A compact and broadband microstrip antenna design using a geometrical-methodology-based artificial neural network

A new artificial neural-network-based methodology for a microstrip antenna design was studied and presented. The methodology is applicable to DCS, GSM, WLL, WLAN, large band planar antennas, and fractals. In this paper, we present five applications of this methodology, three of which are applicable in the WLL, 802.11a, and 802.11b antenna standards. The two others are broadband designs with 500 MHz and 1 GHZ bandwidth, respectively. All the antennas radiate an end-fire beam, and have compact sizes of 29 mm /spl times/ 25 mm, 10 mm /spl times/ 13.5 mm, 10.3 mm /spl times/ 17.2 mm, 35 mm /spl times/ 25 mm, and 35 mm /spl times/ 25 mm, respectively.