In induction motors with Indirect field oriented control (IFOC) scheme, the rotor resistance (Rr) variation results in loss of dynamic performance. Hence the estimation of Rr is essential for high performance vector controlled drives. Neural based estimators are now receiving active contemplation as it has number of advantages over conventional techniques. The learning algorithm of the neural network determines its estimation speed, accuracy of estimation, and viability for digital implementation. In this work an enhanced learning algorithm is proposed with superior performance involving less computational complexity. Also the proposed neural based estimator has been implemented in TMS320F2812 digital signal processor (DSP) and Spartan 3E field programmable gate array (FPGA) and the results are presented. Multilevel inverter (MLI) has been used as it is an attractive alternative over conventional two level inverters in medium and high power applications. The enhanced performance of the MLI fed green drive with the proposed Rr estimator is presented. The impact of estimation time and estimation accuracy of the two implementation schemes on the MLI fed drive has been explored. The drive performance with DSP and FPGA based estimator has been analysed for a wide range of variations in Rr. © Research India Publications.