FMCW radars play an important role in autonomous driving and emerging applications, such as motion sensing and healthcare. To achieve accurate and precise target detection, FMCW signals with wider bandwidths, larger chirp slopes, and better linearity are preferred. In the context of autonomous driving, multiple vehicles may occupy the same radar band, causing interference. A frequency-hopping strategy inspired by bats has been proposed to mitigate the interference between different FMCW radars [1]. For loT applications, a multi-mode FMCW radar that supports various detection modes (e.g., a long-range, low-resolution mode, and a short-range, high-resolution mode) within a burst frame is preferred to accommodate different application scenarios and reduce power consumption. The above-mentioned applications put forward the need for fast convergence of the FMCW DPD to fit the unknown nonlinearity existing in the new frequency bands.