Channel Measurement and Resource Allocation Scheme for Dual-Band Airborne Access Networks
MetadataShow full item record
The aerial base stations (ABSs) can be quickly deployed to provide emergency communications and airborne network infrastructures. How to ensure wide coverage, reliable links, and high throughput for ground users under the conditions of limited onboard power supply, large propagation distance, and restricted frequency resource is a critical and challenging issue. In this work, we propose a hybrid-spectrum scheme for ABS-based airborne access networks, named dual-band aerial access (DBAA) where the ABS employs both the UHF and S-bands to provide connectivity for ground users. The DBAA can improve coverage range and reliability by taking advantage of the preferable radio propagation characteristics of the low-frequency band and meanwhile improve network throughput by exploiting the large spectrum bandwidth in the high-frequency band. Following the cross-layer approach, we first conducted a measurement campaign on the large-scale fading of the air-to-ground (A2G) channels at 785 and 2160 MHz simultaneously. We installed an ABS with two antennas on an airship that hovered at several altitudes from 50 to 950 m. We measured the signal power attenuation from the ABS to a ground terminal that moved in rural, suburban, and urban scenarios with the horizontal distance up to 70 km from the airship. Based on the measurement data, we establish the large-scale fading channel model for ABS at different operating frequencies. Then, we design the joint spectrum-and-power allocation algorithm to maximize the network throughput for the dual-band airborne access network. We evaluate the performance of the optimal resource allocation based on the proposed channel model. The simulation results show that the DBAA scheme with the optimal resource allocation can achieve substantial performance improvement in comparison with the single-band solution given the total spectrum bandwidth and onboard power supply. - 2013 IEEE.
- Computer Science & Engineering [825 items ]