A SVR Learning Based Sensor Placement Approach for Nonlinear Spatially Distributed Systems

Figure 7

Control performance comparisons with different sensor placement schemes: (a1)–(a3) using proposed sensor locations (eight optimal locations); (b1)–(b3) using eight evenly distributed sensor locations. (1) denotes catalyst temperature evolution profile, (2) denotes manipulated input signal, and (3) denotes catalyst temperature under steady state, where may be (a) or (b).