Cost-Minimized Crowdsourced Spectrum Sensing

Abstract

Cooperative spectrum sensing which enhances the sensing accuracy is an important research issue for cognitive radio networks, especially in complicated environment. Considering the extensive use of mobile intelligent terminals such as smart phones and tablets, crowdsourced spectrum sensing, which assigns spectrum sensing tasks to mobile terminals, can take advantage of mobile terminals' cooperation and obtain the accurate sensing results. In this paper, crowdsourced spectrum sensing is studied to propose assignment scheme of spectrum sensing tasks in large geographical areas. There may be several kinds of terrains affecting sensing in large-scale regions. Hence, according to the terrains, we divide a large region into several sub-regions and introduce sensing effect function to evaluate the sensing accuracy based on the number of sensing sub-regions. Furthermore, considering energy consumption is an important issue which mobile terminals focus on, we use the relative energy consumption to evaluate the cost of mobile terminals during spectrum sensing. Then, we formulate the crowdsourced sensing problem to minimize the total cost while keeping sensing effect not lower than the predefined threshold to maintain sensing accuracy. Since the problem is NP-hard, a heuristic algorithm is proposed to solve the crowdsourced sensing problem. At first, our algorithm arranges all sensing tasks in a priority queue based on their urgency. Then, sensing tasks are sequentially assigned to terminals with higher energy to prolong their survival time under makespan and energy constraints. To obtain the lowest system cost, we introduce remaining time and reassign sensing tasks from high-cost terminals to low-cost terminals based on the remaining time. Simulation results show our algorithm achieves higher performance than the other algorithms.