Broadband Signal DOA Estimation Using TCT Algorithm

The TCT (Time Domain Covariance Transformation) algorithm is specifically designed for Direction of Arrival (DOA) estimation of broadband signals, suitable for complex scenarios involving multiple sources and coherent signals. Its core principle involves decomposing broadband signals into multiple narrowband components, enhancing signal subspace identification capability through time-domain covariance matrix transformation, thereby addressing the performance degradation of traditional subspace-based algorithms in broadband environments.

The algorithm's key advantage lies in effectively handling coherent signals, avoiding subspace ambiguity caused by signal correlation that plagues conventional methods. Implementation typically involves three key steps: first, frequency-domain segmentation of array received data; second, calculating covariance matrices for each frequency bin and constructing focusing matrices; finally, extracting signal subspace through eigenvalue decomposition to complete DOA estimation. From a coding perspective, this requires implementing frequency-domain partitioning using FFT, followed by covariance matrix computation and joint optimization across frequency bins.

Compared to narrowband algorithms like MUSIC, TCT significantly improves estimation accuracy under low SNR conditions by aggregating broadband signal energy, making it particularly suitable for practical systems like radar and sonar where multipath effects exist. The computational complexity primarily stems from joint optimization of frequency-bin covariance matrices, representing a typical subspace algorithm balancing computational load and performance. Code implementation should focus on efficient matrix operations and parallel processing of frequency bins to handle the computational demands.