Adaptive Multi-Channel Offset Assignment for Reliable IEEE 802.15.4 TSCH Networks

More and more IoT applications require low-power operations and high reliability (close to 100%). Unfortunately, radio transmissions are unreliable by nature since they are prone to collision and external interference. The IEEE 802.15.4–2015 TSCH standard has been recently proposed to provide high reliability through radio channel hopping and by appropriately scheduling all the transmissions. Since some of the radio channels still suffer from external interference, blacklisting techniques consist in detecting bad radio channels, and in privileging the good ones to transmit the packets. MABO-TSCH is a centralized scheduling algorithm which allocates several channel offsets to allow each radio link to apply a localized blacklist. However, such strategy is inefficient for large blacklists. In this study, we propose to allocate the channel offsets dynamically at each timeslot according to the number of parallel transmissions, while still avoiding collisions. We evaluate the performance of our solution relying on a real experimental dataset, highlighting the relevance of dynamic and per timeslot channel offset assignment for environments with high external interference, such as a smart building.