Computing Electronics Market to Further Develop with Concept of Near-Sensor and In-Sensor Technology
- Analysis
- 13-January-2021
The presence of sensory nodes on the Internet of Things is swiftly increasing. As per research, this number may reach up to 45 trillion by the year 2032. The information that is generated from sensory nodes is equal to 1020 bit/second. Hence, the need for the hour is that computation tasks should shift from cloud computing centers to edge devices, at least partly. This will help in reducing energy consumption and time delay and, in turn, improving data security and privacy while saving communication bandwidth.
Devices that connect to the internet keep on increasing day by day—in turn, increasing the quantity of data that is transferred between computing units and several sensory terminals. Computing approaches can help decrease power consumption and, at the same time, process redundant data more efficiently by interceding in the vicinity of or inside sensory networks.
Recently, researchers have brought forward a study that outlines the concept of in-sensor and near-sensor computing. These two approaches facilitate the partial transfer of computation tasks to sensory terminals. It my even lead to a decrease in power consumption and an increase in the performance of algorithms. This is a massive development in Computing Electronics Market as it may help bring this in-/near sensor computing architecture into reality. These concepts may also be able to integrate themselves with other types of sensors, such as those that can detect biological signals, pressure, acoustic chemical, or even stain.
Sensors and Computing units have different roles and so are usually made up of different materials. They have different device structures, processing, and design systems. Conventionally, sensors, and computing units are kept separate from each other in the sensory computing architecture. However, the team has eliminated or significantly lessened this difference in near and in-sensor computing architectures in the new concept.
In near-sensor computing systems, accelerators or processing units are placed right beside sensors. This results in processing units or accelerators executing specific operations at the sensor endpoint. Increasing the overall functioning of the system and minimizing the transfer of unneeded data.
Whereas, in the case of in-sensor computing architectures information collected by the individual sensors or multiple connected sensors is directly processed. This removes the requirement for having an accelerator or processing unit as it combines functions performed by sensors and computing systems.
Both concepts revolving around computing are interdisciplinary in nature and cover architectures, materials, integration technologies, algorithms devices, and circuits. These architectures have an intricate design because they come across a huge amount of data and several types of signals, that too in different scenarios. In order to successfully install near-/in-sensor computing, the co-development and co-optimization of algorithms, devices, sensors, and integration technologies are required.
The researchers have suggested possible ways in which the realization of integrated sensing and processing units. This would help provide a dependable definition of near and in-sensor computing.
In the future, this study could inspire further research works aimed at realizing these architectures or their hardware components using advanced manufacturing technologies. Until now, the research has been focused on vision sensors but may come to be integrated with other types of detecting sensors as well.
Other Related Reports:
Global Entertainment Consumer Electronics Market 2020 by Manufacturers, Regions, Type and Application, Forecast to 2025
Global Visual Computing Market 2020 by Company, Regions, Type and Application, Forecast to 2025
Global Hyperscale Computing Market Growth (Status and Outlook) 2020-2025