Cognitive Radio is the Future

In the future internet of things, wireless technologies will represent a large part of the market in weak competition with power line or infrared communications. Up to now, standards have been developed with a bottom-up approach and the radio spectrum sharing policies are mostly static. Since billions of objects are expected to use wireless links, the present way the wireless medium is shared has to be revisited. For such, radio systems, algorithms and protocols have to be deeply transformed. The most promising approach stems from the cognitive radio paradigm, which relies on three complementary mandatory properties: radio systems real-time reconfigurability, wireless environment awareness behavior, and self-organization capability. The first item is referred to as software defined radio (SDR), the second as cognitive radio (CR) and the third as self-optimization networks (SON).


A valuable platform able to test realistic future scenarios, should offer the possibility to evaluate simultaneously these three items to support the actual theoretical developments from an experimental point of view. Unfortunately, a platform that contemplates all of these topics at the same time does not currently exist. CorteXlab aims to fill in this gap. The CorteXlab testbed will be hosted at INSA-Lyon in France, benefiting from the Senslab (now FIT/IoT-lab) experience and from the 5 years experience on developing a MIMO (Multiple Input Multiple Output) high data rate reconfigurable platform. CorteXlab will use the network architecture developed in IoT-lab and will integrate SDR nodes to offer a remotely accessible development platform for distributed Cognitive Radio (CR). Reconfigurability, compatibility, coexistence and even cooperation between SDR nodes will be evaluable. A large set of heterogeneous SDR nodes (MIMO nodes, SISO nodes and Wireless Sensor Network (WSN) nodes) together with classical sensor nodes will permit a full experimental evaluation.

Cognitive Radio at the Reach of everyone

CorteXlab will allow remote users to test their own algorithms on the existing nodes, but the architecture will be also opened to industry third party to deploy their own front-end (RF or UWB) or baseband systems to test and validate their developments. A clear expected result is offering a remote access to all equipments in a comprehensive way, such that many scenarios can be evaluated by remote users. A second target is to create a “network of SDR” development community including people from digital communications, networking and embedded systems, to provide a complete set of functionalities and also enroll interesting industrial partners to include in the platform new SDR or front end components.


Tanguy Risset

Sense City

Sense-City is an EquipEx platform lead by Université Paris-Est. It provides a set of equipment for prototyping and evaluating the performances of micro and nanosensor systems for sustainable cities. Sense-City is a « climatic mini-city », a 400m² large, mobile and reconfigurable hall, able to host realistic urban experimentations in controlled environmental and climatic environment. The main urban features are available such as buildings, infrastructures, urban furniture, distribution networks, and soils. The Urbanet team is involved in several experiments, including « smart-road » sensor networks for low-cost vehicle detection, and atmospheric pollution sensor deployments and self-calibration.


Hervé Rivano

NVRAM for Transiently-Powered Systems

Non Volatile Random-Access Memory (NVRAM) is an umbrella term covering several upcoming memory technologies like with groundbreaking properties. These include ferroelectric memory (FRAM), magnetoresistive memory (MRAM), phase-change memory (PCM), and the memristor. Non-volatility means that data stays intact when power is lost. Random access means that each byte can be read or written directly by the CPU.
These two properties together make the classical RAM+Flash architecture obsolete, and enable the design of tiny embedded systems running on intermittent power. This is very attractive in the context of energy-constrained scenarios, for instance systems harvesting their power from the environment. But working with NVRAM also poses novel challenges in terms of software programming. For instance, application state consistency must be guaranteed accross reboots, even though the system includes both NVRAM and volatile elements (e.g. CPU, hardware periperals).
In this context, we are developing and studying novel operating system mechanisms for NVRAM-based embedded systems.


Guillaume Salagnac

Kevin Marquet

Sense In The City

Sense in the city is a lightweight experimentation platform for wireless sensor networks in development. The main objective of this platform is to be easily transferable and deployable on the field. It allows a simplified deployment of the code running on the sensors and the collection of logs generated by the instrumentation of the code on a centralized database. In the early stage of the platform, the sensors are powered by small PCs, e.g. Raspberry Pis, but we are investigating the integration of energy harvesting capabilities such as solar panels.


