With increasing competitiveness, the sophistication and flexibility of the chemical processes have increased considerably. This has significantly increased the complexity of the encountered control problem, thus leading to the development and implementation of advanced control strategies such as MPC, GPC etc. Practical experience with advanced control has demonstrated that process identification is an important factor in the resulting control performance and in the design of advanced controllers for industrial processes; considerable (80-90%) amount of time is spent in process identification. Identification is however, expensive in terms of time, energy and potentially off-spec product quality. While a persistently rich excitation with high signal to noise ratio is theoretically preferred, operational, safety, environmental and economic considerations have to be taken into account during identification. The concept of "plant friendly identification" has recently gained attention in the process systems community. These objectives are inherently conflicting. Our research is focused on a synthetic procedure for input signal design that explicitly addresses these multiple objectives and constraints. We are also exploring applications of input and auxiliary design for model discrimination and fault diagnosis.