Spectrum Allocation For Real-time Data Collection In Industrial Wireless Networks

With the integrative development of industrial automation and information technology,industrial wireless sensor networks(IWSNs)have attracted wide attention of the business community.Compared to traditional wired network,IWSN can be deployed in high-temperature,high-humidity,and corrosive environments to monitor harsh industrial process;and can also be deployed in large scale industrial site for the ubiquitous perception.However,IWSN is still facing many challenges.The most prominent one is the real-time transmission required by industrial automation.Since ISM(Industrial,Scientific,Medical)2.4GHz unlicensed band only covers 85 MHz bandwidth,how to schedule the limited spectrum resource to meet real-time industrial data transmission,is still now a hot research issue.This dissertation focuses on the IWSNs for the monitoring of multi-stage industrial processes.Taking hot strip mill as the study object,it is observed that the rolling equipments are independently installed but the rolling stages are tightly cascaded.Then,the industrial cascaded FieldNets are proposed,where each FieldNet is a field sub-net corresponding to one process stage.Within the framework of the industrial cascaded FieldNets,this dissertation aims at the real-time data collection for process industries.First,the end-to-end delay optimization for data collection is considered.Based on the topology of cascaded FieldNets,a nonlinear mixed integer programming problem formulated by both(i)channel allocation among FieldNets and(ii)multichannel transmission scheduling within each FieldNet.To solve it,a two-level resource allocation method is proposed,by which we prove that sub-problem(i)can be solved independently from the sub-problem(ii).Then,the optimal solution of the original problem is obtained.Compared with standardized IWSN protocols,this method achieves lower end-to-end delay.Second,the delay of heterogeneous data collection is considered.In IWSNs,it is still a challenging problem due to the complexity of multichannel access collision avoidance and transmission priorities coordination.To address it,we propose a separate design principle(SDP)for priority-aware transmission scheduling.In SDP,the multichannel superframes are separated designed for sensors with same priority based on IEEE 802.15.4e protocol,where each superframe fully utilizes the slotted channels.As a result,each time slot on each channel is repeatedly scheduled to sensors of different priorities.Then,a priority-aware transmission coordination mechanism is devised to rule the sensors transmitting in predefined priority order.Both theoretical analysis and experimental results demonstrate that SDP obtains lower transmission delay of each priority sensor than the TDMA-based scheduling.Third,the fast spectrum access problem is considered.It is observed that the evenness of spectrum usage affects channel accessing delay.To this end,a new concept of equilibrium is defined to represent the achievable best evenness of spectrum usage,and a set of rules called Local EQuilibrium guided Autonomous Channel Switching(LEQAutoCS)are then devised.With LEQ-AutoCS,each accessed sensor autonomously equalizes the local channel occupations within its range of spectrum sensing without exchanging the spectrum sensing reports.It is further proved that the equilibrium can be achieved by this concessive manner.Theoretical analysis and experimental results show that LEQ-AutoCS rules gains lower spectrum access delay than some existing spectrum access strategies.This dissertation has presented the design of industrial cascaded FieldNets and the optimization for the real-time data collection,which provide a positive reference for the studies on IWSNs for multi-stage industrial processes.