Khaled Boussetta

Robots + IoT (ROBIOT)

Service robotics is an application domain currently emerging rapidly. We are involved in developing observation and surveillance systems, by using ground robots (Turtlebot2 robots) or aerial ones (Parrot Bibop 1 drones).


Olivier Simonin

Radio communication systems experimentation facility

Radio communication systems experimentation facility

The main objective of the CITI’s radio communication systems experimentation facility (the “radio” room) is to provide to the researchers working on antenna and analog radio front ends domains, the opportunity validate their theoretical concepts through experimental setups. This facility features all the hardware and software tools needed in order to conceive, model and test the radio interfaces for the next generations of wireless communications systems. Here, the influence of the radio interface parameters such as noise, sensitivity, nonlinearity, etc. on the transmission quality is quantified thanks to a versatile platform combining hardware and software capabilities and being able perform experiments in RF, IF or baseband domain. Each year, the platform hosts a dozen of graduate students from INSA Electrical Engineering department, PhD and post doc from CITI laboratory.

Facility description

The experimentation facility has two arbitrary waveform generators (Keysight’s ESG4438C and Rohde & Schwarz’s SMBV 100A), able to generate complex waveforms and a two path vector signal analyzer having 36 MHz bandwidth and able to process signals up to 6GHz. By using this equipment, MIMO 2×2 experiments can be conducted. This platform is driven by the Kesight’s ADS software and Error Vector Magnitude (EVM) and Bit Error Rate (BER) measurements can be performed.
The “radio room” is also equipped with a high performance oscilloscope: Keysight’s DSO9064A, 4 analog channels, 10 GSa/s and 600 MHz bandwidth one. Together with this oscilloscope, a current probe is available and power consumption measurements can be performed.
Together with the performance evaluation, the energy consumption is one of the crucial issues of next generation of radio interfaces. The experimentation platform is also equipped with a “homemade” power consumption measurement facility which is built around a National Instruments PXI 5105, 8 channels, 12 bits resolution and 60 MHz bandwidth digitizer. This measurement platform was specially conceived to measure the energy consumption of the communicating objects.
The radio communication systems experimental facility takes great benefit from the new born FIT(Future Internet of Things) CortexLab platform. Indeed, the home made radio front ends conceived and developed in the “radio room” can then be tested in the Cortexlab facility which proposes a complex environment. The two platforms are complementary since Cortexlab is primarily designed have a fixed hardware infrastructure and agile programming capabilities. A MIMO 4×4 National Instruments high performance equipment used jointly by Cortexlab to perform channel model extraction and radio front end characterization. The equipment is built around NI PXIe-5673 VSG and NI PXIe-5644R VST and is able to analyze communications using the most recent communication protocols such as LTE-A.
In addition to that, the” radio room” is equipped with all the tools needed for the analog front end prototyping: a 8 GHz Vector Network Analyzer, soldering station, low frequency generators, DC power supply, high resolution multimeters, etc. Moreover, the Keysight’s ADS together with the VSA software are deployed on a high performance workstation.

Past projects

This experimental facility allowed to characterize the behavior of a double IQ front-end conceived by the researchers of the CITI laboratory in collaboration with the Orange Labs and patented in 2011. This radio front end is able to simultaneously process signals on two communication channels while decreasing the energy consumption by 30% compared to a classical stack radio front end approach.
More recently, in 2015 the same experimental facility allowed building and characterizing a quasi-passive wake-up radio demonstrator which validates the approach proposed by the researchers of CITI laboratory. Our approach consists in using analog processing in order to perform the equipment identification, compared to the classical approach where an ultra-low power energy consumption microcontroller is used. Here the energy consumption and the latencies are drastically reduced.
Novel radio receiver architecture able to reduce the signals dynamic range at the input of the analog to digital converters and so to relax their constraints was proposed by the researchers of CITI laboratory. This receiver was characterized by using the “radio room” facilities and so the influence of a realistic communication channel and of the receiver impairments on the communication performance was measured.
Indoor channel characterization was conducted by using the radio room experimental facility and the model extracted using these measurements were included into the Wiplan simulator. In the same manner body area network channel models were also extracted.


Scientific coordinator: Florin Hutu

Technical issues: Régis Rousseau