Self Sustainable Cognitive Wireless Sensor Networks with RF Energy Harvesting

Cognitive Wireless Sensor Network (CWSN)’s offer a solution to the capacity problem for the next generation Beyond 5G (B5G) Wireless Sensor Network (WSN)’s. Secondary User (SU)’s access the Primary User (PU) network by finding unused spectrum and opportunistically accessing it, thereby increasing system capacity. This requires signal detection, analysis & decision making and spectrum exploitation techniques in order for the CWSN to work effectively. As a result, two acute problematic issues in a CWSN are: signal detection and power loss. The focus of this thesis is on efficiently solving the power loss problem, thereby creating a self sustainable CWSN. 

Harvesting energy from the ambient Radio Frequency (RF) signal has been shown as a viable solution to the power loss problem. Various schemes have been proposed including opportunistic and cooperative RF Wireless Energy Harvesting (RFWEH). Unfortunately, many of these models require RFWEH from low power sources, which is shown in this dissertation to be unsustainable in the long haul. 

This issue is thoroughly investigated and a novel hybrid RFWEH architecture for CWSNs is proposed. Simulation results show that the proposed model after harvesting energy from high power devices, such as TV and radio towers, results in energy sustainability. This is in sharp contrast to existing work, wherein nodes die out eventually and the network collapses. Furthermore, changing the RFWEH architecture to include a power switch, where individual nodes change RF energy sources, increases system performance by up to 72%. 

The design of sustainable MAC protocols for CWSNs using RFWEH is also proposed. These include a distributed, centralized and mobility-aware MAC. Simulation results show that the distributed MAC has a 90% node survival rate while keeping network parameters such as throughput, delay and packet loss within acceptable target limits. The centralized MAC outperforms the distributed MAC by up to 40% and the mobility-aware MAC has a 30-80% node survival rate, depending upon speed of individual nodes. 

Simulation results show that both the stationary (distributed and centralized) and mobile CWSN’s are self sustainable. To the best of our knowledge, this is a unique result as no comparable self-sustainable designs for harvesting energy from high power devices have been found in literature.