Intelligence Management of BLE Sensors by the Edge Device

Reducing energy consumption is an important task to this day. Thanks to the development of energy efficiency technologies it was possible to achieve the possibility of using a large number of personal devices. One such technology that has led to the emergence of personal wireless devices (smart watches, headphones, etc.) is BLE (Bluetooth Low Energy) technology. This technology has significantly reduced the use of electricity by the devices with which the devices were developed. Such devices using BLE technology have become widely used in industrial solutions as iBeacons devices. Today, the topic of reducing the energy consumption of these devices is still relevant due to aggressive environments of data use of devices that reduce the life of these devices. The paper presents an algorithm for controlling BLE sensors using an edge device using a deep neural network. An algorithm based on a set of parameters (distance between BLE and Edge devices, priority, and number of current connections) that describe the BLE sensor, calculates the optimal duration of deep sleep of the microcontroller to reduce average power consumption of the device. The main element in this algorithm is a deep neural network, which returns the individual value of deep sleep for the BLE sensor. The new value is written to the corresponding BLE GATT service and characteristic. The simulation allowed to use the BLE sensor battery capacity by almost 20% more efficiently.