Scholarly article on topic 'Protocol for Latency Reduction of Prioritized Traffic in WSN'

Protocol for Latency Reduction of Prioritized Traffic in WSN Academic research paper on "Computer and information sciences"

Share paper
Academic journal
Procedia Computer Science
OECD Field of science
{"IEEE 802.15.4" / "latency reduction" / event-driven / "service differentiation" / "Contention window differentiation" / "dynamic p-persistent CSMA."}

Abstract of research paper on Computer and information sciences, author of scientific article — Kiran Kumar Pattanaik, Shivani Sharma

Abstract In critical scenarios nodes switch from low data rate of transmission mode to high data rate. Such scenarios are event driven and traffic is highly dynamic in nature. Latency and reliability issues with Quality of Service (QoS) for such applications in IEEE 802.15.4 have been an issue. The earlier schemes did address them but high data rates and channel error rates resulted in increase in contention for GTS request packets and increased data packet drop rate. In our work, we employed a mechanism to deal with isuues arising out of additional contention caused by prioritized nodes in CAP. We present an enhanced MAC scheme that incorporates two-level service differentiation through Contention Window Differentiation (CWD) and dynamic p-persistent CSMA, for mitigating the issues like lower transmission delays and higher success probability to time-critical messages in high channel error conditions. We evaluate our proposed scheme for varying channel and traffic-load conditions by simulation using Omnet+ + .The results show that performance in terms of the mentioned issues has improved significantly. Further, the energy decline rate analysis have been done for varying traffic arrival rates and compared with the original IEEE 802.15.4 and found that our scheme has less energy decline rate thus increased network lifetime.

Academic research paper on topic "Protocol for Latency Reduction of Prioritized Traffic in WSN"

Available online at

SciVerse ScienceDirect PfOCGCl ¡Q

Computer Science

Procedía Computer Science 19 (2013) 265 - 272

The 4th International Conference on Ambient Systems, Networks and Technologies


Protocol for latency reduction of prioritized traffic in WSN

Kiran Kumar Pattanaik, Shivani Sharma

Wireless Sensor Network Laboratory, ABV-Indian Institute of Information Technology and Management, Gwalior, India.


In critical scenarios nodes switch from low data rate of transmission mode to high data rate. Such scenarios are event driven and traffic is highly dynamic in nature. Latency and reliability issues with Quality of Service (QoS) for such applications in IEEE 802.15.4 have been an issue. The earlier schemes did address them but high data rates and channel error rates resulted in increase in contention for GTS request packets and increased data packet drop rate. In our work, we employed a mechanism to deal with isuues arising out of additional contention caused by prioritized nodes in CAP. We present an enhanced MAC scheme that incorporates two-level service differentiation through Contention Window Differentiation (CWD) and dynamic p-persistent CSMA, for mitigating the issues like lower transmission delays and higher success probability to time-critical messages in high channel error conditions. We evaluate our proposed scheme for varying channel and traffic-load conditions by simulation using Omnet++.The results show that performance in terms of the mentioned issues has improved significantly. Further, the energy decline rate analysis have been done for varying traffic arrival rates and compared with the original IEEE 802.15.4 and found that our scheme has less energy decline rate thus increased network lifetime.

© 2013 The Authors. Published by Elsevier B.V.

Selection and peer-review under responsibility of Elhadi M. Shakshuki

Keywords: IEEE 802.15.4, latency reduction, event-driven, service differentiation, Contention window differentiation, dynamic p-persistent CSMA.


Low energy consumption and low cost has made WSNs to be used for wide range of applications ranging from home automation, industry automation and monitoring, environmental monitoring to military surveillance, healthcare etc. Specifically, traffic in emergency response applications of wireless sensor networks are event driven and dynamic in nature. It is desirable that MAC protocol should adapt to connectivity variations effectively and efficiently. Timeliness and reliability of transmitted data is of paramount importance in such applications. The [18] and IEEE 802.15.4-2006 standards do not provide enhanced features to deal with such application specific requirements.

