At present, the design and control of single qubits as well than coupled qubits with superconducting circuits have been achieved. Unfortunately, the coherence time of all these devices is too short resulting in too large error rates and preventing their application as scalable superconducting qubits. The shortness of the superconducting qubit’s coherence time is a direct consequence of the naturally strong coupling of the electronic system to its electrical environment in a solid state circuit. It is believed that physical sources of noise acting upon qubits are local which theoretically makes the task of large-scale quantum computation realistic. Suppression of noise by improving materials involved in the qubit fabrication is a difficult task. Thus, it is very attractive to explore the possibility to decouple the qubit from local noises by tuning the qubit control parameters in such a way that it becomes less susceptible to noise. Flux and charge qubits already use special symmetry points where the qubit is decoupled from noise in linear order. Operating the qubit at these “sweet spots” enhances the coherence time. At the same time these experiments indicate that linear-order decoupling is currently insufficient for running long quantum computations. In this project we propose to realise a novel class of topologically protected qubits that should be decoupled from local noises well beyond the linear order.