1: \begin{abstract}

2: 		Distributed functional scalar quantization (DFSQ) theory provides optimality conditions and predicts 

3: 		performance of data acquisition systems in which a computation on acquired data is desired.

4: 		We address two limitations of previous works: prohibitively expensive decoder design and a restriction to sources with

5: 		bounded distributions. 

6: 		We rigorously show that a much simpler decoder has equivalent

7: 		asymptotic performance as the conditional expectation estimator previously explored,

8: 		thus reducing decoder design complexity.  The simpler decoder has the feature of decoupled

9: 		communication and computation blocks.

10: 		Moreover, we extend the DFSQ framework with the simpler decoder to acquire sources with

11: 		infinite-support distributions such as Gaussian or exponential distributions. 

12: 		Finally, through simulation results we demonstrate that performance at moderate coding rates is well 

13: 		predicted by the asymptotic analysis, and we give new insight on the rate of convergence.

14: \end{abstract}

15: