Views
5 years ago

Synthèse de haut-niveau de contrôleurs ultra-faible consommation ...

Synthèse de haut-niveau de contrôleurs ultra-faible consommation ...

tel-00553143, version 1

tel-00553143, version 1 - 6 Jan 2011 20 Introduction Sensor Subsystem Sensor MCU Power Supply Computation Subsystem Power Subsystem Communication Subsystem Figure 1.1: General architecture of a WSN node. versality. However, when looking more carefully to actual design practices, we observe that the need for flexibility/programmability is essentially geared toward the user application layer, which happens to represent only a small fraction of a WSN node’s processing workload. Whereas most of the processing workload is almost dedicated to the communication protocol stack. Hence, in our opinion, it is worth-studying to explore the hardware specialization approach in WSN node design as well to meet the ultra low-power requirement. In order to reduce the power consumption in a WSN node, we first need to look at the generic node architecture to find out the hotspots for power consumption. The generic architecture of a WSN node is discussed in the next section. 1.1.1 WSN node architecture WSN nodes are low-power embedded devices consisting of processing and storage components (a Microcontroller Unit (MCU) connected to a RAM and/or flash memory) combined with wireless communication capabilities (RF transceiver) and some sensors/actuators. Since these nodes must have small form-factors and limited production cost, it is not possible to provide them with large energy sources [138]. In most cases they must rely on non-replenishing (e.g. battery) or self-sufficient (e.g. solar cells) sources of energy. Figure 1.1 presents the system architecture of a generic sensor node. It is composed of four major subsystems: power supply, communication, control and computation, and sensing. The power supply subsystem consists of a battery and a DC-DC converter and has the purpose to power-up the node. The communication subsystem consists of a radio transceiver for wireless communication. Most of the platforms use a single omni-directional antenna however, cooperative “Multiple-Input and Multiple- Output (MIMO)” technology has also been deployed [99]. The processing subsystem is typically composed of memory to store application program codes and data, and of a microcontroller to control the system and process the data. The last subsystem links the sensor node to the region of interest and has a group of sensors and actuators that depend on the WSN application. It also has an Analog-to-Digital Converter (ADC) to Tx Rx

tel-00553143, version 1 - 6 Jan 2011 Wireless Sensor Network (WSN) 21 convert the analog data sensed by the sensors to digital data that can be used by the processing subsystem. To design such architecture with limited resources, the designers are faced with some tough constraints that are discussed in the following section. 1.1.2 WSN node design constraints Designing a WSN node is a challenging task, since the designers must deal with many stringent design constraints and metrics that are often interrelated. Here we will briefly discuss some of the metrics that are considered while designing a WSN node. � Power is the biggest design challenge to meet while designing individual sensor nodes to implement the applications that require multi-year life-time. � Robustness also becomes an important parameter in WSN node design to support correct functioning of the network. Each node must be designed to be as robust as possible to tolerate and adapt to neighborhood node failures. � Security at application-level is another metric to be considered while designing a node. The individual nodes must be capable of performing relatively complex encryption and authentication algorithms. � Communication bit-rate and range are key design metrics for a WSN node as well. An increase in the communication range (and bit-rate) has a significant impact on the power consumption (and computational requirement) of the node. � Computation workload is another key design metric and it directly influences a node’s power consumption. The more a node would be computationally-intensive, the more would be its overall power/energy budget. � Cost and size The physical size and cost of each individual sensor node has a significant and direct impact on the ease and cost of deployment as well as the size of the energy source available to it. As discussed earlier, since WSN nodes are deployed in huge numbers, they must be of small form-factor and inexpensive. Besides, it is not possible to equip them with large power sources. Hence, ultra low-power becomes the most critical design metric for a WSN node. It is also supported by the fact that WSN nodes may have to work unattended for long durations due to a large number of deployed nodes or a difficult access to them after deployment. If we analyze the power profile of a WSN node, among all the subsystems (Section 1.1.1), communication and computation subsystems consume bulk of a node’s available power-budget [116, 30]. In this work, we are targeting the power optimization of the computation and control subsystem of a WSN node. Indeed, we believe that power and energy savings obtained through our approach could open possibilities for more computationally demanding protocols or modulation schemes which, as a result, would provide better Quality-of-Service (QoS), lower transmission energy and higher

Synthèse, caractérisation et intérêt biomédical de (glyco ...
Synthèse, caractérisation et polymérisation par ouverture de cycle ...
Analyse et synthèse de sons de piano par modèles physiques et de ...
Emission gamma de haute énergie dans les systèmes binaires ...
Martin Teichmann Atomes de lithium-6 ultra froids dans la ... - TEL