Machine Learning Based Data Rate Optimization for Mobile LoRaWAN Sensors
Adaptive Data Rate is a feature of LoRaWAN that allows to optimize the network performance by adjusting the data rate of end devices based on their current channel conditions. Current approaches to ADR optimization algorithms focus on static or low-mobile end devices, and the specification recommends to disable ADR for mobile devices, e.g. location trackers. To let these devices also benefit from ADR adjustments, this thesis suggests to implement a predictive ADR algorithm based on deep reinforcement learning. The algorithm is evaluated on a real-world data set captured in the city of Darmstadt.
Generalized Network Coded Cooperation in High Density LoRa-Networks
The wireless channel is a non-linear and time-varying system. Thus, it represents a harsh environment to conduct transfers of information. One of the variables that predict the outage performance of transmissions over the wireless channel is the diversity order, where systems with higher diversity order experience a lower outage probability at a specific signal-to-noise ratio (SNR).
Diversity can be achieved through several means, with the most simple being repetitions (retransmissions) of the same information over different instances of the wireless medium, i.e. over another time period or frequency. One very relevant means is the use of multiple antennas, which adds diversity by also incorporating a spatial element. However, this element can also be obtained when devices with transmissions to a common destination aid each other with retransmitting their partner's information frames. That is the concept behind cooperative communication: achieving a spatial diversity gain without requiring multiple antennas on each device .
Network Coded Cooperation (NCC) is a more complex cooperative technique whereby devices perform linear combinations of the data contained in their own and their partner's information frame, creating a parity frame. This allows for an even higher diversity order gain without requiring any additional transmissions beyond the standard information and cooperative phases seen in cooperative communications .
This kind of technique can therefore be especially useful in scenarios where multiple devices share a common base station and require energy-efficient communications, such as in LoRa-based networks. LoRa is a prime modulation technique for enabling Low-Power Wide Area Networks (LPWANs), providing adequate interference prevention, relatively low power consumption, and long range. These benefits, however, do not scale well with increases in the network density . Note that, in these high-density scenarios, increasing diversity by simply realizing more transmissions results in an increased collision probability, i.e. even higher interference. For LoRa-networks, this also means the network loses maximum range. Given that the number of connected devices is expected to balloon in this decade, LoRa-based protocols must be adapted to mitigate high levels of interference.
It has been shown that using NCC can produce positive results in the high-density scenario LoRa-based network when associated with a fast inter-device transmission of information frames using high rate frequency shift-keying (FSK). However, previous analyses were purely theoretical and limited to evaluating a two-way cooperation process .
This thesis will tackle the empirical and theoretical challenges of implementing generalized network-coded cooperation on LoRa-based networks. Cooperation will be expanded to include multiple devices within the cooperation range, which will generate a higher diversity order for the uplink transmissions. The student is expected to be programing LoRa devices based on either the SX1272 or SX1276 transceivers to validate their results.
If you have any interest in the described topic, please do not hesitate to get in touch.
 Cooperative communication in wireless networks
 Multiuser Cooperative Diversity Through Network Coding Based on Classical Coding Theory
 Network-Coded Cooperative LoRa Network with D2D Communication