Improving the Timeliness of Bluetooth Low Energy in Dynamic RF Environments

The ability to communicate within given delay bounds in noisy RF environments is crucial for Bluetooth Low Energy (BLE) applications used in safety-critical application domains, such as health care and smart cities. In this work, we experimentally study the latency of BLE communications in the presence of radio interference and show that applications may incur long and unpredictable transmission delays. To mitigate this problem, we devise a model capturing the timeliness of connection-based BLE communications in noisy RF channels by expressing the impact of radio interference in terms of the number of connection events necessary to complete a successful data transmission (n_CE). We show that this quantity can be estimated using the timing information of commands sent over the host controller interface of common BLE devices, hence without additional communication overhead or energy expenditure. We further show that a BLE device can make use of our BLE timeliness model and recent n_CE measurements to adapt its BLE communication parameters at runtime, thereby improving its performance in the presence of dynamic radio interference. We implement such an adaptive scheme on the popular nRF52840 platform and perform an extensive experimental study in multiple indoor environments using three different BLE platforms. Our results show that a BLE application can, indeed, make use of the proposed model and recent n_CE measurements to adapt its connection interval at runtime to increase the timeliness of its communications, reducing the number of delayed packets in noisy RF environments by up to a factor of 40.