260 likes | 396 Views
Author: M. M. Aftab Hossain Supervisor: Riku Jäntti Instructor : Aamir Mahmood Department of Communications and Networking (Comnet). Master Thesis Presentation How much room does WLAN allow for IEEE 802.15.4 Operation?. Date: September 07, 2009. Background.
E N D
Author: M. M. Aftab Hossain Supervisor: Riku Jäntti Instructor : Aamir Mahmood Department of Communications and Networking (Comnet) Master Thesis PresentationHow much room does WLAN allow for IEEE 802.15.4 Operation? Date: September 07, 2009
Background • IEEE 802.15.4 IEEE 802.11 share the same unlicensed ISM Band • IEEE 802.11 transmit power is 15-18 dB higher than that of IEEE 802.15.4 • E.g. WiFi is efficient jammer for IEEE802.15.4 • Depending on the utilized sensing scheme, there is either • High packet drop (CCA mode 2) • Low access probability (CCA mode 1)
Objectives To investigate how much room IEEE 802.11 allows for IEEE 802.15.4 operation To develop an algorithm to rank the channels based on interferer activity and strength
Abusing the term `SINR´ We abuse SINR as the difference between RSSI of recieved packet and RSSI of Interferene as From literature S is the signal power, I is the interference power n is the noise power Case 1: only noise present Case 2: both noise & intereference present
Effect of Packet rate of WLAN-1 • WLAN employs 802.11b mode • WLAN packet payload is fixed at 500 byte and 802.15.4 at 15 byte . • WLAN Packet Rate is varied to get the effect of change of wLAN activity. • If rate is less than 300 Packet/sec, there is some room to use the white spaces. • More than 15 dB SINR* is required to achieve PDR 90%.
Effect of Packet rate of WLAN-2 • WLAN employing 802.11g mode. Packet payload size is fixed at 500 byte and transmission rate is 18 Mbps. • IEEE 802.15.4 packet size is 15 byte. • More than 15 dB SINR* is required to achieve 90% PDR. • 300 packet /sec is enough to bring down PDR very low, if SINR is less than 15dB.
WLAN employs 11b mode and transmission rate is 11M. WLAN packet size is fixed at 500 and packet rate is varied. 802.15.4 payload sizes compared are 24byte and 72 byte. There is significant effect of changing the payload size of 802.15.4. For all other variables constant, PDR is much higher for smaller payload size of 802.15.4. Effect of 802.15.4 payload size
Comparison of different models • Measurement is repeated under the following models • -SISO WLAN model A • -SISO WLAN model D • -TGn model B • There is no significant variation under • Different models holding all other variables constant • SINR* and activity of WLAN interferer • are the main factors that affect PDR.
Algorithm 1Derivation of metrics -1 • Our aim is to rank the channels in terms of interference based on the RSSI samples found from the channel scanning. • As WLAN signal is bursty in nature, initially, we take 1000 sample (passive scan of the channel). • Over the decision window we record the following • Minimum of the RSSI samples • Maximum of the RSSI samples • Average of the RSSI samples
Algorithm 1Derivation of metrics -2 • 1st metric, Strength: • Strength: The difference between Maximum and Minimum of the RSSI recorded in the window. • This metric gives the strength of interfering signal at that channel. • 2nd metric: Activity • The relative difference of average from the maximum and minimum RSSI in the window in the form of Activity = 100 *(Avg of RSSI - Min of RSSI)/(Max of RSSI – Min of RSSI); • The more the average is closer to maximum, higher is the activity of the channel and more the average is closer to the minimum, the channel is less active.
