Green ICT

Ongoing digitalization is making a significant contribution to the further development of the economy and the development of innovative products and business models. However, with digitalization, the total energy consumption of information and communication technology (ICT) has risen continuously in recent years.

Green IT/ICT refers to efforts to use information and communication technology and the entire ICT system in an environmentally friendly and resource-saving manner throughout its entire life cycle. This includes optimizing the consumption of resources during the manufacture and operation of devices.

Below we present some of our activities in the area of Green ICT.

Energy self-sufficient, maintenance-free and wireless systems with unlimited operating time - that is the aim of our energy harvesting technologies. Available energy sources from the environment such as light, heat or movement serve as energy suppliers for wireless sensors, displays and other small consumers.

In the field of energy harvesting, we develop and investigate technologies and systems for using energy from the environment to supply small electronic consumers.

The use of ambient energy eliminates the need to use, change and dispose of batteries.

In addition, we are investigating whether the use of an energy harvester is more advantageous from an ecological point of view than the use of batteries.

Further Information

Massive IoT applications in the smart city sector and IIoT require permanent energy for reliable and robust data transmission. Our energy self-sufficient wide-area network is a unique combination of the new LPWAN standard mioty® with energy harvesting. Telegram splitting makes mioty® technology more energy-efficient. In addition, fewer base stations are required to use the technology compared to other LPWAN solutions.

This not only reduces costs, but above all ensures that fewer base stations need to be manufactured, operated and disposed of. In combination with energy harvesting, mioty® technology therefore increases the service life of transmission systems.

Further Information

We develop and offer technologies that generate electricity from slight temperature differences or barely perceptible vibrations in order to operate sensors or small devices in the IoT in an energy self-sufficient manner - for example, the intelligent screw for monitoring screw connections. In particular, we are looking at the power management function, which ensures energy-efficient coupling of energy harvesting and the IoT sensor node. Power management differs depending on the energy source.

An important criterion for applications in the IoT sector is the lowest possible power consumption. Continuous wireless networking requires a lot of available battery capacity. The operating time of wireless sensor nodes is therefore limited. The RFicient® portfolio enables low-maintenance ultra-low-power connectivity for many years.

RFicient® ultra-low-power receiver technology enables continuous monitoring of a wireless channel with microwatt-level power consumption and responds in milliseconds. As a WakeUp variant, the RFicient® technology has a very low power consumption of < 3 µA. This can be achieved with the RFicient® chip from the Fraunhofer Institute for Integrated Circuits IIS, which can save 99 percent of power consumption. This means that mobile applications can be operated with 24/7 connectivity and an extended service life of up to 10 years.

Further Information  

Machine learning applications in embedded devices are an emerging trend. A large number of AI chips have been announced and the first products for embedded AI are on the market. Current neural network architectures such as deep neural networks require high computational complexity and high power consumption. Neuromorphic hardware, on the other hand, relies on massive parallel processing and performs calculations, e.g. for machine learning, faster and with less power.

Further Information

Edge AI is a research area in the field of machine learning and describes the optimization and execution of AI-based processing chains on embedded systems. ML applications are executed that require up to a thousand times less power compared to a standard GPU. One of the reasons for this is the local execution of energy-efficient AI models optimized for embedded systems. In addition, there is no need to transfer the data to a cloud server. For this reason, such embedded AI-capable devices can be operated for years without batteries, depending on the application.

Further Information

Edge computing is a process in which data, services and application information are moved directly to the logical "edge" of a network. The time-consuming route to and from the cloud, which often forms the "bottleneck" for fast and effective communication, is eliminated.

With edge computing, only data that is actually needed in the cloud to optimize processes is transferred. Location-based data processing through edge computing reduces the amount of data transfer over long distances, which significantly reduces energy consumption. This architecture also makes it easier to comply with security requirements. System availability, short latency times, data backup and encryption are easier to implement.

Further Information

The concept of distributed data processing ensures that resources are used optimally in every application throughout the entire IoT system, from sensor to edge to cloud. To find this out, different processing focuses are compared with regard to the life cycle assessment in the IoT chain.

By using software containers in computing clusters, computing resources are optimally utilized. For example, individual computing units can be switched off at times while others take over this computing power because they have free capacity.

This saves a lot of energy, especially when operating large computing clusters.