An investigation of two mediation strategies suitable for behavioural control in animals and animats

This paper gives details of recent experiments to determine the characteristics of two simple mediation strategies. We have chosen to compare two mutual inhibition strategies; the standard ‘n-flop’ model [Ludlow 76][Shackleford 89], and a scheme suggested by McFarland [McFarland 65]. To study the comparative features of each network we use TAG's ‘non-spiking’ neuron model ( SNF block model [Snaith 89a][Snaith 89b][Holland&Snaith 90a] ) both in simulation and on one of our Hilda series of mobile robots to show the mediation between simple ‘fight’ and ‘flight’ behaviours set up in conflict.