We introduce a scheme to perform universal quantum computation in quantum cellular automata (QCA) fashion in arbitrary subsystem dimension (not necessarily finite). The scheme is developed over a one spatial dimension N-element array, requiring only mirror symmetric logical encoding and global pulses. A mechanism using ancillary degrees of freedom for subsystem specific measurement is also presented.PACS numbers: 03.67.Lx Quantum computation involves control at some level of a large number of quantum subsystems. The traditional circuit based approach is to have a set of subsystems for storing quantum information, denoted q-sites, which are fully addressable not only individually but in subsets [1]. It is known that universal quantum computation is possible, independent of subsystem dimension, when arbitrary single q-site unitaries and 2-q-site entangling gates are available [2, 3]. Yet addressing individual subsystems in large arrays is extremely challenging and can impose significant errors due to miss-alignment of the fields which unintentionally act on neighbors to the target (cross-talk) or miss the target. Thus as a mean to avoid it another option appeared: global control schemes [4,5,6].The philosophy behind global control is to reduce the interaction with the array of q-sites to require only global manipulation, implemented for instance through global fields homogeneously coupled with all q-sites. In general such global control schemes require a natural evolution (time step) for the array, and tailored sequences of global pulses which translate physical asymmetries in the array into control of particular sites or, equivalently, chronological control into spatial control. When the resultant evolution is a set of gates which act on small neighborhoods in parallel, this is also a quantum cellular automata (QCA) model. The models for global control have been so far concerned only with qubits [4,6], however with the development of higher dimensional computational models (using qudit computation and continuous variables(CV), also called qunat computation) that show advantages in terms of efficiency and robustness [7], the natural direction is to find a way to implement such models in a globally controlled fashion.The aim of this paper is to develop such a model in one spatial dimension, inspired by a previous protocol restricted to qubits [6] and recent results on globally controlled transport of qudits and qunats [8]. The main technical difficulty is the presence of more complex phases that appear in higher dimensions, in contrast with the {1, −1} phase in the qubit case, and solving equations for discrete variables which are defined modulo the dimension d of the logical q-site under consideration. Fortunately, in [8] we developed most of the tools we will need as well as revising some of the known results avail- * E-mail: gpazsil@ics.mq.edu.au able in the literature [9].The paper is organized as follows: in Sec.I we review some mathematical results on bases for the operator space particularly exploring a...