Current research topic: Adaptive QoS in
wireless sensor networks
The main objective of our currently ongoing project is
to specify, develop and evaluate algorithms
and mechanisms for providing the adaptive QoS
(Quality of Service) in WSN (Wireless Sensors Network) that support
real-time applications and therefore to demonstrate the effective use of the WSN
technology for building CPS
(Cyber-Physical Systems) and the Internet of Things. If the monitoring of the
physical world benefited from promising results obtained within the last ten
years on WSN, these WSN have to be enriched by adding actuators, forming thus a
WSAN (Wireless Sensor and Actuator
Network), as soon as we consider applications that require not only monitoring
but also reacting on the physical world with high precision, support of long
unattended operations, promptness of reaction. Moreover and most importantly,
the QoS of the existing WSN must be enhanced since
monitoring and acting form together a closed loop within which the control
decision should be made in real-time.
Although WSN technology is economically a very
interesting solution for building CPS, unfortunately its current QoS is not sufficient for supporting such applications. In
this project, we defend the idea that currently existing WSN can be
enhanced to provide desired QoS. This could be
achieved in two coordinated directions: one is to develop the on-line
adaptive QoS management in network to cope with
the time varying performance requirement of an application; another is to
enable applications to adapt to the network working condition changes if
they go beyond the network QoS control range.
In this project, we follow a pragmatic approach by
assuming the use of the COTS components (e.g. IEEE802.15.4) at the lower
levels. The adaptive QoS are mainly studied at the routing level with
MAC-Routing cross-layer optimization and by defining and developing a QoS middleware allowing the necessary on-line
interaction between the network and the application.
Different recent PhD work that I supervised (or still
supervise) contributes to this objective:
- A network-MAC cross-layer framework for packet scheduling, congestion
control and energy consumption minimization has been developed as part of the PhD
work of B. Nefzi.
The work is based on a very simple idea of “collecting-and-burst transmitting”
scheme. The underlying MAC protocol is the widely adopted and product-supported
CSMA/CA of IEEE802.15.4. An algorithm is designed making the network
self-adapts to the dynamic traffic changes and provides improved performance.
This is typically a robust solution without or with minimized configuration
requirement. An ideal solution for large-scale soft real-time applications.
- Completely at opposite to the above-mentioned approach, we also have
explored, similar to the industrial wireless network initiatives (e.g. WirelessHART, ISA SP100, IEEE802.15.4e), the optimal
configuration schemes based on TDMA. The aim is to find dynamic TDMA cycle
configuration, optimizing both time-slot and frequency channel assignment of
networks using UWB (IEEE802.15.4a). Typical application could be large-scale
hospital networks. This work is part of J. Ben Slimane’s
PhD thesis.
- The PhD work of N. Boughanmi aimed at
developing wireless networked control
systems by an application-network co-design approach. CSMA/CA MAC protocol
is reinforced using blackburst scheme to introduce
priorities. According to the observed quality of control (e.g. tracking error),
the priorities are online reassigned, guaranteeing thus the required quality of
control.
- In addition to the industrial communication domain, another interesting
application domain is the use of WSN for providing an ambient assisted living
environment to elderly people at home (PhD work of S. Nourizadeh
under LORIA-MEDETIC contract). We also addressed the problem of the integration
of WSN into the existing home automation networks. A first demonstration
platform has been settled down at LORIA. To solve the interoperability problem,
a gateway called MPIGate has been developed. The aim
is also to develop a middleware between applications and the networks for
dynamic QoS negotiation and management.
- The PhD work of Y. Li provides a large spectrum of solutions for
providing QoS in WSN including asymmetric link
channel modeling of WSN, node deployment strategies (sensing coverage heuristic
solutions with probabilistic connectivity), multi-hop based geographic routing,
and data fusion.
To experiment and further validate the above
solutions, several WSN platforms are also established composed of nodes under either
TinyOS or Contiki, or with
industrial owner systems (e.g., Jennic, Libelium).