Hardware and software techniques to linearize the frequency sweep of FMCW radar for range resolution improvement.

摘要

The climate has been globally changing and sea level rise has been identified as a noteworthy indicator. One of the largest unknown in the sea-level rise predictions is the change in the mass of Greenland and Antarctica ice sheets. It has now been shown that Greenland and Antarctica have lost ice over the past 5 to 10 years and if the Greenland ice sheet completely melted, it would lead to a sea level rise of about 7 m. Surface monitoring method and the output accumulation method are the two ways to measure mass-balance of ice sheets. Ice cores and pits were used to gather information on accumulation rates but there could be a 20--24% uncertainty in the data. Remote sensing is a more practical approach for collecting these accumulation rate data. FMCW radars combined with Synthetic Aperture Radar (SAR) techniques can produce lightweight, cost and power effective sensors which can give high-resolution and high-quality images. However, nonlinearities present in the frequency sweep of the FMCW radar causes range resolution deterioration. Since the nonlinearities are more prominent over longer distances, the application of FMCW systems gets restricted to short distances. The primary objective of this thesis was to linearize the FM sweep of the FMCW radar. Techniques both in hardware and software were explored. The hardware technique involved correction of the Voltage Controlled Oscillator's tuning curve. A modified voltage ramp was fed to the VCO to produce a linear sweep of the output frequencies. The software technique involved estimating and eliminating nonlinearities from the resulting beat signal. Theory of sub-band linearization was analyzed and developed in Matlab. The beat signal was divided into sub-bands and frequency analysis of each sub-band was performed. Frequency deviation of each sub-band was estimated and corrected. Simulations worked successfully. The algorithm was then implemented on the beat signal produced by delaying and mixing the VCO's chirp. This was the VCO used for the hardware correction. The algorithm further improved the beat signal by removing a lot of its nonlinearities.