Minimizing Convergecast Time and Energy Consumption in Green Internet of Things

Real-time surveillance systems with green wireless sensor networks (WSNs) are vital for maintaining high energy efficiency in many situations. This paper considers a scenario utilizing green WSNs to monitor the situation of Internet of Things (IoT), which constitute one of the most crucial sources of electricity consumption in information and communications technologies (ICT). More specifically, we focus on optimizing the cluster structure to minimize the delay and energy consumption for aggregation convergecast in green WSNs. We first find the optimal value of the network cluster radius for minimizing the delay through theoretical analysis. We then propose a novel cluster network architecture in which clusters that are far from the sink are small, allowing inter-cluster data aggregation to be processed earlier, and clusters that are near the sink are relatively large to allow more time for intra-cluster data aggregation. Hence, the sensor nodes can be scheduled in consecutive time slots to reduce the number of state transitions, consequently achieving the goal of minimizing both delay and energy consumption. Simulation results indicate that the proposed Algorithm outperforms previously reported solutions in terms of both schedule length and lifetime, thereby demonstrating its effectiveness.

I. Introduction

Internet of Things (IoTs) [1]–[3] as well as cloud computing [4]–[6] leverage the ubiquity of sensor-equipped devices to collect information at low cost [7], provide a new paradigm for solving the complex sensing applications from the significant demands of critical infrastructure [8], [9]. It is vital to maintain high energy efficiency in data gathering [2], [6]–[8]. To this end, this work focuses on the design of a novel aggregation convergecast cluster structure to minimize both the required number of aggregation time slots (delay, for short) and the energy consumption. In WSNs, convergecast [10]–[12] is one of the most commonly used data aggregation patterns [10]–[12]. The Time Division Multiple Access (TDMA) scheduling approach is adopted in this work because it is more effective than CSMA, especially under medium to high traffic loads [13]. In the convergecast data collection process, the data sensed at all nodes should be aggregated and sent to the sink within the minimal number of time slots and with minimal energy consumption. However, it is well known that the minimal convergecast time (MCT) problem is NP-hard [14]. In addition, Wu et al. [15] have noted that a wireless sensor node will typically wake up periodically to sense the environment and produce new data or to receive or send data to the sink node or other sensor nodes, after which it will return to sleep mode. The corresponding state transition of the processor consumes a significant amount of energy. For instance, the ATMega128L processor of a Mica mote sensor requires 4 ms for start-up, and the RFM radio used by the Mica sensor requires 12 μs to switch between sending and receiving, whereas the raw bit time is 25 μs [15]. Thus, a proposed scheduling scheme should reduce the frequency of state transitions, thereby increasing the network lifetime.