Lecture
Intelligent sensors guided by biology
- at -
- ICM Saal 4b
- Type: Lecture
Lecture description
R. Brederlow, E. Korek, M. Ochs,
There are lots of sensor analytical problems where nature and biology have inspired optimal technical solutions. This talk will discuss biologically inspired ways to design microelectronic analytical sensor systems to identify information useful to mankind. After an introduction into a practical sensing problem, basics of sensor needs and their data processing, as well as a brief explanation of biologically inspired neural networks used to process data, I will show practical examples for both sensor and circuit technologies enabling such function and their relation to classical physical sensors. To have impact on society and markets such electronic sensor systems however are not good enough yet: sensor systems need to be versatile in their information content, power efficient and cheap to be placed everywhere where sensor data are of interest. Their data need evaluation capabilities in the application context have to be meaningful. Therefore, understanding and learning in the application context is a mandatory step for the usefulness. Using similar data evaluation as used for sensor postprocessing also in system level context we may furthermore address application related problems - closing the loop towards well known artificial intelligence algorithms.
There are lots of sensor analytical problems where nature and biology have inspired optimal technical solutions. This talk will discuss biologically inspired ways to design microelectronic analytical sensor systems to identify information useful to mankind. After an introduction into a practical sensing problem, basics of sensor needs and their data processing, as well as a brief explanation of biologically inspired neural networks used to process data, I will show practical examples for both sensor and circuit technologies enabling such function and their relation to classical physical sensors. To have impact on society and markets such electronic sensor systems however are not good enough yet: sensor systems need to be versatile in their information content, power efficient and cheap to be placed everywhere where sensor data are of interest. Their data need evaluation capabilities in the application context have to be meaningful. Therefore, understanding and learning in the application context is a mandatory step for the usefulness. Using similar data evaluation as used for sensor postprocessing also in system level context we may furthermore address application related problems - closing the loop towards well known artificial intelligence algorithms.
