Singlehop Collaborative Feedback Primitive: Threshold Queries in Wireless Sensor Networks

To speed-up the collaborative feedback collection in a singlehop wireless sensor network deployment, initiator can exploit the receiver-side collision detection. The initiator node queries a predicateP in its neighborhood in constant time by making all the nodes answer simultaneously. Despite the collisions the initiator is still able to infer useful information from a broadcast (an activity in the channel means the predicateP holds for at least one node, while silence indicates that P does not hold at any queried node). Using this primitive, we investigate the threshold querying problem, where the initiator has to deterministically and precisely learn whether P holds in the network at least threshold t number of nodes in singlehop of the initiator. To answer the threshold queries in an efficient fashion, an adaptive querying mechanism that dynamically re-groups the queried nodes in the network is proposed. We evaluate our method on Kansei testbed and carry out a number of simulations to contrast our approach with the traditional techniques. The experiments reveal that we can perform threshold queries much faster.