Research On Some Key Issues In MAC Layer Of Wireless Sensor Networks

Wireless sensor network, with no infrastructure, has broad application prospects in many fields. So far, CSMA/CA is the primary MAC protocol in most wire-less sensor network applications, but it doesn't work in large-scale and super-dense networks because of the extremely large number of collisons among the contending sensors. The MAC layer protocol design is the key technology in wireless sensor networks to achieve higher throughput, lower latency, and bet-ter fairness performance.In order to improve the performance of CSMA/CA protocol in large-scale and super-dense wireless sensor networks, this paper studies the key technolo-gies in MAC layer and controlls the collisions by reasonably allocating and scheduling the wireless channel resources. Innovative technologies are pro-posed in the following aspects:Firstly, the paper proposed a random channel access control mechanism in large-scale and super-dense wireless sensor networks based on a directional data transmission model. By enabling the sensors in different area to access the channel in different time duration, such random access control mechanism can effectively improve the network throughput performance by decreasing the channel contending level.In use of the distinct directivity in wireless sensor network, A directional data transmission model is constructed by expressly restraining the path of data forwarding. Based on such directional data transmission model, a space-time based random medium access mechanism is proposed. By enabling sensors in different area to access the channel in different time duration, the contending sensors are effectively reduced as well as the channel contention level. An in-terference model is proposed to analyze the interference and collisions from potential contending sensors in such networks. By reasonably confige the car-rier sense range of the sensors, interference from serious contending nodes are suppressed. Simulation results show that such method can effectively improve the network throughput performance by decreasing the collisions.Secondly, the paper proposed an efficient wireless channel contention level estimation algorithm based on the theorem of interval estimation. By reason-ably control the estimation cycle, the estimation accuracy and efficiency can be significantly improved. Such algorithm is the fundamental of the efficient channel access scheduling in super-dense wireless sensor networks with ever-changing channel contention levels.A biased estimation analyzing model for channel contention level estima-tion is constructed in this paper to analyze the distribution of the biased values between the real channel contention level and the corresponding estimation re-sults. Based on the theorem of interval estimation, a fast and accurate channel contention estimation algorithm is proposed to handle these biased estimations.By restraining the biased estimations into a fault-tolerant interval, the confi-dence level of the estiamtion reuslts is improved and by adjusting the estiamtion cycle the estiamtion efficiency can also be optimized.Thirdly, the paper proposed an adaptive contention window tuning strat-egy based on a normalized channel contention model. The contention window tuning amplitude is adaptively chosen in such strategy according to the channel contention level estimation results to improve the contention window tuning eff-icency and accuracy. Such strategy determines the efficiency of channel access scheduling in super-dense wireless sensor networks with ever-changing channel contention levels.A channel contention analysizing model is constructed to evaluate the con-tention window tuning strategy for different channel contention status. The an-alyzing model significantly decreases the evaluation complexity by normalized mapping different channel contention status into a reference case. Based on such processing, an adaptive contention window tuning strategy is proposed to dynamically set the contention window tuning amplitude according to the chan-nel contention level esitmaiton reuslts. The small backoff factor in such strategy improves the contention window tuning accuracty, and the continuous tuning processing in a single estimation cycle extends the contention window tuning amplitude, which effectively improves the contention window tuning efficiency towards to optimal value.This paper is aiming to achieve efficient communications with low colli-sions and high throughput in large-scale and super-dense wireless sensor net-works. Different collisions have been effectively relieved by three aspects of progressive studies, including the channel access control mechanism, the wire-less channel contention level estimation algorithm and the contention window tuning strategy. This paper can provide both theoretical and technical support for advanced wireless sensor networks.