Research On Algorithm Of Wireless Channel Optimization In Coal Mine Internet Of Things

With wide utilization of Internet of Things(Io T) mainly supported by wireless communication technology in coal mine, variety of safety monitor&control data through wireless transmission has shown a trend of exponential growth. The electromagnetic interference on wireless communication channel is the biggest constraint of broadband wireless transmission throughput(channel capacity), and the impact of channel interference with multi-path effect on wireless systems is much more serious especially in coal mine closed wireless environment. Strong co-channel & adjacent channel interference will seriously affect the quality of communication and substantially increase loss of monitor&control data and transmission time latency, which is a direct hazard to mine safety. Therefore, research on channel allocation minimizing co-channel & adjacent channel interference and intelligent channel switch algorithm avoiding severe electromagnetic interference has very important, profound and lasting significance.First, this thesis studies self-organized distributed wireless channel allocation of coal mine Internet of Things(Io T) under the condition of no time-varying channel states.Based on in-depth analysis of existing distributed wireless channel allocation algorithms, this thesis proposes an innovative fast distributed orthogonal Channel Allocation scheme based on Largest Edges of Network Topolgy(CALENT) with the combination of sensor nodes’ distribution and architecture features, and determines minimum transmitter power of nodes and detailed procedure of CALENT. By comparing with distributed channel allocation Dis-Link algorithm, CALENT has better performance in co-channel interference index and convergence velocity.Based on in-depth analysis of existing physical and protocol interference models, this thesis proposes an innovative Sender-Receiver based Protocol Interference Model(SRPIM) and distributed orthogonal wireless channel allocation algorithm CALENT-BE with bidirectional communications of peer nodes. Due to asynchronous channel allocation mechanism, CALENT-BE overcomes the slow convergence of original algorithm CALENT in data back mode.To improve spatial reuse efficiency of Partially Overlapped Channels(POC), this thesis proposes an innovative distributed wireless channel allocation algorithm CALENT-POC under non-orthogonal wireless channel sets of Wi-Fi communications. Through effective avoidance of excessive channel allocation proliferation by channel allocation control mechanism, CALENT-POC ensures the diversity of channel allocation strategies and provides global optimal solution of minimizing total number of channels needed while acheiving no channel interference in the whole network.Next, this thesis studies intelligent wireless channel switch algorithm of Internet of Things(Io T) terminals avoiding severe electromagnetic interference in coal mine specific areas of time-varying wireless environment.Based on in-depth analysis of traditional single fixed channel communication and Q-learning channel switch, this thesis proposes an innovative Experience-Weighted Attraction(EWA) wireless channel switch algorithm. It establishes more detailed and reasonable multi-layer reward feedback organization and can learn communication channel characteristics of radio environment online. By accumulating history of channel switch experience, EWA is able to predict, select and change current optimal communication channel, dynamic ensuring the quality of communication links while reducing system communication outage probability.In the premise of maximally guaranting optimal channel selection accuracy, this thesis also proposes an innovative fast wireless channel switch algorithm EWAS to solve the shortcomings of original EWA algorithm’s complexity, higher energy consuming, and nodes’ hardware restrictions of real-time data processing capabilities. To avoid existing exponential operation, EWAS simplifies the calculation procedure by updating the objective function directly.In the end, this thesis summarizes the whole research work while analyzing problems and shortcomings of current research, then outlooks future research directions and prospects.