We give a strategy for nonlocal unambiguous discrimination (UD) among N linearly independent nonorthogonal qudit states lying in a higher-dimensional Hilbert space. The procedure we use is a nonlocal positive operator valued measurement (POVM) in a direct sum space. This scheme is designed for obtaining the conclusive nonlocal measurement results with a finite probability of success. We construct a quantum network for realizing the nonlocal UD with a set of two-level remote rotations, and thus provide a feasible physical means to realize the nonlocal UD. nonlocal, unambiguous discrimination, linearly independent nonorthogonal states, positive operator valued measurement (POVM), remote rotation PACS number(s): 03.67.Hk, 03.65.Ta Citation: Chen L B, Lu H. Nonlocal unambiguous discrimination among N nonorthogonal qudit states lying in a higher-dimensional Hilbert space. SciSparked by the fast development in quantum information theory and its applications, an interest in the theory of general quantum measurement beyond von Neumann measurements has been arising. Quantum measurement is an inevitable part of all quantum information processing tasks. In many situations, the choice of an optimal measurement depends on the task to be performed. For instance, in the case of quantum cryptography [1] and quantum teleportation [2], a central problem is the capability of discriminating among different outcomes. In general, these outcomes appear as nonorthogonal quantum states, which can either be spatially close together, or be located at widely separated places of a quantum network. The formalism of positive operator valued measurement (POVM) has been developed to describe general quantum measurement and allows us to discriminate a quantum state in more general ways. There are, in fact, two different optimal strategies that have been developed to accomplish quantum state discrimination tasks:the "minimum-error discrimination (MD)", where each measurement outcome selects one of the possible states and the error probability is minimized; and the "unambiguous discrimination (UD)", where unambiguity is paid by the possibility of getting inconclusive results from the measurement [3]. UD was pioneered two decades ago by Ivanovic-DieksPeres (IDP) for finding the optimal probability of conclusive discrimination between two nonorthogonal states with equal a priori probability [4][5][6]. The proposal involves the coupling of the original system to an ancillary system via unitary evolution. Postselection (von Neumann projective measurement) of the ancilla induces an effective nonunitary transformation of the original system. By the appropriate design of the entangling unitary, this effective nonunitary transformation can turn an initially nonorthogonal set of states into a set of orthogonal states with a finite probability of success. Jaeger and Shimony [7] have generalized this result by assuming that the dimensionality of the Hilbert space of the system of interest is greater than two. In this