Computation reduction of smart sensors for EBSE using Mahalanobis sampling

Event-based sampling strategies allow for reducing the amount of communications between a sensor and the estimation module. This reduction is interesting specially when the sensors are linked by a shared wireless network. This paper proposes new event-based sampling methods based on the Mahalanobis distance concept. Combined with an Event-Based State Estimator, they provide the same level of performance than other advanced sampling methods but keeping the computation simple enough to be executed by a smart sensor. The mentioned improvements of the algorithm respect other proposals are validated in simulation.