates, causes base stations to broadcast more power at each handset to counter noise and the number of handsets that can be supported simultaneously is reduced. All of this means greater expenses for operators.
“The economies are so big in this space that operators cannot ignore [interference],” John Thomas, founder and CEO of TensorComm, a developer of interference cancellation technology (ICT). “The operators are now going out with so much conviction about it. A forty percent improvement in capacity means a big capital expenditure savings for a lot of these operators.”
Hence, network operators are looking for more sophisticated interference mitigation methods. Now the question is: Do network operators address interference in the handset, in the base station or both?
Each approach has pros and cons, experts say.
“My understanding is that most of the carriers are favoring interference mitigation through the handset,” Andy Fuertes, analyst with Visant Strategies said. “If the technology is integrated, it can be done relatively inexpensively. They want to avoid the cost of putting more antennas on the base station.”
Relatively inexpensive signal processing technology is becoming more prevalent in the cellular world as operators try to boost the performance of existing networks without the need for upgrading existing base stations. Soon, such technologies could find their way into homeland security, public safety and other sectors where faster data rates are coveted.
Venture-capital backed TensorComm, which has researched network interference for years, has developed silicon that it says mitigates interference unique to CDMA-based systems. CDMA networks inherently suffer from having to shout over a noisy environment. More noise in the network requires the base station to raise the power level so that the signal can be heard over that noise. The company’s technology is capable of canceling interference from all channels of a signal as well as from adjacent base station sectors and from multiple paths within the serving sector, Thomas said. Doing this not only improves the quality of service on a voice call, but it also improves the data rates and reduces the number of base stations necessary to cover an area.
Thomas said field trials with a CDMA operator yielded significant improvement in network capacity. Field data collected by Hutchison 3 UK, a W-CDMA/HSDPA operator, concluded that interference is present on 3G networks and is discrete, identifiable and cancelable. Post-processing of the collected data showed 40% to 60% improvement in handset performance for mobile devices enabled with ICT.
But TensorComm must convince baseband suppliers to include its receiver chipset-based technology in their products. It’s an ambitious goal considering the handset suppliers are looking to keep costs down as much as possible, even though the cost to implement TensorComm’s technology is less than 3% of a single silicon in a handset.
“Cost is not the issue here. The issue is educating the industry,” Thomas said. “You have to have the operators convinced of this value, and so many operators are now starting to understand what the interference looks like on their networks, and they are now communicating that very strongly to their handset suppliers.”
Fabless semiconductor company Magnolia Broadband has been doing the same type of evangelizing with operators as it pushes its single-chip solution — dubbed DiversityPlus — designed to take advantage of two separate radio frequency signals to improve network capacity by 40%, increase geographic coverage by up to 60% and double data rates from the handset back to the network. The beam-forming technology directs energy off the handset, lowering the noise and interference in adjacent cells, said Larry Wasylin, Magnolia’s vice president of marketing and sales.
“If you approach a handset OEM, they’ll say it’s nice technology, but they don’t want to increase the cost of the phone by one cent because they are under constant pressure to reduce the cost of the phone,” he said. “The only way to do that is to sell the benefit to the carrier, and it will tell the handset suppliers to integrate it into the phone. … The incremental cost to the handset is in the dollars range. That’s minimal compared to the benefit the operator will see in terms of capacity and data-rate improvement.”
Although Magnolia initially developed DiversityPlus with the CDMA2000 cellular standard in mind, the company now is applying it to UMTS, Wi-Fi and WiMAX, with other radio systems to follow. Multiple original equipment manufacturers have built demonstration handsets using the technology, which already is being used by South Korea wireless carrier SK Telecom. Magnolia expects to integrate its technology in commercial products this year, with the first products being data cards because the most immediate benefit is increased data speeds.
Talk to network infrastructure suppliers, and they’ll say that placing intelligent antennas in the base station is the most efficient method for mitigating interference, as handset solutions increase processing power, eat up more battery life and increase costs.
