Section: New Results
Real time services and protocols
In this area, we developed, on the one hand, policies for managing the quality of service of operating support (mainly, networks and protocols) in order to meet the properties required by real time applications (hard real time, weakly hard real time) and, on the other hand, strategies for scheduling activities and memory management.
Network-MAC cross-layer framework for differentiated QoS in wireless sensor networks
Participants : Hugo Cruz Sanchez, Bilel Nefzi, Ye-Qiong Song.
Self-adaptive QoS mechanism is preferable in large-scale wireless sensor networks because of frequent network condition changes and the difficulty to statically configure the network parameters. A network-MAC cross-layer framework has been developed for facilitating packet scheduling, congestion control and energy consumption minimization. The work is based on a very simple idea of “collecting-and-transmitting burst” scheme, called CoSenS (Collecting and Sending burst Scheme). The underlying MAC protocol is the widely adopted and deployed unslotted CSMA/CA of IEEE802.15.4. An algorithm is designed making the network self-adapts to the dynamic traffic changes. CoSenS provides a simple but efficient improvement of the MAC layer of IEEE 802.15.4 in terms of reliability, delay and throughput. The network layer uses static routing or hierarchical tree routing of ZigBee standard [47] . A scheduling algorithm GVF (Greatest Velocity First) is also developed above CoSenS. CoSenS itself has been extended to a scheme called SCSP (Send Collect and Sleep Protocol) which supports active/inactive periods of nodes, allowing thus to save energy. SCSP uses a modified version of the Zigbee hierarchical tree routing protocol [34] .
QoS in UWB-based sensor networks
Participants : Jamila Ben Sliman, Mounir Frikha [ INIT, SupCom, Tunisia ] , Anis Koubâa [ ISEP-IPP-Politechnic Institute of Porto, Portugal ] , Ye-Qiong Song.
IEEE802.15.4a provides higher data rates with smaller energy consumption thanks to the UWB (Ultra Wide Band) technology. However there exist few solutions on how to optimally exploit the great potential of this new standard. Similar to the industrial wireless network initiatives (e.g. WirelessHART, ISA SP100, IEEE802.15.4e), we are developing optimal configuration schemes based on TDMA. The aim is to find dynamic TDMA cycle configuration that optimally assigns both time-slots and frequency channels. Typical applications could be large-scale mesh networks. Hospital WSN introduced in [20] is an example where the system is organized in three tiers (Body Sensor Network, Personal Area Network and UWB/Wifi Mesh cellular network) using UWB and Wifi technologies. To ensure an efficient channel allocation, we proposed in [19] a control and data channel allocation which decomposes the frequency allocation problem into two sub-problems: static control channel allocation to ensure a permanent control channel frequency per PAN, which avoids the control channel congestion problem, and dynamic data channel allocation based on PANs duty cycle's information and spatial frequency reuse to avoid the underutilization of spectrum resource. Moreover, we proposed in [18] a prioritized multi-channel multi-time slot MAC protocol which, according to the network configuration and the available spectrum resource, tries to efficiently assign time slots per channel in response to received resource requests taking into account priority levels and some other QoS requirements.
Wireless Networked control systems (WNCS)
Participants : Najet Boughanmi, Hugo Cruz Sanchez, Eric Rondeau [ CRAN UMR 7039, Nancy ] , Ye-Qiong Song.
With recent technology progress, it is becoming attractive to use wireless solutions for industrial process monitoring and control. Our approach for developing wireless networked control systems (WNCS) is based on the application and network co-design principle [17] . The idea is to on-line adjust the network parameters according to the needs of the control loops (typically represented by the tracking error). For achieving thus on a WSN (Wireless Sensor Network) which is based on CSMA/CA MAC protocol, several enhancements must be done. In [23] several possible solutions are investigated including probabilistic priority by adjusting the minimum waiting time, deterministic priority using black burst mechanism in IEEE 802.15.4/ZigBee and the beacon-enabled mode using the Guaranteed Time Slot (GTS) mechanism. [22] presents online adaptation of the IEEE 802.15.4 parameters for WNCS. Following the tracking error of a control loop, the macMinBE parameter of the IEEE 802.15.4 MAC protocol (non beacon-enabled mode) is adjusted to provide necessary priority to the nodes of the control loop. As part of GIS 3SGS Conecs and CPER Cownecs projects, in [40] our solutions of WNCS have been extended to include multi-hops and applied to the control of an industrial process (a high speed sliding crane).
Wireless networks for ambient assisted living systems
Participants : Claude Deroussent [ MEDeTIC ] , Shahram Nourizadeh, Ye-Qiong Song, Jean-Pierre Thomesse.
Wireless sensor networks have a great potential for contributing to build the ambient assisted living environment to elderly people at home (PhD work of S. Nourizadeh under LORIA-MEDETIC contract). However several problems have to be addressed for the integration of WSN into the existing home automation networks. The first problem we have addressed is the optimal data routing in the WSN which is under multiple constraints (energy, mobility, node reliability, link reliability). In [39] and [38] a new routing protocol based on dynamic clustering and fuzzy logic is proposed. Simulations showed its good performance. The second problem is the high false alarm rate of the system, partially because of the frequent sensor node failures. Always based on the fuzzy logic approach, a node failure detection algorithm has been proposed [37] . This algorithm exploits the node failure history to assign a confidence level to the nodes, preventing thus the use of too frequently failing nodes. For testing our solutions, a platform has been built including both home automation and health monitoring parts ([36] , [35] and [43] ).
QoS routing protocols
Participants : Fares Ameur, Chung Shue Chen [ University of Hong-Kong ] , Yanjun Li [ Zhejiang University ] , René Schott, Ye-Qiong Song, Zhi Wang [ Zhejiang University ] .
For supporting time-constrained or more generally performance-requiring applications, a two-hop neighborhood information-based routing protocol is proposed [11] . Similar to SPEED protocol, the packet deadline is mapped to a velocity. The routing decision is based on two-hop velocity integrated with energy balancing mechanism. Initiative drop control is embedded to enhance energy efficiency. Simulations show that the new protocol has led to lower packet deadline miss ratio and higher energy efficiency than two existing popular schemes. The result indicates a promising direction in supporting real-time QoS for wireless sensor networks. Connectivity is a fundamental issue in multi-hop wireless sensor networks. However, node sleeping periods, unreliable and asymmetric links have a great impact on the global quality of connectivity and the routing protocol performance. The classic result established by F. Xue and P.R. Kumar on the number of neighbors needed (lower bound) for connectivity is no longer hold. By simulations we studied the 1-connectivity probability in different network conditions including asymmetric links and varying duty cycles. Further work is under going to find the required minimal neighbors in function of duty cycles.
Wireless sensor network testbed
Participants : Ahlam Bencheikh, Hugo Cruz Sanchez, Shahram Nourizadeh, Ye-Qiong Song.
Simulation models make necessarily assumptions. Some real world network and application behaviors are hard to be totally captured. To experiment our solutions, in addition to simulations, we developed a WSN platform. This platform includes not only nodes under TinyOS but also the state-of-the-art nodes of ZigbeePRO and 6loWPAN. It allowed us to test the robustness of our previously developed routing protocols, while other proposals such as N-MAC protocol are under integration.
