Investigation of Bluetooth Communications for Low-Power Embedded Sensor Networks in Agriculture
Andrew D. Balmos, Alexander W. Layton, Aaron C. Ault, James V. Krogmeier, Dennis R. Buckmaster · 2013 Kansas City, Missouri, July 21 - July 24, 2013 · 2013
Abstract. Wireless sensor networks that monitor agricultural equipment can provide useful data to optimize a farm’s productivity. To become widely adopted, such a sensor network requires sensors with battery lives at least as long as a full farming season. The standard low power wireless sensor communication platforms, e.g. Zigbee and ANT, lack common hardware support in consumer devices. Using more ubiquitous technology that is already owned by many farmers, such as cellphones and tablets, can significantly reduce cost and complexity. Nearly all mobile devices currently on the market have Bluetooth built-in, making it a viable wireless protocol choice. In addition, the recent release and initial adoption of the Bluetooth 4.0 Low Energy (BLE) standard is a strong candidate for use in very low power, long battery life sensor networking. This work investigates how various BLE configurations affect network size, network throughput, and sensor battery lifetimes. We present a strategy to select the best network parameters that achieve a certain sensor Update Interval (UI) and Connection Latency (CL), as well as minimize a sensors power requirements, reduce data latency, and conform to size and throughput constraints. We present a simple BLE power model created from current consumption measurements of the commercially available TI CC2540 BLE module. Combining this BLE power model with a simple capacity model of a battery allows estimation and validation of sufficient battery life. Sensor lifetimes on the order of multiple years can easily be achieved for large UI and CL sensors.