Comparison of polynomial network and look-up table for an avionics application

Summary form only given, as follows. The problem of pilot overload under missile attack is discussed. A pilot advisor is proposed to indicate whether the threat is serious and what action is best. A missile-aircraft simulator generates data for constructing a look-up table and a polynomial neural network. It is shown that the network is a powerful tool for reducing the cost of generating the look-up table. Further, the authors expect the network to be more cost-effective than a look-up table for complex situations requiring very large tables.<>