Logical Designs has a strong background in digital signal processing methods. Logical Designs was one of the first companies to show the capabilities of neural networks in the area of signal processing. Using back-propagation we trained a neural network to remove additive random noise from an EKG signal. We saw the use of the back-propagation neural network as a logical extension of traditional adaptive filter theory.

While specializing in adaptive techniques, we have learned how to build fixed coefficient filters to optimize your specific phase and frequency requirements. Since implementation cost is always an issue, these filters can be optimized with integer coefficients. These methods work with most traditional kinds of filters. We then add to these development capabilities, the ability to design nonlinear and neural network filters. Hybrid filter types can be designed with performance not achievable with any single method.

The analysis of your project requirements will determine what kind of filter or signal processing is required. Logical Designs can implement FFTs, DFTs, Gabor transforms, and deconvolutions.

We have worked with a variety of filter forms including Finite Impulse Response filters, Infinite Impulse Response filters, LMS and Kalman adaptive filters. Logical designs will build simulations for modeling the filter and verifying the filter performance prior to hardware implementation.

Logical Designs can also use these techniques in image processing applications. Image filtering, normalization, deconvolution and compression techniques, and neural network methods are some of the tools that can be applied to your image processing task.

Logical Designs is not limited to the theoretical aspects of signal processing. We have developed heuristic methods for handling some practical applications. Working with EKG signals we developed algorithms for detection of the QRS complex used to determine heart rate, and methods for baseline elimination for EKG signals that minimized the distortion of the desired signal. These algorithms were implemented in hardware on the TMS32010 processor.

We can implement the solution in custom designed hardware and software, or in efficiently coded assembly language on your hardware. We have built custom hardware and software for the TMS320 and DSP32 processors.

All designs, analysis and software are always completely documented and maintainable. Logical Designs has the ability and experience to make your project a success.