The lifetime of wireless sensor networks (WSNs) has been studied extensively. Lifetime is application specific, and long lifetime is desirable in general. In barrier coverage, it is desirable to have a large number of barriers active as well. Energy harvesting is an active research area in WSNs. In this paper we study the barrier coverage problem in energy harvesting WSNs. We develop an algorithm that finds a repeating sleep/wakeup schedule that can provide the continuous operation of the maximum number of barriers in an energy harvesting WSN. I. INTRODUCTIONThe lifetime of wireless sensor networks (WSNs) has been studied extensively [4]. Long lifetime is desirable. For instance, it can make the deployment of a network along a difficult border much less cost-prohibitive as it would reduce the need for frequent maintenance. However, network lifetime is application specific and thus necessitates individual study for different applications.One WSN application is detection of intruders for infrastructure protection and border control. For this application, barriers made of sensors are set up to detect intruders' crossing of the barriers. The k-barrier coverage problem is to set up k disjoint barriers in a wireless sensor network with an objective of achieving the longest operation time [12], [13]. The network lifetime for k-barrier coverage is the longest time of having k operational disjoint barriers. The parameter k has a large impact on the quality of the barrier coverage: the larger the k, the more robust the barrier coverage.Harvesting energy in WSNs has become an active research area [2], [6], [8], [9], [14], [17], [16]. A recent survey on energy harvesting in sensor networks [17] observes that "energy harvesting techniques have the potential to address the trade off between performance parameters and lifetime of sensor nodes." It is obvious by looking at the harvesting and sensing costs [16] that a naive approach to barrier coverage will not work. In a network that gathers energy with vibrational sensors harvesting at most .1mW that also employs light sensors using 1.155mW , it is not hard to see that such a network cannot be powered indefinitely in a barrier coverage application. Energy harvesting rates are much lower than energy consuming rates in sensor nodes. With energy harvesting techniques, there is still research to be done on extending network lifetime and increasing coverage.