SEMIoTIC project main goal is investigating and providing integrated solutions to distributed sensing networks and the Internet of Things, which contribute to minimize environmental, energetic and material costs.

In order to achieve this objective, we plan to evaluate, design and experimentally test multiple integrated microelectromechanical sensors and the associated electronics using standard CMOS technology to implement monolithic or quasi-monolithic systems for environment sensing.

When power consumption of such systems is reduced to sub-milliwatt levels under active operation, it is possible to envisage self-powered, autonomous devices, also called motes, that harvest energy from the ambient and are able to provide simple sensing operations and transmit this information to a wireless network. Thus, it is intended to demonstrate the possibility of using sensor systems in some applications without any maintenance and being able to run indefinitely except for aging.

The project is structured in three levels:

A. Development of VLSI transduction and energy harvesting devices.

B. Sensor conditioning, ultra low-power and energy harvesting circuits for autonomous sensors.

C. Autonomous Systems.

Building on the experience of the research group, various MEMS (Micro-Electro-Mechanical Systems) sensors will be developed. The distinguishing fact is that they will be implemented in standard CMOS technology (mainly membranes) using the BEOL (Back End of Line) process metallization layers and applying a postprocessing etching in order to release the metal layers.

Currently, a 3D accelerometer is available and it is proposed to develop at least two new MEMS sensors: a pressure sensor and a magnetometer based on Lorentz force. In addition, CMOS photoelectric sensors, for energy harvesting, thermal sensor, for temperature compensation, and magnetic sensors using giant magnetic resistance materials or GMR (Giant Magnetic Resistance) will be also developed.

MEMS Brownian noise and low signal level demand precise circuits to maximize sensor performance. In particular, pressure sensor and Lorentz force magnetometer conditioning will be based on closed-loop oscillation of mechanical elements. Thus, this requires a chain of signal processing circuits, analog and digital.

In many distributed sensor applications, the system must be able to work without any external connection. Thus, the required energy will be kept to the minimum necessary and collected from the environment through energy harvesting techniques. The power management stages must be optimized to handle very low power and voltage.

The integration feasibility of environmental energy harvesting systems together with power management circuits will be considered. Of course, this will constrain the duty cycle of sensing operation and tradeoffs on the sensing performance will also arise.

Depending on their power balance, a subset of the transducers, circuits and devices developed above will be combined together with control circuitry, power management and energy harvesting. The ultimate purpose is to develop a demonstration sensing prototype mote that:

• Allows obtaining enhanced information from the synergy of multiple sensors
• Meets the requirements of integration and energy. When not possible to integrate all the elements, the size of the external ones will be minimized and connected to the multisensor chip in multi-chip modules.