DoA Estimation Performance and Computational Complexity of Subspace- and Compressed Sensing-based Methods

We investigate the Direction of Arrival (DoA) estimation for small- and large-scale antenna arrays with a small and a large number of antenna elements, respectively. Two classes of algorithms are considered, namely subspace- and compressed sensing (CS)-based algorithms. We compare those algorithms in terms of both the DoA estimation performance and the computational complexity based on different parameters such as number of antenna elements, number of snapshots and quantization. From this comparison, we conclude that the subspace-based method ESPRIT is well suited for small-scale antenna systems while the CS-based method IHT is advantageous for large-scale antenna systems.