POPULATION CODING IN A NEURAL-NET FOR TRAJECTORY FORMATION

In this study we investigate the time evolution of the activity in a t opographically ordered neural network with external input for two type s of neurons: one network with binary-valued neurons with a stochastic behaviour and one with deterministic neurons with a continuous output . We demonstrate that for a particular range of lateral interaction st rengths, changes in external input give rise to gradual changes in the position of clustered neural activity. The theoretical results are il lustrated by computer simulations in which we have simulated a neural network model for trajectory planning for a multi-joint manipulator. T he model gives a collision-free trajectory by combining the sensory in formation about the position of target and obstacles. The position of the manipulator is uniquely related to the clustered activity of the p opulation of neurons, the population vector. The movement of the manip ulator from any initial position to the target position is the result of the intrinsic dynamics of the network.