Real-time multimedia applications require quality of service (QoS) provisioning in terms of bounds on delay and packet loss along with soft bandwidth guarantees. The shared nature of the wireless communication medium results in interference. Interference combined with the overheads, associated with a medium access control (MAC) protocol, and the implementation of a networking protocol stack limit the available bandwidth in IEEE 802.15.4-based networks and can result in congestion, even if the transmission rates of nodes are well below the maximum bandwidth supported by an underlying communication technology. Congestion degrades the performance of admitted real-time multimedia flow(s). Therefore, in this paper, we experimentally derive the IEEE 802.15.4 channel capacity using an unslotted CSMA-CA MAC protocol. We experimentally derive channel capacity for two cases, that is, when the CSMA-CA protocol is working without ACKs and when it is working with ACKs. Moreover, for both cases, we plot the relationship of offered data load with delay and packet loss rate. Simulation results demonstrate that the parameters that affect the choice of a CSMA-CA MAC layer protocol are end-to-end delay and packet loss requirements of a real-time multimedia flow, data load within the interference range of transmitters along the forwarding path, and length of the forwarding path.

/pdf/contiki-based-ieee-802-15-4-channel-capacity-estimation-and-4cqknj6mdd.pdf

Contiki-Based IEEE 802.15.4 Channel Capacity Estimation and Suitability of Its CSMA-CA MAC Layer Protocol for Real-Time Multimedia Applications

In the wireless sensor networks domain, one of the most used standards is IEEE 802.15.4. This standard has been made available on many low power operating systems such as TinyOS and Contiki OS. It is crucial for the implementation to be compliant with the specifications of the standard. In the case of Contiki OS, the provided version of the main medium access algorithm, unslotted Carrier Sensing Multiple Access with Collision Avoidance (CSMA/CA), presents many flaws. In this paper, we discuss the implementation of the provided CSMA/CA algorithm and we point out to the parts that do not respect the standard specifications. We also propose and implement a compliant version of this algorithm and show through simulation the correctness of the implementation.

Implementation and performance evaluation of IEEE 802.15.4 unslotted CSMA/CA protocol on Contiki OS

Although the ZigBee technology is massively deployed the performance question still remains unanswered. There is a lack of serious performance evaluations and thus this paper provides a comprehensive and unbiased study of the ZigBee PRO technology. An extensive real measurement was conducted to accomplish this task. We found out a big performance gap among ZigBee implementations even though they are compliant with the ZigBee standard. The presented results and analysis could be beneficial for many wireless sensor network developers.

Comprehensive Performance Analysis of ZigBee Technology Based on Real Measurements

With an expeditious increase in the vehicle transport in the major cities, it has become more difficult to find the free parking space to park the vehicles. Constant rise in the number of vehicles on the roads has led to increase in the unwanted traffic, drivers frustration, delay in the transportation and increase in the noise and air pollution. Due to lack of proper knowledge of parking space, it has become unable to utilize most of the nearby parking space. This work proposes an innovative approach to guide the vehicle drivers to the parking space using Hierarchical Wireless Sensor Networks. A three sensors per parking slot based technique is used to distinguish the type of parked vehicles and to find the free parking slots in the parking area. Free, partially-occupied and occupied parking slot states information is stored at the Parking Server in the form of parking location and the parking area layout with parking slot states marking on it. Parking Server makes the parking information available to the vehicle drivers by the mean of web and mobile applications for guiding vehicles for parking. The results show that the parking detection with multiple passive sensor nodes per parking slot system is more energy efficient and vehicle parking detection is as accurate as the single active sensor per parking slot system in the existing parking designs.

M-parking: Vehicle parking guidance system using hierarchical Wireless Sensor Networks

As there is a rising demand on energy efficiency the proper network stack is an essential for every wireless sensor network application. This paper gives a comprehensive view on performance comparison of $$\hbox {Atmel}^{\circledR }$$ Atmel ® Lightweight Mesh (LWM) and ZigBee Pro technology based on extensive real measurements. We focused mainly on application layer throughput, routing latency and self-healing capability. Although the LWM clearly outperforms ZigBee Pro in most aspects, the LWM has particular drawbacks. Particularly, the LWM provides smaller footprint and faster stack processing along with higher application throughput. The most notable throughput difference where the LWM outperforms the BitCloud was by as much as thirteen times.

Comprehensive Performance Analysis of Lightweight Mesh and Its Comparison with ZigBee Pro Technology

This paper presents the throughput results obtained from a Wireless Sensor Network testbed, with single and multiple sources in different network deployments and routing architectures. The experimental testbed deployed at IIT Madras consisted of commercially available Crossbow TelosB and MicaZ nodes and a custom-built sensor node based on the DigiNet Xbee chip, with all nodes implementing the Zigbee standard. The networks were deployed in uniform grid topologies in three different deployments with up to 228 nodes. The main aim of the experiments is to analyze the throughput and packet delivery ratio observed with single and multiple sources. The experimental results show that delivery ratio reduces with increase in data rate due to collisions and help characterise the network capacity limits.

Sreejith Vidhyadharan

Papers

Testbed based throughput analysis in a Wireless Sensor Network