Radars work on the principle of transmitting and receiving radio waves to ascertain the position, direction and speed of distant objects. Modern radar systems require high performance, high resolution and high anti-jam features that classical analog radar cannot provide. In today’s digital radars, front-end functions such as beamforming, pulse compression and Doppler processing are also performed in the digital domain. As such the performance demand on the digital processor continues to grow. Also, because radars need to work with a large range of energy that is reflected back (from objects nearby as well as those very far away) the processors at the heart of these systems must consequently deal with very large dynamic ranges. The greater precision of output required, along with the need to support larger dynamic ranges of data, make a 1.25-GHz floating-point engine like the TMS320C6678 ideally suited for radar processing. Variable size, floating point and Fast Fourier Transforms (FFT) are key in radar systems and perform very well on a multicore DSP. Depending upon the system needs, the use of DSPs can lower the number of processing boards required and power dissipation of the radar by an order of magnitude.
Designers of radars are under pressure to differentiate their product by meeting ever tighter size, weight and power (SWaP) constraints, while supporting increased processing performance and functionality. As radar systems find more and more commercial applications such as weather radars, air traffic control systems, mapping of terrain and emerging automotive applications, it appears these constraints will continue to grow. Use of a small form factor, power efficient, scalable device such as TI’s multicore DSPs can be can be an important part of a comprehensive strategy to meet SWaP goals.
So the question to readers is... Are the performance improvements large enough to push designers towards a digital system (including retrofitting existing applications) or is the front end analog infrastructure and architecture so ingrained that it’s hard to uproot. That is, will most near term improvements to existing radar systems be incremental upgrades?
Great to see that TI is pushing floating point DSP's to higher and higher speeds and now into multicore devices, floating point applications are numerous and after a bit of a hiatus it's good to see floating point performance being pushed forward in a number of TI devices.
Tom
http://www.mitydsp.com
Hi Tom,
Thanks for the note. Yes, we are really excited about performance push on the floating point performance across TI devices. We are looking forward to a continued push on both fixed and floating point instruction sets as we move forward. As we couple that with more cores and higher GHz, the hope is that more applications will find their way to these power efficient multicore DSPs.