Design Challenges for Future Wireless Systems
ANDREA GOLDSMITH
Department of Electrical Engineering
Stanford University
Stanford, California
A wireless communications system that provides a high-speed, high-quality exchange of information between portable devices located anywhere in the world is the communications frontier of the next century. The potential uses of wireless technology include multimedia, Internet-enabled cell phones; home entertainment networks; smart homes and appliances; remote learning and telemedicine; autonomous sensor and robot networks; and automated highways. The exponential growth of wireless communications worldwide, coupled with the potential impact of future wireless systems on people everywhere, has captured the attention of the media and the imagination of the public. However, many technical challenges will have to be overcome for the promise of wireless communications to become a reality. In this paper I will outline a vision for future wireless systems and then describe some of the inherent technical challenges to realizing this vision.
WIRELESS VISION
Wireless networks will enable people on the move to communicate with anyone, anywhere, at any time, via a range of multimedia services. Wireless networks will form virtual offices anywhere in the world from small hand-held devices that provide seamless telephone, modem, fax, and computer communications. Wireless local area networks (LANs) will connect palmtop, laptop, and desktop computers anywhere in an office building or on a campus, as well as in the corner cafe. In the home, LANs will enable a new class of intelligent home electronics that can interact with each other and with the Internet. Video teleconferencing will be possible between buildings that are blocks or continents
apart; teleconferences can include travelers as well, from salespeople who miss their plane connections to CEOs sailing in the Caribbean. Wireless video will create remote classrooms, remote training facilities, and remote hospitals anywhere in the world. Wireless technology will also enable developing countries to build up telephone and computer networks quickly and cheaply, effectively bypassing the high cost and delays associated with building wired infrastructures. Sensors with wireless transceivers will have the ability to self-configure into networks, make localized decisions, and relay information to centralized decision centers. Potential applications for sensor networks include environmental monitoring, energy-efficient temperature control, security systems, and automated highways.
THE WIRELESS CHANNEL
The wireless communications channel poses the biggest obstacle to building high-performance wireless networks. Because the fundamental characteristics and capacity limits of the wireless channel affect all aspects of the wireless network design, designers cannot just borrow from the networking technologies evolving to support the explosive demand for wired networks, which typically operate over fiber-optic channels with very high data rates and very low probabilities of signal corruption. Neither of these characteristics is enjoyed by wireless channels.
One key aspect of radio-wave propagation that affects system design is path loss, which dictates that the power in a radio wave diminishes with the distance it travels. Thus, the coverage area of a wireless system is dictated by how far wireless devices can be separated and still maintain reasonable levels of received signal power. In addition, signals propagating out from a transmitter may be reflected off surrounding buildings or objects and arrive at the receiver after some delay and with phase shift relative to the original transmission. This causes self-interference between the signal transmission and its reflection, which can give rise to large amplitude fluctuations (fading) in the received signal and interference between subsequently transmitted bits. Self-interference can severely degrade the quality of the received signal. A signal propagating through a wireless environment may also be blocked by buildings or other objects, resulting in a very weak signal at the receiver. In addition, because radio is a broadcast medium, all transmissions that share the same signal bandwidth interfere with each other, and this interference must be managed through the system design. Moreover, as devices move around, their signals and interference propagation characteristics are subject to random changes. In short, the wireless radio channel as a medium for reliable high-speed communication poses severe challenges to designs of wireless systems. Not only is it susceptible to noise, interference, signal blockage, and reflections, but the impediments also change over time in unpredictable ways.
CAPACITY LIMITS OF WIRELESS CHANNELS
The nature of wireless signal propagation, coupled with limited available bandwidth, severely limits the rate at which data can be sent over wireless channels. Although the fiber-optic cable that supports most wired network traffic can send data at a rate measured in gigabits (109 bits) per second, indoor wireless systems can only send data at the rate of megabits (106 bits) per second, and outdoor wireless systems at the rate of kilobits (103 bits) per second. Thus, the data rates for wireless systems today are from three to six orders of magnitude lower than the data rates of their wired counterparts, and the gap is growing. Given that end-users will expect the same performance for both types of networks and that applications for wireless systems will be mostly designed relative to wired network constraints (because wired networks are more prevalent), overcoming and adapting to the data rate limitations of wireless channels will be essential for the success of high-performance wireless systems.
Channel capacity is a fundamental characteristic of a wireless channel that defines the channel’s maximum possible data rate. One way to increase the capacity of a wireless channel is to use more bandwidth—a larger portion of the radio spectrum—because channel capacity is typically proportional to the bandwidth of the channel. Unfortunately, bandwidth is a very expensive and tightly regulated commodity, so increasing capacity by increasing channel bandwidth typically comes at a very high price. A less costly approach is to develop techniques that maximize channel capacity in a given bandwidth. However, the random propagation and interference characteristics of the wireless channel conspire to reduce this capacity. In fact, these characteristics make it difficult to derive the fundamental capacity limits in the first place. Thus, determining the capacity limits of wireless channels and developing strategies to achieve transmission data rates close to capacity remains an area for active research.
