This paper addresses fault-tolerant topology control in a heterogeneous wireless sensor network consisting of several resource-rich supernodes, used for data relaying, and a large number of energy-constrained wireless sensor nodes. We introduce the k-degree Anycast Topology Control (k-ATC) problem with the objective of selecting each sensor's transmission range such that each sensor is k-vertex supernode connected and the total power consumed by sensors is minimized. Such topologies are needed for applications that support sensor data reporting even in the event of failures of up to k - 1 sensor nodes. We propose three solutions for the k-ATC problem: a k-approximation algorithm, a greedy centralized algorithm that minimizes the maximum transmission range between all sensors, and a distributed and localized algorithm that incrementally adjusts sensors' transmission range such that the k-vertex supernode connectivity requirement is met. Extended simulation results are present...