Target Coverage Problems in Radio Frequency Harvesting Networks

Internet of Things (IoTs) networks have wide ranging applications, where they are deployed to improve the productivity and efficiency of various industries. They will also help facilitate the creation of smart cities. IoTs networks contain sensor devices that are responsible for monitoring or sampling an environment or targets. The collected samples are then sent to a fusion center or cloud to be processed, visualized and acted upon by network operators. For example, these samples or information can indicate traffic congestion or pollution, which can then be used by city planners to optimize traffic flows and volume in a smart city. To ensure high sensing quality, these sensor devices must ensure target(s) remain monitored by a sensor device. To this end, of interest to network operators is complete targets coverage, whereby all targets in a sensing area are monitored by at least one sensor device. This coverage requirement is especially important for applications such as security and assets management, where sensor devices can be deployed strategically to monitor entries or exits of a building or expensive items such as a painting. Guaranteeing complete targets coverage involves a number of challenges. First, there is likely to be coverage outage if sensor nodes have a higher energy expenditure than their energy harvesting rate. Second, the energy arrivals at sensor nodes vary over time and space. This has a direct impact on the choice of sensor nodes selected to monitor targets. Moreover, the choice of sensor nodes may cause some of them to experience energy outage in future time slots. Also, the amount of harvested energy and their frequency of activation will determine the number of samples collected by a sensor node. Consequently, there is a need to optimize the energy used for targets monitoring or sampling and data transmissions. Third, the number of set covers used to monitor targets grows exponentially with the number of targets and sensor nodes. A key problem is identifying the set covers used over time whilst ensuring sensor nodes have sufficient time to replenish their battery and thus ensure long coverage lifetime. Fourth, in order to construct set covers, a gateway/sink requires the battery level information of sensor nodes, which it then uses to select those sensor nodes with sufficient energy to be active for a given time period. However, in practice, collecting battery level is expensive in large-scale IoT networks. Hence, it is desirable to construct set covers using imperfect battery level information as doing so minimizes communications cost.