Communications

Engineering Tasks

• System Architecture
• Baseband Signal Processing
• Analog and Mixed-Signal Design
• Control Logic Design
• Advanced Technology Research
• Optical Networking Design

Baseband Signal Processing Design

All modern communication systems use digital signal processing techniques. Benefits include increased throughput, reduced bit error rate, and greater stability over temperature and process variation. Central to signal processing is the representation of a band limited analog signal as a sequence of samples equally spaced in time. This technique is commonly referred to as sampled data processing. Algorithms acting on the sample stream not only yield the correct output, but also produce it at the correct time.

Historically, communications engineers treated modeling and design as separate activities. Modeling was defined as developing specifications and validating against requirements (via capture and simulation) while design was defined as implementation, programming, and synthesis. Modeling was devoid of any notion of an implementation or design.

As communications systems have become more complex, this separation has led to non-optimal designs and unacceptable time-to-market results. In addition, communications engineers used general-purpose programming languages that were not designed for communication or signal processing tasks. This is inefficient because it is difficult to capture and visualize a model of a concurrent system in a procedural, text-based language like C. Further, text or C-based specifications can be ambiguous and add translation errors.

A Better Approach: Model-Based Design

An improved approach would be to begin with end in mind. Start by considering the design aspects, such as implementation, programming, and synthesis, at the beginning of the modeling stage. The MathWorks calls this merged approach model-based design. Model-based design anticipates that multiple sample rates in the model result in multiple clock rates in the design. Frames correspond to memory, buffers, and buses. However, for model-based design to be effective, the design environment must support the concepts that will be implemented in the design phase. In this case, the concepts to be captured and simulated include: concurrency, time, framing, multi-rate, and discrete-time sampling. Simulink® software with Signal Processing Blockset™ software supports model-based design of baseband signal processing and offers verification links to implementation tools when you enter the design phase.

Next