“The industry has known for a long time that it will be a hard sell to get handset-based solutions adopted for cost reasons,” said Joe Tarallo, director of systems engineering for Alcatel-Lucent.
The most talked about antenna solutions are beam-forming and MIMO (multiple input/multiple output) technologies. Beam-forming can be used to strengthen weak antenna signals. Beam-forming, as the name implies, enables antenna transmissions to be dynamically shaped to avoid the interference that weakens them. The technology, like the handset interference solutions, doesn’t require any changes to the CDMA standard to implement it.
Alcatel-Lucent this year will be rolling out beam-forming technology in its CDMA 1x base stations for the North American market. The move is driven by customer demand for capacity and coverage improvements in high-traffic areas, and the potential to gain an improved link budget of 3 dB. (A link budget is the accounting of all of the gains and losses from the transmitter, through the medium to the receiver.) An intelligent antenna solution for 1xEV-DO is more problematic because EV-DO works like a time division system, said Hank Menkes, chief technology officer for Alcatel-Lucent’s wireless business.
The general drawback to beam-forming is that the technology has a limited distancing signal on a narrowing area. That means if a wireless signal is bouncing off buildings or moving around too much, there isn’t much of an improvement.
“In the worst case, there is no improvement,” Tarallo said.
But beam-forming is merely the first step for Alcatel-Lucent and other vendors. MIMO appears to be the Holy Grail, but its timing is unclear as it requires a change to the CDMA standard and an entire ecosystem of support, from the chipset supplier on down. However, MIMO is built into the WiMAX standard.
Based on the concept of multi-antenna arrays deployed at both transmitter and receiver locations, MIMO takes advantage of a traditional weakness of wireless technology by harnessing the once-unusable multi-path propagation that occurs in a single antenna transmission and using it to create additional transmission channels within the same frequency. When companies talk about MIMO in the modern sense, what they usually mean is a system with two antenna arrays at the transmission location and two antenna arrays inside the receiver, whether that is a modem, handset or other device. This translates to a doubling of bandwidth capacity and more efficient use of a single frequency.
Most network operators are resistant to MIMO. “Moving to MIMO is not wanted by mobile carriers at this point. That means putting up more hardware in towers,” Visant’s Fuertes said.
But as operators move to the next iterations of technology, the incorporation of orthogonal frequency division multiple access, or OFDMA, networks, they will be hitting the Shannon’s Law limit in terms of spectral efficiency, Alcatel-Lucent’s Menkes said. “The only way to improve the signal-to-noise ratio is through fundamental improvements to the modulation scheme through intelligent antennas,” he said.
However, Richard Lowe, head of Nortel Networks’ mobility and converged core networks group, said MIMO would improve capacity of OFDMA networks by a factor of three on the low side, five on the high side. The vendor has taken the lead on MIMO, working with chipset suppliers and device manufacturers to integrate the technology.
“The economies of scale will go into WiMAX, and by the time 3G operators deploy OFDMA, they can take advantage of that,” he said.
The road to faster data speeds
User throughput (kb/s) | ||
---|---|---|
Pedestrian | Vehicular 19 mph | |
Current receiver (RAKE) * | 433 | 595 |
Type 2 future receiver (LMMSE equalizer) | 976 | 988 |
Type 1 dual antenna receiver (RD) | 1181 | 1220 |
Type 3 (RD + LMMSE equalizer) | 1651 | 1674 |
ICT ** | 1825 | 1899 |
*RAKE by itself will not provide meaningful data throughputs. For many channel conditions and geometries, advanced receivers will be necessary for any realistic applications beyond voice. **Interference is modeled as a true node B, per 3GPP 25-101 |
Definitions:
LMMSE: Linear minimum mean squared error
RAKE: The digital section of a CDMA receiver that permits the phone (or cell) to separate out the relevant signal from all the other signals.
Receiver RD: Receive diversity (dual antenna receiver)
ICT: Interference cancellation technology
Source: TensorComm