Abstract
Quantum computers have the ability to perform some problems that would be intractable on a classical computer. The fundamental building block of quantum computers is the qubit, which is continuous as opposed to classical bits. These qubits become intrinsically sensitive to the environments, and parameter drift makes them lose some of their performance.
This thesis is focused on the robust control of transmon qubits, a superconducting qubit architecture, which is one of the leading choices of qubits for implementing quantum computation. These systems are naturally many-level systems and need to be controlled in a way that prevents detrimental effects of these other levels.
Robust solutions can offer protection against some effects that impede the high performance required of a qubit to perform quantum computation.
In this pursuit, pulses robust to different sources of error, such as a detuning error or a control amplitude error, are presented.
This thesis treats single-qubit and two-qubit gates and analyses the control landscape by exploring possible solutions, and investigates the optimisation traps which prevent obtaining high-fidelity solutions. Also, an exploration of penalty functions which are used to encourage different beneficial properties of the pulse. For instance, penalising the populating of higher energy levels in the qubit.
Furthermore, these numerical solutions, which have higher complexity than other control techniques, are tuned in hardware on the IBM qubits. A scheme was used to ensure the control performs as expected without implementation errors.
The experimental results also verify the model's validity and show high performance.
Addressing problems in qubit control has applications in scaling quantum computers and extracting more performance from current hardware.