WIRELESS SYSTEM DESIGN CHALLENGES
Challenges to making the wireless vision a reality arise at all levels of system design, including the hardware, link, network, and application levels. Future wireless networks will require tightly integrated, adaptive designs that transcend the boundaries of these design layers to deliver the best possible performance given the random variations in the constraints of the overall system.
The challenge for designers of wireless hardware is to enable terminals with multiple modes of operation to support different applications. Desktop computers currently have the capability of processing voice, image, text, and video data for small, lightweight, hand-held devices; however, breakthroughs in circuit design will be necessary before multimode operation can be implemented. Because most people will not carry around a 20-pound battery, the signal processing and communications hardware of portable terminals must consume very little power. Many of today’s signal processing techniques that increase channel capacity and
mitigate channel impairments require a lot of processing power. Thus, breakthroughs in efficient algorithms, as well as in battery design, will be necessary to overcome the limitations of hardware power.
Another major design challenge will be overcoming the capacity limits, interference levels, and random variations of the wireless channel. Significant breakthroughs have been made in this arena over the last decade, driven mainly by commercial cellular technology. These breakthroughs include: multiple antennas at the transmitter and receiver to increase channel capacity; sophisticated coding strategies to correct channel-induced bit errors; multiuser detection techniques to reduce interference; equalization, spread-spectrum, and multicarrier modulation to reduce self-interference from signal reflections; adaptive modulation to optimize performance over time-varying channels; and dynamic resource allocation for sharing power and bandwidth among multiple users in the system as channel conditions and requirements change. Although these breakthroughs represent significant accomplishments, much more work will be necessary to improve wireless channels.
Wireless networking also presents a significant challenge. The network must be able to locate a given user among millions of mobile terminals and route a call to that user, which could be moving at speeds of up to 100 mph. The finite resources of the network must be allocated fairly and efficiently to meet changing user demands and locations. Today, a tremendous infrastructure has been developed for wired networks: the telephone system, the Internet, and fiber-optic cable, which should also be used to connect wireless systems into a global network. However, because wireless systems with mobile users are not likely to be competitive with wired systems in terms of data rate and reliability, the design of protocols to provide interfaces between wireless and wired networks with vastly different performance capabilities remains a challenge.
Wireless systems must support wireless applications, which may have very different requirements (e.g., voice mail and email). It is impossible to design a “one-size-fits-all” wireless network that can support all of the applications that exist today, let alone the applications that will evolve in the future. Moreover, it is impossible to guarantee fixed performance metrics (e.g., data rate or a harddelay constraint) for a wireless network because of the underlying random channel and network dynamics. Thus, wireless applications will have to be adapted to these dynamics to deliver the best end-to-end performance. For example, a wireless video application might require a data rate of 10 megabits per second for very high picture and sound fidelity. However, if the underlying network cannot support this rate, the resolution could be scaled back to a rate commensurate with system capabilities. Under extremely poor network conditions, the video might revert to a sound-only mode. Although ideally applications could always deliver high performance, the ability to scale back quality on the fly in response to degraded network conditions will be an essential characteristic of robust wireless system design.
Perhaps the most significant technical challenge to wireless network design is an overhaul of the design process itself. Most wired network designs are based on a layered approach; that is, the system hardware design, link design, network design, and application design are all created independently of one another with baseline mechanisms that interface between the design layers. Although this methodology leads to some inefficiencies and performance loss because the global design is not optimized, it also greatly simplifies the overall system design. For wired networks, because they have large capacity and good reliability, and because the performance loss resulting from layering is fairly low, the layered design works well. The situation is very different for wireless networks. Not only can wireless links exhibit very poor performance, but this performance, along with user connectivity and network topology, changes over time. In fact, the very notion of a wireless link is somewhat fuzzy because of the nature of radio propagation, and because of the dynamic nature and poor performance of the underlying wireless communication channel. High-performance wireless systems will have to be optimized for this channel and must adapt to its variations as well as to user mobility. Thus, wireless systems will require a tightly integrated and adaptive design that transcends hardware, link, network, and application layers. Given the underlying constraints and dynamics of the channel and network, as well as the application requirements, each layer of the system design, as well as across layers, will have to adapt for the system to deliver the best end-to-end performance.
SUMMARY
Wireless communications could have an enormous impact on our daily lives, not only in the way we communicate, but also as a technology enabler for other large-scale systems. However, many technical challenges will have to be overcome for us to realize this potential. This will require an interdisciplinary approach to wireless system design and creative thinking about the large-scale systems that wireless technology could enable.
FURTHER READING
Fasbender, A., R. Reichert, E. Geulen, J. Hjelm, and T. Wierlemann. 1999. Any network, any terminal, anywhere. IEEE Personal Communications 6(2):22–30.
IEEE Communications Magazine. 2001. QoS and resource allocation in the 3rd generation. IEEE Communications Magazine (Special Issue), February 2001.
Katz, R.H. 1994. Adaptation and mobility in wireless information systems. IEEE Personal Communications 1(1):6–17.
Negus, K., A. Stephens, and J. Lansford. 2000. HomeRF: wireless networking for the connected home. IEEE Personal Communications 7(1):20–27.
Noble, B. 2000. System support for mobile, adaptive applications. IEEE Personal Communications 7(1):44–49.