A GPU-based implementation of a sensor tasking methodology

In this paper, we present a graphics processing unit (GPU) based implementation of a receding horizon solution to the optimal sensor scheduling problem. The optimal sensor scheduling problem can be posed as a Partially Observed Markov Decision Process (POMDP) whose solution is given by an Information Space (I-space) Dynamic Programming (DP) problem. In references [1], [2], we proposed a simulation based stochastic optimization technique that, combined with a receding horizon (RH) approach, obviates the need to solve the computationally intractable I-space DP problem. In this paper, this RH sensor tasking approach is implemented using GPUs allowing us to greatly increase the number of simulations that we can perform to estimate the gradients in the stochastic gradient descent underlying the technique. This allows us to drastically reduce the variance of the technique, thereby greatly improving its performance. The technique is tested on a 48 object space situational awareness (SSA) problem and it is shown that the average uncertainty in state of the objects is reduced over hundred times when using the GPU based RH sensor tasking strategy as opposed to a myopic policy.