Telephone operators hold that their technology and infrastructures can support about 10-15% of the devices that the IoT paradigm plans to deploy by means of their current GPRS/3G/4G solutions. WiFi technology can span an additional significant percentage, but it’s clear that in unpowered environments – not only in Smart Cities, but also infrastructures and agriculture – both options are not the solution.
IoT has big challenges ahead. Gus Vos, Chief Engineer at Sierra Wireless, summarizes it very well in the so-called 3 C’s – Cost, Current, and Coverage – and personally I would add a fourth C for Cybersecurity.
Regarding the energy consumption, optimizing it in unpowered systems is kay for the feasibility of wireless networks. A bad use of the technology or the devices will cause, in the short or half term, in unaffordable costs for its maintenance. In a system with hundreds or thousands of battery-powered dispersed points, the change can be time-consuming and expensive. That’s why, regardless of the chosen technologies, the equipment in our IoT network must accomplish the following set of policies in order to increase the batteries lifespan whenever possible:
– OPERATING MODE: the normal mode of a node must be the idle mode, i.e., almost switched off. It must be always on exclusively when the system enables it to wake up in response to a stimulus. Once this has occurred, the node switches on completely and performs the programmed functions, typically acquiring, processing and sending information.
– STRATEGY: the use of PUSH is prioritized from PULL, so the node ensures the transmission of useful information. This avoids the node to be activated periodically if there aren’t new data.
– WRITING: the drawback with a PUSH strategy is that it doesn’t enable parameters to be written. If necessary, a procedure is applied so the node periodically – once a day, a week, a month… – requests its gateway whether it wants to overwrite. If a response is not received immediately, the communication is closed and goes back to idle mode. If the gateway has to change a parameter, it immediately sends a response back with the message. One received by the node, it processes it, makes changes and goes back to idle mode.
– ARCHITECTURE: it must be ensured that, for each transmission of information, a minimum number of nodes must be turned on. This implies that, beyond mesh or tree architectures, the star architecture must be prioritized since the node sends all information directly to the gateway.
– COMMUNICATION: transmission times must be minimized, since they require the node to be on and generating a signal, which requires high energy consumption. For this reason, the number of messages must be minimized – always ensuring its transmission and reception – at the highest possible speed. This means that it’s better to send one big message than three smaller messages, and always with the highest speed enabled by the channel.
– CODE: the node programming code must be optimized, minimizing code lines to be run since typically each processor runs each command line at a periodic speed between 1 and 1000 µs. For this reason, any instruction set that may be reduced will imply a shorter node enabling time. Obviously, all instructions that leave the device in standby – a WHILE structure, for example – are highly unadvisable. If used they must have always a loop output for timeouts.
– POSSIBLE POWER SUPPLY: depending on the type of collected and measured signal, it’s possible to extract energy from it in order to power the node. This involves additional layers at the signal acquisition and rectification node, but it provides a big advantage to achieve a significant amount of energy. Two conditions must be met for it to be feasible. The first is that the acquired signal should allow it, i.e., it’s feasible if temperatures or pressures are measured, but it’s more difficult for a humidity or pollution signal. The second condition is that the acquisition should not affect the measurement. The veracity of information is prioritized over the node supply.
– WARNINGS: the node must transmit the battery level on each message or when this level falls below a certain threshold. It’s not an energy savings procedure, but it can help to take decisions for parameter writing – reduce the information transmission frequency – or to plan preventive maintenance tasks.
Surely any network of devices not complying with most of these features will pose problems. However, it’s so relevant for the devices to be focused on energy savings as the right technology used to transmit the information. Remember that the information transmitted by means of 3G/4G consumes 2,500 as much energy as if it were based on IEEE 802.15.4.