Florent Carlier

Multi-user detection for CDMA communications based on self organized neural networks structures

In the upcoming high transmission rate system, based on code division multiple access, standard multi-user detection methods are restricted because of performance and implementation complexity issues. To end up with this issue, self-organized neural networks are proposed. An optimal multi-user detector can be implemented, by using Growing Grid network.

Dynamic bridge generation for IoT data exchange via the MQTT protocol

Abstract Infrastructures create many problems in the security, economic or health sector. The high level of connectivity on Internet of Things (IoT) field is at the center of new networks and requires interoperable solutions that promote redundancy, and that fault tolerance. Many standard protocols currently allow extensive exchanges between different IoT groups. In this paper, we are interested in reconfigurable links management between sets of IoT linked by Message Queue Telemetry Transport (MQTT). This approach is based on the use of multi-agent systems in charge of arranging network topologies.

Multi-user detection in MC-CDMA systems and unsupervised neural network: why?

This paper is about a MC-CDMA detector based on unsupervised neural network. Systems based on MC-CDMA techniques generally require quite a complex implementation to achieve high data rates. The solution we will consider here, offers less complexity as well as realtime execution. The idea is to allow the implementation of a simulated system in a realtime system within the field of standard multi-user detection methods.

Kohonen neural networks for multi-user detection in cdma systems

Performance and implementation complexity issues restrict standard multi-user detection methods in the forthcoming high transmission rate system, based on code division multiple access. We propose self-organized neural networks to cope with this issue and suggest that an optimal multi-user detector can be implemented by using a Kohonen network.