Yi Yu

Cette thèse étudie le problème de planification des ressources pour les réseaux IoT longues portées basés sur les technologies NB-IoT et LoRa. Dans les deux cas, on suppose que les capteurs et les collecteurs sont distribués suivant des processus de points de Poisson spatial indépendants marqués par le caractère aléatoire du canal. Pour le NB-IoT, nous élaborons un modèle de dimensionnement statistique qui estime le nombre de ressources radio nécessaires en fonction du délai d’accès toléré, de la densité des nœuds actifs, des collecteurs et de la configuration de l’antenne. Pour le réseau LoRa, nous proposons une technique d’allocation de plusieurs sous-bandes pour atténuer le niveau élevé d’interférence induit par les nœuds qui transmettent avec le même facteur d’étalement. Pour allouer dynamiquement le facteur d’étalement et la puissance, nous présentons une approche d’apprentissage automatique avec multi-agents qui permet d’améliorer l’efficacité énergétique.

In this paper, we focus on the radio resource planning in the uplink of licensed Orthogonal Frequency Division Multiple Access (OFDMA) based Internet of Things (IoT) networks. The average behavior of the network is considered by assuming that active sensors and collectors are distributed according to independent random Poisson Point Process (PPP) marked by channel randomness. Our objective is to statistically determine the optimal total number of Radio Resources (RRs) required for a typical cell. On one hand, the allocated bandwidth should be sufficiently large to support the traffic of the devices and to guarantee a low access delay. On the other hand, the over-dimensioning is costly from an operator point of view and induces spectrum wastage. For this sake, we propose statistical tools derived from stochastic geometry to evaluate, adjust and adapt the allocated bandwidth according to the network parameters, namely the required Quality of Service (QoS) in terms of rate and access delay, the density of the active sensors, the collector intensities, the antenna configurations and the transmission modes. The optimal total number of RRs required for a typical cell is then calculated by jointly considering the constraints of low access delay, limited power per RR, target data rate and network outage probability. Different types of networks are considered including Single Input Single Output (SISO) systems, Single Input Multiple Output (SIMO) systems using antenna selection or Maximum Ratio Combiner (MRC), and Multiuser Multiple Input Multiple Output (MU-MIMO) systems using Zero-Forcing decoder.

In this paper, we consider an IoT dedicated network corresponding to a non licensed LoRa Low Power Wide Area Network. The LoRa network operates in the unlicensed 868 MHz band within a total bandwidth of 1 MHz divided into 8 orthogonal channels of 125 kHz each. Despite the high level of interference, this network offers long range communications in the order of 2 to 5 km in urban areas and 10 to 30 km in rural areas. To efficiently mitigate this high level of interference, LoRa network essentially relies on a Chirp Spread Spectrum (CSS) modulation and on repetition diversity mechanisms. The LoRa CSS modulation spreads the signal within a band of 125 kHz using 6 possible spreading factors (from 7 to 12) to target data rates (starting from 5 kbps for the closest node to 300 bps for the furthest ones). The repetition diversity mechanisms enable the data recovery when the transmission is subject to bad channel conditions or/and high interference levels. Although the CSS modulation protects edge-cell’s devices from the high level of interference induced by nodes in the proximity of the gateway, it fails to protect nodes at the edge of a given SF region and several trials are required to recover the packet. In this paper, we propose an adaptive multi-channels allocation policy that attributes multiple adjacent channels of 125 kHz for nodes situated at the edge of SF zones. We study the impact of this adaptive sub-band allocation on the gateways’ intensities, the rate distribution and the power consumption. Our results are based on a statistical characterization of the interference in the network as well as the outage probability in a typical cell.