Representatives from the CSIC team headed to Berlin in April for the International conference and exhibition covering Wireless Sensor Networks (WSN), Real Time Locating Systems (RTLS) and other forms of Active RFID: WSN and RTLS Conference, Berlin. While there, the team also presented at the British Embassy as part of the British delegation trade mission.
One of CSIC's highly talented researchers, Heba Bevan, reports:
Over the past decade the demand for wireless communication devices in various applications has skyrocketed as increased functionality and affordability have dovetailed to make these devices a ubiquitous feature of our society. Wireless sensors are often used to monitor quantitative physical changes. Readings are translated into a signal and relayed through communication points, or nodes, until it reaches a base. This is a reliable method for monitoring and transmitting change data.
High performance, small size and low power consumption are qualities that have yet to be simultaneously integrated in a WSN device. Successfully creating such a device is a lengthy process. Lab testing, while necessary, is insufficient to fully replicate real-life conditions and environments.
There are three major research challenges for the successful deployment of wireless sensor networks.
Hardware: picking a CPU that gives high performance, long-term usability and low power consumption to manage the device; and choosing the right combination of resistors, transistors, capacitors and sensors that gives a balanced level of high performance and low power consumption.
One of the most important aspects of designing an optimal sensor network device is selecting the right CPU. For decades the 8-bit CPU has been a staple of sensor networks such as Atmel 1281. Simple to program and easy to use, the 8-bit processor’s low power consumption made it a natural choice for system architects. In comparison, traditionally 16-bit and 32-bit processors consumed a great deal more power, requiring frequent manual battery changes - a difficult task when devices were located in difficult to reach places. 16-bit and 32-bit processors also used to cost a great deal more than 8-bit processors. Today, however, the 8-bit CPU is no longer the clear-cut choice for sensor network devices. Advances in power management have made 32-bit processors viable and indeed vital components in advanced wireless system networks, such as STM32W, Atmel SAM3U and MeshConnect ZICM357SP based on ARM Cortex CPU.
Software: choosing an OS with a small memory allocation footprint on the CPU, low power consumption and open source libraries providing long-term usability that enables the user to further customise the device without licensing for optimal application performance example for OS used in wireless sensor network TinyOS, Nano-RK, MANTIS, Lite OS and Contiki.
Networking: using a communication protocol that is universally accessible.