Submitted: March 2021

Abstract

Bluetooth low energy (BLE) is a low power wireless technology well established in consumer products. BLE is also used for industrial applications. While BLE is designed to transfer data reliably, the worst-case transmission delay is unbounded. This is problematic for real-time systems which require predictable transmission delays. In this thesis, we investigate probabilistic time-bounds for BLE employing a model-based approach. As a use-case, we explore a wireless biking application that requires a predictable transmission delay below 50 ms. We measure BLE transmission errors and latency within a residential area. These measurements help us to understand BLE performance and are the basis for our modeling. We build a Modest model for BLE which conforms with the experimental measurements. We use this model to verify that BLE is reliably able to transfer data within the 50 ms deadline. We also modify the model to explore the novel Isochronous channels which were introduced only recently to the Bluetooth specification. The model can show significant latency improvements if using Isochronous channels compared to regular BLE.