Research On Fairness And Interference Resistance In LoRa Networks

LoRa,one of the most potential LPWAN(Low-Power Wide Area Networks)techniques,has been paid much attention recently as it offers radio coverage over a very large area by way of adapting transmission rates,modulation,duty cycles,etc.,such that end-devices incur a very low energy consumption.Its unique nature(long range and low power)makes it an excellent candidate for asset tracking,water meter,fire detection,etc.Since these applications mainly require tens of hundreds of end-devices to collect data,a fair,interference resistant network is badly needed.However,different transmission rates suffer unfairness if they are allocated to equal number of end-devices.The main reason is that lower transmissions seem to be much more prone to collide with each other.Moreover,current interference avoidance scheme could cause collisions again,especially when the the duty cycle of interference is high.In other words,current scheme is not interference resistant.Rising to these problems and challenges,our main research and contributions include:1.In this thesis,we propose Fair Spreading-factor Allocation(FSA),a transmission rate allocation scheme with fair collision probability to ensure the different transmission rate fairness.Based on the fact that in fewer transmissions scenarios,fewer collisions would occur,FSA models the relationship among packet reception ratio,spreading factor and allocation ratio,and allocates fewer transmission end-devices to the spreading factor with high collision probability since spreading factor determines the rate and coverage of LoRa.Moreover,FSA proposes a novel allocation scheme for networks with different coverage and deployment.Later simulations reveal FSA not only achieves fair transmissions among different data rates,but has consistent improvements in packet reception ratio and throughput.2.Based on the observation that interference transmisions also have some idle spaces,we propose an interference-awareness concurrent transmission model LoRaSense,which is interference resistant.This model can identify a particular interference with the knowledge of its fingerprints and choose the proper data length based on the idle period to transmit concurrently.Later experiments show that LoRaSense can achieve 10%-15% packet reception ratio improvements since it can transmit data successfully even with interferences.