Double-Spreading Modulation Scheme Picks Up Where CDMA And TDMA Leave Off

Multiple-access wireless systems are under constant pressure from customers demanding higher data rates and increased service quality. Unfortunately, regulatory requirements limit the bandwidth and output-power levels available. In turn, this minimizes the throughput possible for any given transmission scheme. The throughput also is restricted by the minimum signal-to-noise ratio (SNR) needed to meet application specifications.

At present, code-division, multiple access (CDMA) and time-division, multiple access (TDMA) are the two most popular schemes in operation. Even they possess technical limitations with respect to flexibility, scalability, and noise immunity, however. To tackle these pitfalls, Nanotron, located in Berlin, Germany, came up with a solution. The company developed a new multiple-access method called multi-dimensional, multiple access (MDMA).

Based on double-spreading principles, the technique is capable of providing a low-cost path to increased throughput. This would occur by optimizing use of the allotted spectrum and improving noise immunity. MDMA also offers dynamic flexibility to meet future needs.

The subscriber base for current wireless communications channels is growing exponentially. Consequently, such channels are fast reaching their limits in terms of data throughput per user and overall quality. A strong demand for service provisioning also is becoming the mantra for providers, so that they can charge varying rates according to usage. CDMA and TDMA face numerous technical challenges when it comes to satisfying these requirements.

CDMA, for instance, transmits to all subscribers simultaneously, separating the customers using orthogonal spreading techniques. A positive aspect is that it offers minimal interference suppression via system gain. But this causes the service quality to suffer, as part of the channel capacity is dedicated to achieving this gain. According to Manfred Koslar, founder and president of Nanotron, the system's flexibility also is limited when it comes to meeting future demands.

For its part, TDMA sends data time-sequentially, providing each user with a certain time slot. While these systems are flexible, they're also highly susceptible to reflections. Eliminating the resulting signal degradation requires the precise estimation of channel parameters, to avoid limiting distance and failure rate. As the data rate and distance increases, the channel estimation becomes increasingly complex and expensive.

MDMA's double-spreading technique fully enhances channel capacity by integrating the quality of CDMA with the flexibility of TDMA. The method is based on the allocation of symbol energy over the entire bandwidth, continuously (frequency spreading), while dispersing this energy in the time domain. This results in the transmission of constant power over time. Energy levels can be controlled to meet the regulated symbol-energy limits. They can be adjusted so that the channel power used for each spectral line will not be higher, or lower, than stipulated. Doing so optimizes the otherwise limited channel capacity.

Based On FM Pulses At the core of MDMA is the use of "chirp" signals. These are defined as frequency-modulated pulses of constant amplitude, with duration T. Within this time, T, the frequency is rising or falling monotonically between a lower and a higher frequency. The difference between the lower and upper frequency represents, approximately, the bandwidth of the pulse. Though not restricted to linear modulation, linear will be assumed for the purpose of explanation.

Used since the 1940s, chirp pulses have one key feature. When a linear-frequency-modulated pulse is compressed via a dispersive filter, frequency modulation is converted into amplitude modulation, or vice versa. As these pulses are sent over an air interface, they represent broadband pulses, with a long duration and spread in time. In order to achieve a much higher energy density, a dispersive filter is used to compress the distributed energy at the receiver. The shape represents the shortest accessible pulse for a given bandwidth. As a result, the pulse gives the highest resolution on the time axis for a sequential system.

The transformations are bidirectional, exhibiting the same energy and spectral energy density. Only the distribution of energy on the time axis varies. The chirp pulses can overlap, due to the superposition law, with the compressed pulses separable on the time axis.

Hardware testing has been used to demonstrate the benefits of chirp-modulation. One of its features is that it combines the advantages of both correlative signals and time-domain-based TDMA systems, without causing a reduction in flexibility and capacity. In addition, chirp modulation makes it possible to keep the short-term average power constant and lower than the maximum value allowed. Furthermore, it's relatively easy to accomplish in hardware, making it a viable option for new transmission systems (see the figure).

Frequency spreading by time-continuous signal processing can be done using quasi-Dirac pulse shaping with bandpass filtering. If time-discrete signal processing is applied, frequency spreading can be achieved by up-sampling. The symbol rate is controlled to create the system gain and attain desired performance. For each symbol rate, the total maximum frequency bandwidth is used.

Time spreading can be accomplished using a dispersive surface-acoustic-wave (SAW) filter for time-continuous processing. In time-discontinuous systems, folding with a correlative sequence will do the spreading. The received signal, including noise, distortions, and multi-path reflections, will be compressed in time by folding it with the conjugated complex correlation sequence.

In analog circuitry, this is accomplished by a dispersive filter complementary to that used in the transmitter. The pulse sequence at the output of the compression filter supplies very precise channel coefficients. This allows the application of all known equalizers for high symbol rates. Frequency must be compressed during the last step. A sample-and-hold or an integrate-and-dump block can be used for this purpose. In general, the processing can be done digitally or by using analog techniques.

MDMA offers many advantages over CDMA and TDMA standards. Unfortunately, most providers are now flocking to established standards in order to take advantage of the installed base. Regardless, MDMA's low cost of implementation still makes it a viable alternative.

For more information, contact Manfred Koslar at +49 (0) 30-399 954 0, or go to www.nanotron.de.