In [18] standard the failed GTS packets face delay of duration of inactive period till they can transmit again in the next superframe. This wait period is unacceptable in any event driven emergency response

Email addresses: kkpatnaik@iiitm. ac. in (Kiran Kumar Pattanaik), shivani. vce@gmail. com (Shivani Sharma)


1877-0509 © 2013 The Authors. Published by Elsevier B.V. Selection and peer-review under responsibility of Elhadi M. Shakshuki doi: 10.1016/j.procs.2013.06.038

application. The swap and retransmission mechanism proposed by [1] allowed the nodes with failed GTS transmission to re-transmit in CAP thus increasing the number of contending nodes. This results increased collision and not suitable for applications under consideration. In Extended CFP scheme [2] the introduction of Extra Contention Free Period (ECFP) resulted in reduced CAP in the superframe structure. Both the approaches allowed the nodes to transmit in the same superframe rather than waiting for the next superframe and improved response time as compared to the basic IEEE 802.15.4 standard. But reduction in CAP [2] causes collision and overall decrease in WSN lifetime with increase in packet arrival rate and channel error rate.

Our hypothesis debates on two viewpoints:

i. The contending WSN nodes can be made aware of their probability of acquiring the channel before contending. This helps in reducing unnecessary traffic for channel access and subsequent reduction in collision and increase in overall network lifetime. In this context we proposed modification to the IEEE 802.15.4 MAC superframe structure to deal with the latency and reliability issues. The p-CSMA approach is used for obtaining channel access probabilities in dynamic channel error conditions using Hui's ratio [3] for both prioritized and non-prioritized WSN nodes.

ii. To accommodate service differentiation for the priority based contention the entire set of WSN nodes is categorized into prioritized and non-prioritized class. This is accomplished naturally through Contention Window Differentiation (CWD) mechanism. This limits the number of nodes trying to transmit at a same time thus by reducing collision.

The rest of the paper is organized as follows. Section 2 summarizes the related work on latency reduction and service differentiation of IEEE 802.15.4. Section 3 describes the proposed protocol. The simulation setup, result and performance analysis of the proposed model follows in Section 4. Finally, Section 5 concludes the paper.

2. Related work

The latency issues, collision avoidance and the related effect on the overall working of WSN especially in emergency response applications has been addressed in the literature by various approaches such as:

i. Ndih et al. suggested different set of access parameters and different frame length for different priority classes of nodes [4],

ii. Kim et al. suggested service differentiation with the use of Priority Toning strategy along with Frame Tailoring strategy [5]. Similarly swap and retransmission strategy in [1] and ECFP in [6] was implemented to study latency reduction in emergency response applications.

In all the related research to the best of our observation, countable effort doesnot exists which cooperatively treat the collision avoidance and channel error conditions for examining the reduction in latency and improved WSN lifetime. Hence, our work attempts to study the above by introducing the approach of dynamically calculating p-persistent CSMA. on the basis of channel error conditions.

2.1. Adaptive MAC related work

Several algorithms have been proposed to adaptively tune the MAC parameters to improve different metrics. Before sending data a node waits backoff period (BP) UnitBackoffPeriods where BP is a randomly chosen number between 0 to 2BE - 1. [7] proposed a delayed backoff algorithm (DBA) to improve the channel utilization and power consumption. Instead of a random backoff time the algorithm assigned different backoff time for the next transmission of the sensor node according to its demand during the current transmission. The results have shown improved throughput, delivery ratio and reduced power consumption for the varying network load. [8] introduced adaptive FEC algorithm to reduce the retransmission traffic thus improving the throughput, packet delivery ratio, delay, packet loss, and error rate. [9] used physical layer optimization approach and focussed on reliability aspect of data transmission by controlling the activity periods of sensor nodes using Bernoulli's approach. Managing the sleep periods enabled less contention for

channel access thus increased reliability as well as improved network life time. [10] proposed a probability based scheme to set the Contention Window (CW) size according to the number of nodes reporting a common event. A minimum initial CW size was used and halved with probability 'f' after each successful transmission. The algorithm had shown the results with reduced collision probability with the improvement in network throughput and latency. [11] also proposed a dynamic tuning algorithm of CW size with the evaluation on good throughput, reliability and average delay.

Most of the strategies discussed above concentrated primarily on collision avoidance or throughput improvement. However, [12] considered optimizing all MAC parameters at once to minimize the power consumption by estimating busy channel and channel access probabilities for various changing traffic regimes and number of active nodes. The results confirmed the objective function by showing a longer lifetime of the network. In another work, [13] presented an adaptive differentiated p-persistent CSMA protocol aimed for 802.11. It has shown improvement in throughput by 20 by alleviating the effect of channel degradation. Therefore, this algorithm and its approach is basis of study in our work for IEEE 802.15.4.

