Laboratory based structural testing encompasses many experimental techniques. This thesis will focus on systems where actuators and controllers are used to perform experiments on a range of test specimens. Accurate control of these experiments is extremely important. The control problem is often further complicated by specimen degradation during testing. Adaptive controllers which can take account of specimen failure are therefore desirable. However conventional adaptive controllers are limited by their performance and particularly robustness under such challenging operating conditions. A robust adaptive controller presents significant advantages over both fixed gain or conventional adaptive controllers. A conventional model reference adaptive controller was modified through the use of localized eigenvalue analysis to form novel robust adaptive controller. This algorithm, the ρϕ modified MRAC algorithm, has several advantages over conventional model reference adaptive controllers. It is extremely robust in the presence of noise, non-linearities and other destabilizing influences. The nature of the reformulation of the algorithm allows the user to tailor the performance of the algorithm in the frequency domain. And, as a result, known disturbances can be excluded. The ρϕ modified algorithm was applied to three structural testing applications, each of which presented different control challenges. The ρϕ modified algorithm showed significantly improved performance over previous control approaches on all of these experiments. The synchronization subspace method, which is also developed in this thesis, was used to quantify this. To further demonstrate the performance of the algorithm it was used to control a real time dynamic substructuring experiment. Real time dynamic substructuring requires very accurate control as the stability margins are often exceptionally small. The pep modified algorithm was able to control a real time dynamic substructuring experiment even in the presence of considerable actuator noise. Overall this new algorithm gives significantly improved performance for structural testing applications. It could also potentially be applied to other applications where robust adaptive control would be desirable.