Algorithm 1Schematic diagram of metrics Min of RSSI b a Mean of RSSI a c Max of RSSI First metric, strength = a Second metric ,activity = b / c * 100 A
Algorithm 1Ranking the channels • Group 1: Ranking based on signal strength • All the channels will be ranked based on the 1st metric i.e. the difference between maximum and minimum of the RSSI samples. • Group 2: Channels with strong interference level • We will sort all the channels according to strength (metric 2) and according to activity (metric 3) separately. • we will add the position of the same channel in two arrays to get a new metric based on which the channels will be ranked in group two. • When two channels will end up with same number, priority can be given to better channel by introducing threshold (example: less than 10% activity or 13 dB difference)
Algorithm 2Derivation of metrics -1 Classification of channels based on interference level and interferer activity: • In our first algorithm we use maximum, minimum and average of the RSSI samples to rank the channels according to channel interference level and activity • This time we will use the overall average of the RSSI and Average of the RSSI excluding the only noise samples with hope of improved reliability. • Over the decision window we need to record the following • Min of the RSSI samples • Mean of all the scanned RSSI samples • Mean of the interference RSSI ( Excl. noise samples)
Algorithm 2Derivation of metrics -2 • 1st metric: • The mean RSSI of the interference (excl. noise samples) This metric gives the strength of interfering signal at that channel. 2nd metric: • The relative difference of overall RSSI and interference RSSI mean as Activity = 100 *(Overall mean of RSSI - Min of RSSI)/(mean of intereference RSSI - Min of RSSI); • The more the gap between the two mean, the less is the channel activity and vice versa.
Algorithm 2Schematic diagram of metrics Min of RSSI a Overall Avg of RSSI b Avg of interference RSSI (excluding noise samples) Max of RSSI First metric , strength = b Second metric activity = a / b * 100
Algorithm 2Ranking the channels • Group 1: ranking based on signal strength • All the channels will be ranked based on the 1st metric i.e. the mean of interference RSSI • Group 2: Channels with strong interference level • We will sort all the channels according to both metric separately to place the channels in two different arrays. • we will add the position of the same channel in two arrays to get a new metric based on which the channels will be ranked in group two. • When two channels will end up with same number, priority can be given to better channel by introducing threshold (example: less than 10% activity or 13 dB difference) • When there is no suitable channel in Group1 one, only then channel will be picked from Group 2.
Ranking channels according to algorithms • Emulated wireless channels • Wirelss channel at the corridor
Emulated wireless channels • Channel simulator is used to emulate the wireless channels. • Packet rate of WLAN and Output gain (attenuation) is varied. • 16 channels are emulated as the combination of different attenuation i.e. 45, 55, 65 and 75 dB and different packet rate i.e. 100, 200, 300 and 1000 packet/sec. • Each combination of packet rate and attenuation represents a channel with a specific strength and activity. • For each combination, passive scanning is performed by the 802.15.4 nodes to estimate the channel properties. • According to the algorithms, channels are ranked.
Ranking of emulated channels based on Algorithm 1 Ranking based on strength/activity Combined ranking • The 1st rank is based on strength and 2nd • based on activity • Strength rank of any channel is less than all • the channels of its left columns • Activity rank of any channel is less than that • of any channel of its lower rows. When the ranking are combined for Group 2, the channels at upper right zone are found to be safe zone as expected
Ranking of emulated channels based on Algorithm 2 Ranking based on strength/activity Combined ranking • Algorithm 2 also ranks the channels without any error.
Wireless Channels at the corridor • A WLAN is transmiting in the corridor and is sensed by a • 802.15.4 node. • Packet rate of WLAN and distance of the WLAN transmiter relative to IEEE802.15.4 node is varied • Each combination of Packet rate and distance represents a particular channel with a specific strength and activity • combinations of five packet rate 100, 200, 300, 500 and 1000 • packet /sec and three different distannce P1(6m), P2 • (15m) and P3 (18m, next wing) represent 15 channels. • For each combination, passive scanning is performed by the 802.15.4 node to estimate the channel properties. • According to the algorithms, channels are ranked.
Ranking of channels at the corridor Algorithm 1 Ranking based on strength/activity Combined ranking • The 1st rank is based on strength and 2nd • based on activity • The channel ranked 13th in strength rank and • the channel ranked 10th in activity rank is • wrongly placed. • Ranking of other channels are correct • In combined ranking, the best channels are still seperated in the upper right zone of the table.
Ranking of channels at the corridor Algorithm 2 Ranking based on strength/activity Combined ranking • Algorithm 2 ranks all the channels without any • exception