2.2. Service Differentiation related work

The performance of slotted CSMA can be varied by changing values of any of the four MAC parameters: the minimum backoff exponent (macMinBE), maximum backoff exponent (aMaxBE), CWinit length and maximum number of backoffs (macMaxCSMABackoffs). The underlying idea behind this is to provide differentiation and priority to group of nodes. [14] proposed the differentiation by using Priority Toning and Frame Tailoring strategy to enable fast delivery of high priority frames. But the solution is incompatible with the standard as requires the change in the MAC protocol. [15] provides service differentiation by assigning different values of MAC parameters for data and command traffic. The results have shown the improved QoS for time critical messages. [5] presented a discrete time Markov Chain mathematical model to provide service differentiation by varying Backoff Exponent (BE) and CW. Their results showed that BED is better for channel utilization while CWD provides low latency. [16] implemented TRADIF on IEEE 802.15.4 stack in ERIKA. The results showed that adequately tuning the parameters of slotted CSMA/CA leads to an improved QoS for time-critical messages. MAC scheme by [1] proposed the swap and retransmission scheme where the position of CAP and CFP were interchanged in the superframe structure to allow the retransmission of failed GTS frames in the same superframe. In continuation with this work, an extended CFP was appended at the end of CFP by [2], which consists of GTS slots called as extended GTS. This was done in order to provide prioritized channel access to failed GTS transmissions to ensure timely response of critical messages. Both the approaches suffer from degraded performance and high energy consumption at high channel error rates, because of increased collision in the CAP due to retransmissions. Moreover, we find that the high congestion may result in failure of GTS request messages in CAP and hence delay in allocation, summing to overall latency for time-critical traffic.

3. Proposed Protocol

For applications with strict latency requirements, swap and retransmission method proposed in [1]reduced latency by considerable amount. But for high data and channel error rates, the large numbers of GTS frames suffer from high transmission delay. The performance degradation was observed in [6] too in the form of increased frame drop rate and transmission delay at high channel error rates. Moreover, long wake-up times of the nodes to sense the channel idle for transmission make them lose their battery power fast which is also not desirable.

The design goals of the proposed protocol for time-critical applications of WSNs are: Implementation simplicity, Adaptability to dynamic channel error conditions, Low latency for prioritized data, High-throughput in terms of received data, Energy-efficiency in terms of increased network lifetime.

We have addressed the above design goals in the following manner:

• Implementation complexity is low as the superframe structure proposed by [1]is used instead of complex approaches followed in [14] and [6].

• Adaptability to dynamic channel error conditions due to use of p-CSMA instead of CSMA in the original standard

• Low latency for prioritized data by introducing CWD

• Decreased packet drop due to improved channel contention mechanism

• Increased network lifetime due to better management of active/inactive periods of nodes for channel access

The principles behind the design goals concerning dynamic p-CSMA and CWD are elaborated further in the following sections. The other design goals are excluded in the discussion due to their implicit presence in the two principal goals under consideration.

3.1. Contention Window Differentiation

The CWD is to classify the nodes in order to differentiate nodes that transmit critical data against the ones that are not. The number of such nodes remains static throughout once identified. Prioritized nodes send critical data in Guaranteed Time Slot (GTS). Failed GTSs frames and GTS request messages from the prioritized nodes will contend along with non-prioritized nodes for the channel access in CAP. This helps the nodes to transmit critical data with minimal delay. The delay caused in transmitting non critical data is acceptable in event driven emergency response applications. In order to accommodate both classes of nodes we visualize two options in terms of setting CW size [15].

In the first, the size of CW has been increased to 3 which is optimum [15], while for the prioritized nodes it is set to default CW size of 2. In the simulation we performed tests for different values of CW with reference to non prioritized nodes and the results are depicted in Fig. 1 in Section IV. This causes low priority nodes get a back seat while accessing channel as compared with prioritized nodes. However, this situation is of less importance while transmission of critical data is of higher priority. Hence this overhead can be accommodated at the cost of criticalness of emergency response applications. However, the wait period for transmitting data for non prioritised nodes may vary depending on the traffic from prioritized nodes. In the second, reducing the CW to 1 for prioritized nodes and keeping CW as 2 for non prioritized nodes will cause collision with acknowledgment frames.

3.2. Link Quality Indication

Reliable and fast communication is important factor for QoS of emergency response applications but erroneous wireless link and limited bandwidth pose challenge to achieve this.

In this paper, we propose to improve the QoS of prioritized traffic on the basis of Link Quality Indicator (LQI). The LQI is the measurement of the quality of the received packet at the physical layer and ranges between 110 and 50 [17]. LQI is the Signal to Noise Ratio (SNR) and is transmitted from physical layer to MAC layer with each packet.

The Bit Error Rate (BER) is calculated which signifies varying channel error conditions for different LQIs. The bit error probability pb is the expectation value of the BER. A packet is assumed to be incorrect if at least one bit is incorrect. Hence estimation of packet error rate (PER) takes into account the number of incorrectly received data packets against the number of transferred packets.

3.3. Dynamic P-Persistent CSMA-CA Algorithm

In WSN, variations in channel conditions are inherent. Hence use of p-persistent CSMA may not be perfect for analyzing throughput in WSNs. The use of dynamic p-CSMA is justifiable due to the reason why it incorporates the dynamicity of channel conditions. The proposed algorithm is the modification of the basic IEEE 802.15.4 in terms of CW differentiation and dynamic p-CSMA calculation in order to study their implication in transmitting prioritized data with an objective to minimize delay and increase throughput in emergency response applications. The original algorithm is reproduced in Algorithm 1 given below with necessary changes therein.

Algorithm 1: Dynamic p-Persistent CSMA

If packet ready; CWP := 2, CWn := 3; NB := 0; if BatteryLifeExtension == 1 then | BE := mm(2,macMinBE);

I BE := macMinBE',

backoff := rand(2BE - 1) x unitBackoff Period; do CCA; if channel idle then r := rand(0,l); Calculate dynamic-p; if r < p then CW := CW - 1; if CW == 0 then | Transmit; else

I Go to 10; end

| delay one backoff period; Go to 10; end else

NB := NB + 1; BE := min(BE + 1, aMaxBE); CWP :=2, CW„ :=3; if A© < macMaxCS MABAckoff sthen I Go to 9; else

I Failure: end end

In the algorithm, we have firstly initialized the CW value according to the priority of nodes. For the prioritized nodes, CWp = 2 and for non-prioritized nodes, CWn = 3.

The optimal p value for varying link conditions is estimated by using Hui'sRatio [3]. In our case for the two different traffic types, the Hui's approach for calculating network throughput can be simplified as

S.-.S^-P^:-^- (1)

1 - p1 1 - p2

where, p1 be the persistence value for prioritized nodes and p2 for non-prioritized nodes. A throughput ratio of 2:1 signifies the priority levels of traffic.

For the dynamic p-CSMA, wireless-channel conditions decide the p value. Hence the dynamic-p persistent value, pd, for the two traffic types can be calculated on the basis of PER, Pe as given below:

pd = p1(1 - Pe) (2)

pd = p2(1 - Pe) (3)

Pe = 1 - (1 - pb)L (4)

where L is the length of the packet and pb is the bit error rate. 4. Simulation Setup

In this paper, we used the Castalia 3.0 Simulator for WSN. It is based on the OMNeT++ platform and can be used to model WSNs and test distributed algorithms and/or protocols in realistic wireless channel

and radio models, with a realistic node behaviour especially relating to access of the radiolike modelling of Received Signal Strength Indicator (RSSI), carrier sensing, path loss variation etc.

A total of seven CFP nodes were considered for worst case performance analysis. The other MAC parameters like CSMA backoff exponent, inter-frame spacing (IFS), macMinBE, macMaxFrameRetriesetc. are set to the default values as specified in the standard. The superframe duration is 61.44 ms with CAP of 33.94 ms and CFP of 26.8 ms. The simulation duration was of 300 seconds. We worked on the 2.4 Ghz ISM band with maximum permitted data rate 250 Kbps. For the radio parameters we have used CC2420 file, which defines all the basic operational properties of a radio. The transmit power level was set to -5 dB. The MAC packet is of the maximum size as specified by the standard, i.e. 300 bits for data payload and 104 bits for the MAC.

4.1. Results And Performance Analysis

This section discusses the simulation results obtained for our proposed scheme and analyses the out-come.Primarily we analyse the outcome from these viewpoints viz. low latency for prioritized traffic, throughput of received data, and overall network lifetime.

Fig. 1. Delay incurred with CWD

Fig. 2. Energy consumption with comparision to basic IEEE 802.15.4

In this analysis the emphasis has been given to the effect of varying CW in emergency response application. As from the Fig. 1 we can see that the delay experienced by nodes increases with the increase in CW size and with traffic arrival rate. As mentioned in the simulation setup the duration of the CAP is 33.94 seconds, so for high value of CW chances of getting channel access in same superframe are much diminished due to high delay observed. As our work requires the service-differentiation to nodes, so we have opted for the next higher value of CW i.e 3 to the value 2 specified in standard. This will ensure the categorization of nodes without much affecting the transmission delay.

Fig. 3. Energy Consumption with different traffic arrival rates

Fig. 4. Channel access probability with different traffic arrival rate

Fig. 2 and Fig. 3 shows the energy consumption of each node for varying CSMA traffic arrival rate. The cases for low traffic arrival rate and high traffic arrival rates have been considered individually. As we used two AA batteries in the simulation, the initial energy is 18720 joules. It is evident from Fig. 2 that the energy decline rate for IEEE 802.15.4 MAC is fast as compared to our proposed scheme for low traffic arrival rate. This is due to the fact that at high channel error rates, the non-prioritized nodes tend to defer the transmission

more because of low dynamic-p, less waste of energy due to collision avoidance. The prioritized channel access in case of prioritized nodes ensures their early sleep mode following every successful transmission, thus conserving energy. This shows overall improvement in the energy consumption. Due to increased packet generation or high packet arrival rates the nodes remain active for considerable duration causing decline in stored energy (Fig. 3). This is worth mentioning to avoid confusion arising out of whether this decline in energy level is due to retransmissions and collisions. Our proposed model has been able to successfully address the issues related to retransmissions and collisions.

Fig. 4 shows the channel access probability with different traffic arrival rate. At low traffic arrival rates significant improvement in channel access probability is not observed for our proposed mechanism at low channel error rate. This is because there are very little chances of nodes accessing the channel at same time However at high error rates, the channel access probability greatly improves with the proposed scheme. This is because of high p value assigned to retransmitted packets relative to the traffic from the non-prioritized nodes. High CW and low dynamic -p value limits the number of nodes contending for the channel and thus high channel access probability. At high traffic arrival rate relative degradation in performance is observed, but still improvement over the basic scheme. It can be accounted for the reason that at high contending population and channel error probability, chances of channel access for non-prioritized nodes are very feeble, hence a drop in the overall channel access probability.

Fig. 5. Packet delivery ratio with packet arrival rate

Fig. 6. Throughput with different traffic arrival rates

Fig. 5 shows the relation between the packet delivery ratio and packet arrival rate. This is in order to visualize the amount of successful deliveries. The degree of successfulness of packet delivery at sink is the main concern in emergency response systems. At very low CSMA traffic arrival rate, it is observed that prioritized nodes have slightly high delivery rate as compared their counterparts. This difference grows with increasing traffic rate. It can be explained by the fact that non-prioritized nodes are assigned very low p at high contention and probability of channel error. This is further aggravated by high CW which defers the non-prioritized nodes to next transmission while prioritized nodes gain the channel access.

Fig. 6 illustrates the throughput of the network. Throughput can be defined as the time taken by the sink node to receive all sent packets. At low packet error rate the throughput is fairly high. As the channel error rate increases the throughput begins to decrease. At very high error rate drop is observed in throughput but by very small amount. This can be accounted to the reduced collisions and retransmissions by the application of proposed algorithm.

?t = 0)

?t-0) ?t-02 I

60 .. . * • •« . . • . . '



(0 20 3.0 40 50 60 70 10

Traffic arrival rate fPkts/sec)

Fig. 7. Latency in transmission

The transmission delay with different traffic arrival rate has been shown in Fig. 7. At low data rate, the transmission delay is similar between different channel error rates. However, at high data rate, low channel access probability and large CW for non-prioritized nodes make them to wait for long before attempting transmission. Hence the overall high transmission delay is observed at high channel error rates.

5. Conclusion

In this paper, the issues pertaining to transmission requirements and network lifetime of WSNs meant for emergency response applications were discussed. The simulation studies shown that through node prioritizing and dynamically calculating p-CSMA we are able to achieve improved packet delivery which is a natural consequence of the modification to the basic MAC scheme proposed in IEEE 802.15.4. Secondly, the effect of proposed channel access mechanism on the overall lifetime of WSN while meeting the transmission requirements for emergency response applications have been justified. In addition, the effectiveness of our proposed mechanism against the original protocol from various other dimensions are analysed which suffice our argument.


[1] Bhatti, G., Mehta, A., Sahinoglu, Z., Zhang, J. and Viswanathan, R. Modified beacon-enabled IEEE 802.15. 4 MAC for lower latency. Global Telecommunications Conference (IEEE GLOBECOM). 2008, 6(4): pp.1-5

[2] Mehta, A., Bhatti, G., Sahinoglu, Z., Viswanathan, R. and Zhang, J. Performance analysis of beacon-enabled IEEE 802.15. 4 MAC for emergency response applications. IEEE 3rd International Symposium on Advanced Networks and Telecommunication Systems (ANTS). 2009, pp.1-3

[3] Hui, J. and Devetsikiotis, M. Designing improved MAC packet schedulers for 802.11 e WLAN. Global Telecommunications Conference (IEEE GLOBECOM'03). 2003, 1: pp.184-189

[4] Ndih, E.D.N., Khaled, N. and De Micheli, G. An analytical model for the contention access period of the slotted IEEE 802.15. 4 with service differentiation. IEEE International Conference on Communications (ICC'09). 2009, pp.1-6

[5] Kim, M. and Kang, C.H. Priority-Based Service-Differentiation Scheme for IEEE 802.15.4 Sensor Networks in Nonsaturation Environments. IEEE Transactions on Vehicular Technology. 2010, 59(7): pp.3524-3535

[6] Mehta A. , Bhatti G. , Sahinoglu Z. Viswanathan R. and Zhang J. A modified beacon-enabled IEEE 802.15. 4 MAC emergency response applications. IEEE Symposium on Computers and Communications (ISCC). 2010, pp.261-267

[7] Lee, B.H. and Wu, H.K. A delayed backoff algorithm for IEEE 802.15.4 beacon-enabled LR-WPAN. 6th International Conference on Information, Communications & Signal Processing. 2008, pp. 1-4

[8] Ebenezar jebarani, M.R. and Jyanthy,T. An Analysis of Various Parameters in Wireless Sensor Networks Using Adaptive FEC Technique. International Journal of Ad hoc, Sensor & Ubiquitous Computing (IJASUC). 2010, 1(3), pp.33-43

[9] J. Misic, S. Shafi and V.B. Misic. Maintaining Reliability Through Activity Management in 802.15.4 Sensor Clusters. IEEE Transaction on Vehicular Technology. 2006, 55(3), pp.779-788

[10] Shakya, R.K., Singh, Y.N. and Verma, N.K. Optimizing Channel Access for Event-Driven Wireless Sensor Networks: Analysis and Enhancements. Arxiv preprint arXiv:1203.5874. 2012

[11] Pang, A.C. and Tseng, H.W. Dynamic backoff for wireless personal networks. Global Telecommunications Conference (GLOBE-COM'04). 2004, IEEE, 3: pp.1580-1584

[12] Park, P., Fischione, C. and Johansson, K.H. Adaptive IEEE 802.15.4 protocol for energy efficient, reliable and timely communications. ACM/IEEE IPSN, Stockholm, Sweden, 2010

[13] Abukharis, S. and O'Farrell, T. A new adaptive differentiated p-persistent CSMA protocol that reduces the effect of the transmission errors on the system throughput performance. 7th International Wireless Communications and Mobile Computing Conference (1WCMC),1EEE. 2011, pp.2181-2185.

[14] Kim, T.H. and Choi, S. Priority-based delay mitigation for event-monitoring IEEE 802.15. 4 LR-WPANs. IEEE Communications Letters. 2006, 10(3), pp.213-215

[15] Koubaa, A., Alves, M., Nefzi, B. and Song, Y.Q. Improving the IEEE 802.15.4 slotted CSMA/CA MAC for time-critical events in wireless sensor networks. 5th International Workshop on Real-Time Networks (RTN 2006). 2006

[16] Severino, R., Batsa, M., Alves, M. and Koubaa, A. A Traffic Differentiation Add-On to the IEEE 802.15. 4 Protocol: Implementation and experimental validation over a real-time operating system', 13th Euromicro Conference on Digital System Design: Architectures, Methods and Tools. 2010, pp.501-508

[17] Lee, Y.D., Jeong, D.U. and Lee, H.J. Performance analysis of wireless link quality in wireless sensor networks. 5th International Conference on Computer Sciences and Convergence Information Technology (ICCIT), 2010, pp.1006-1010

[18] IEEE Standard for Information Technology, Part 15.4: 'Wireless Medium Access Control (MAC) and Physical Layer (PHY) Specifications for Low-Rate Wireless Personal Area Networks (LR-WPANs)', 2003