Signal Processing and Communications
Advances in semiconductor technologies have made it practical to implement sensors and smart devices that capture, process, and transmit signals of increasing bandwidths. In order to gain an edge in the market, products need to handle audio, image, video, and data signals efficiently. Complex algorithms must be implemented in hardware so that requirements in speed, power, and cost are met. Hardware subsystems need to be interfaced properly with software subsystems. Signals flowing through networks inside and outside of a device must be simulated at the right level of abstraction to avoid an unnecessarily long verification process while still achieving desired accuracies.
MathWorks Consulting Services leverages industry background and technical expertise gained from working with hundreds of companies to help you quickly model, test, and deploy your signal processing algorithms.
Developing and verifying FPGA implementation of signal processing algorithms
With industry experience in embedded system design for high performance signal processing, MathWorks Consultants help you convert your floating-point algorithm into fixed-point algorithm, design appropriate hardware architecture, and generate HDL code for FPGA implementation. We guide you through HDL synthesis, cosimulation with third-party HDL simulators, and FPGA-in-the-loop verification. MathWorks Consultants also help you optimize your algorithm to reduce quantization effects and hardware resource utilization, and to meet real-time processing requirements. We teach you how to automate steps that minimize the time required to test incremental changes to an existing design.
Analyzing performance via system level simulation
MathWorks Consultants help you model your system at an appropriate level of abstraction via algorithmic simulation, time-based simulation, event-based simulation, or any combination thereof. We have broad experience in simulating various layers of communication systems and in modeling effects such as clock jitter in analog-digital interfaces.
Minimizing simulation time to test complex algorithms
MathWorks Consultants work with you to identify opportunities to parallelize your algorithm and reduce simulation time by using a GPU or a PC cluster. We assist you in setting up an automation framework to use a PC cluster to do parameter sweep and Monte-Carlo analysis for your system.
MathWorks Consulting Services works with you to:
- Deploy your signal processing algorithms on FPGA and acquire the know-how to go quickly from algorithm to FPGA implementation
- Model your signal processing systems in a suitable architecture to achieve accurate and efficient simulations
- Set up test framework to verify your designs with confidence
Paul Peeling is a consultant engineer who specializes in signal processing, machine learning, and code generation for embedded hardware. He works with MATLAB and Simulink users to develop algorithms and model systems in multiple domains, and deploy the code to real-time targets. Prior to joining MathWorks, Paul worked applying pattern recognition techniques to detect and combat online fraud. Paul has a Ph.D. in statistical signal processing from the University of Cambridge.
Related Conference Papers and Technical Materials
Solving Large Geometric and Visualization Problems with GPU Computing in MATLAB - Technical article
Prior Structures for Time-Frequency Energy Distributions - Conference paper
Poisson Point Process Modeling for Polyphonic Music Transcription - Conference paper
Simulink to Hardware (42:47) - Video
Profiling C Code Generated by MATLAB Coder (Technical article)