1: \begin{abstract}

2: Projected gradient descent and its Riemannian variant belong to a typical class of methods for low-rank matrix estimation.

3: This paper proposes a new Nesterov's Accelerated Riemannian Gradient algorithm using efficient orthographic retraction and tangent space projection.

4: The subspace relationship between iterative and extrapolated sequences on the low-rank matrix manifold provides computational convenience.

5: With perturbation analysis of truncated singular value decomposition and two retractions, we systematically analyze the local convergence of gradient algorithms and Nesterov's variants in the Euclidean and Riemannian settings.

6: Theoretically, we estimate the exact rate of local linear convergence under different parameters using the spectral radius in a closed form and give the optimal convergence rate and the corresponding momentum parameter.

7: When the parameter is unknown, the adaptive restart scheme can avoid the oscillation problem caused by high momentum, thus approaching the optimal convergence rate.

8: Extensive numerical experiments confirm the estimations of convergence rate and demonstrate that the proposed algorithm is competitive with first-order methods for matrix completion and matrix sensing.

9: 

10: \keywords{Low-rank matrix estimation \and Local convergence analysis \and Riemannian optimization \and Nesterov's accelerated Riemannian gradient \and Adaptive restart scheme}

11: 

12: %\PACS{PACS code1 \and PACS code2 \and more}

13: %\subclass{MSC code1 \and MSC code2 \and more}

14: \end{abstract}

15: