Performance enhancement of overlapping BSSs via dynamic transmit power control
© Lei and Rhee; licensee Springer. 2015
Received: 26 February 2014
Accepted: 22 December 2014
Published: 24 January 2015
In densely deployed wireless local area networks (WLANs), overlapping basic service sets (BSSs) may suffer from severe performance degradations. Mobile stations in a BSS may compete for channel access with stations that belong to another BSS in such environment, and it reduces overall throughput due to the increased collision probability. In this paper, we propose a new scheme for transmit power control, which enables mobile stations to dynamically adjust their transmit powers. Using our mechanism, those stations in different BSSs will have more chances of simultaneous transmissions and thus improve their performances by enhancing spatial reuse. We develop a Markov chain model to analyze the performance of the proposed scheme and also perform extensive simulations. Both the analytical and simulation results show that our mechanism effectively improves the network performance of WLANs.
As IEEE 802.11 wireless local area networks (WLANs) have been widely deployed in homes, offices, and public places , the high density of WLANs has posed a great concern on the problem of co-channel interferences. Thus, the overall network performance of WLANs may be severely degraded unless an efficient scheme is provided to reduce the interference.
A WLAN basic service set (BSS) is typically formed by an access point (AP) and a number of stations associated with the AP , and in that case, data transmissions are allowed only between the stations and the AP. When the coverage of nearby co-channel BSSs overlaps with each other, they are called overlapping BSSs (OBSSs) . In case a station located at the overlapping area transmits frames, other stations of the neighbor BSS may sense the transmission and refrain their transmissions. Also if they cannot sense the transmission, then they will become hidden terminals to the transmitter. Therefore, the chance of simultaneous transmissions among OBSSs is reduced, and thus the whole network may suffer from the poor spatial reuse of OBSSs.
Many solutions have been suggested so far to dynamically control the transmit power of WLAN stations and thus to improve the overall throughput of the network [1,4-8]. By adopting those schemes, stations are able to reduce their transmission ranges using only proper amount of transmit power, such that more stations can simultaneously transmit and thus the overall throughput is increased. The previous works, however, may not be adopted in a practical WLAN system: For example, the problem of how to determine the proper power level is not fully investigated in [1,4]. Also  is based on assumptions that may not be possible in real world,  requires the real-time adaptation of a measurement algorithm, and power control scheme in  is limited to use only in ad hoc mode.
In this paper, we propose a method for dynamic transmit power control to enhance the throughput of OBSSs. First, we study the four different radio ranges in 802.11 systems and how OBSSs interfere with each other in a density WLAN. Based on these observations, we propose a new power control scheme such that every station keeps a table for recording the path loss between itself and the neighbor BSS stations from which request to send/clear to send (RTS/CTS) frames can be overheard. Utilizing the information, those stations adjust their transmit powers and data frames are delivered using only the proper powers. We develop a discrete-time Markov chain model in order to verify that our proposed method provides the OBSSs with more opportunities of simultaneous transmissions and thus increases spatial reuse. In addition, simulation results are presented to validate our proposed scheme and its analytical model.
The remaining part of this paper is organized as follows. We discuss the related previous works and study the interference occurred in OBSSs in Section 1. The details of our proposed power control method are addressed in Section 1, and the Markov chain model is investigated in Section 1. The extensive simulation results are reported in Section 1, and finally, concluding remarks are drawn in Section 1.
2 Problem definition and related works
2.1 Problem definition
2.2 Related works
Several attempts have been made to improve the performance of 802.11 MAC by utilizing transmit power control scheme. Since the transmit power control (TPC) method standardized in the IEEE 802.11 suffers from inaccuracies, Oteri et al.  propose a fractional CSMA/CA scheme by combining the TPC with user grouping and inter-BSS coordination to improve the performance of overlapping BSSs. However, their approach lacks of a mechanism for determining the proper transmit power. In , an iterative power control algorithm is proposed to increase the number of concurrent transmissions in the dense wireless networks. This proposal is based on the assumptions that every node has complete knowledge of the network topology and current configuration, which may not be possible in real world.
In , a run-time self-adaptation algorithm is proposed based on packet loss differentiation, which can jointly adapt both transmit power and physical carrier sensing (PCS) threshold. The problem of this scheme is that it requires metrics such as PER and interference level to be measured in real time which can increase the burden of system. Also Cesana et al.  present an interference aware MAC for ad hoc mode network, in which each station transmits using RTS/CTS procedure, and the information about reception powers of RTS frames and interference levels is inserted into CTS packets. Utilizing the information, stations which overhear a CTS can tune their transmit powers such that they can transmit simultaneously without interfering with each other.
For performance enhancement of OBSSs, many recent works have provided different approaches. Li et al.  propose an interference avoidance algorithm to mitigate the interference from the neighbor BSS operating at the same channel. However, this scheme enables AP to drop its defer threshold to energy detect threshold when transmitting to stations located at overlapping area. Thus a hidden terminal to AP can sense the transmission from AP and the collision probability is reduced. Fang et al.  propose a PCF-based two-level carrier sensing mechanism which adopts two NAVs in stations, namely self BSS network NAV (SBNAV) and OBSSs network NAV (OBNAV). When a transmission processes in one of the BSSs, the station which senses it sets the value of its NAV to be either SBNAV or OBNAV, whichever is bigger. If there are no OBSSs, the OBNAV is set to 0. In , an interference packet detection scheme in link layer is proposed, in which a receiving station that detects interference packets reports the existence of another BSS to AP. Then the AP announces channel switching to all stations in its BSS for avoidance of interference. There is no guarantee that the chosen channel is free from interference though.
3 Dynamic transmit power control
Our proposed dynamic TPC (DTPC) scheme is presented in this section. In the DTPC, the stations located at overlapping area are referred as interference prone (IP) stations adopting the notion in . As all the stations continually monitor the ongoing transmissions, combining with the information recorded in the path loss table, a station can determine whether it can start a concurrent transmission. Then all the stations which try to start concurrent transmissions adjust their transmit powers to proper levels and compete for channel access. If one station is successful to access the channel, then since its transmission uses a low power, more stations may become hidden terminals to the transmitter. Thus we propose all the stations to use RTS/CTS procedure where the RTS/CTS frames are exchanged using their maximum powers.
Our DTPC scheme enables the performance enhancement in two aspects: First, when a transmission from an IP station is ongoing, another station which belongs to a neighbor BSS and is not a hidden terminal to the IP station can start a simultaneous transmission after tuning the transmit power. Second, if a hidden terminal starts a transmission in parallel with the IP station, the neighbor AP can adjust its transmit power for timely ACK response, which means a successful transmission.
3.1 NAV reset timer modification
A timer named RESET_NAV is defined in the IEEE 802.11 MAC for NAV update . The stations overhearing an RTS set their NAVs and also set the timer RESET_NAV with a duration of CTS_Time+2SIFS_Time+2Slot_Time. Here, CTS_Time is calculated from the length of the CTS frame and the rate at which the CTS frame is transmitted. After setting the timer, the stations will reset their NAVs if they overhear DATA frame from the RTS sender.
As the RTS/CTS frames are transmitted on the maximum power and data frames are transmitted on a tuned power, some stations which set NAVs according to an RTS frame may not overhear the data frame, and the timer of these stations will expire within the time-out. We modify the NAV reset timer as follows: A new timer D_RESET_NAV is added, and the duration of this timer is the same as the duration field of RTS. Thus if a station overhears an RTS of a station that belongs to its BSS, it sets D_RESET_NAV, otherwise it sets RESET_NAV. It makes sense because in 802.11 WLAN, a station is supposed to receive all the incoming frames and at least decode the MAC header part unless it is in the sleeping mode. Moreover, in infrastructure architecture, the direct transmission is only possible between AP and stations. Thus, a station can check the address fields of a received packet to confirm whether the sender belongs to the domestic BSS. This modification in the NAV reset timer guarantees the domestic stations which set D_RESET_NAV timer will not experience time-out until the ongoing transmission terminates. Our DTPC proposes that after RESET_NAV timer expires, stations enter into the back-off (BO) process directly. The station whose BO counter decreases to 0 will access the channel.
3.2 Path loss recording
In DTPC, each station keeps a table for recording reception power of beacon frame and path loss between itself and the neighbor BSS stations from which it can overhear an RTS/CTS frame, i.e. < n o d e id ,p a t h l o s s ij ,p rev >. AP keeps a table for its own BSS stations and neighbor BSS stations located at overlapping area. When a station overhears an RTS/CTS frame, it updates the record related to the sender. If there is no record for the sender, it adds a new record into the table.
3.3 Tuning transmit power
3.4 Transmissions in a non-hidden terminal environment
3.5 Transmissions in a hidden terminal environment
4 Performance analysis
In order to analyze the performance of the proposed scheme compared to 802.11 MAC, we develop an analytical model using the discrete-time Markov chain in this section.
4.1 Markov chain model
While an ongoing transmission in a BSS prevents transmissions in a neighbor OBSS in the legacy 802.11 MAC, in our proposed scheme, however, the OBSSs are allowed to transmit simultaneously. Thus, in order to compare the channel utilization of the proposed scheme (DTPC) and the legacy MAC, we make an assumption that the co-channel is divided into two sub channels, and each BSS may occupy one of them. Adopting slotted time, in order to make the model Markovian, we suppose that the packet lengths which are integer multiples of slot durations are independent and geometrically distributed with parameter q (i.e., packet duration has a mean of 1/q slots) . Also we assume that devices always have packets to send to AP in each time slot, and each device attempts to transmit with probability p. In addition, it is assumed that there are no hidden or exposed terminals in domestic BSS.
where S n is the number of new transmissions successfully started at time n, and T n is the number of terminations at time n. Note that S n =1 if a new transmission starts successfully in time slot n and S n =0, otherwise. If X n =2, which means that both BSSs are processing transmissions, then S n =0 with probability 1. The number of terminations T n at time n ranges from 0 to X n . If X n =0, then T n =0 with probability 1.
4.2 Capacity analysis
Number of sub channels
Channel bit rate
4.3 Additional issues
In the Markov chain model, we have assumed that there are no hidden or exposed stations in the domestic BSS and a transmission is completed successfully. In a practical network, however, the hidden and/or exposed stations may exist and they will introduce collisions. In order to make the analysis more accurate, our future works will include the study on how to model the probability that a transmission completed successfully in a time slot.
Also, we have analyzed two overlapping BSSs. Modeling the performance of multiple OBSSs, however, becomes more challenging, as the transition probabilities in both legacy scheme and proposed solution are dependent on the network topologies. Especially, based on the location of a transmitting station, the possibility whether a neighbor BSS can process a concurrent transmission in a time slot becomes more complex. We plan to investigate this issue in our future work.
Path loss exponent
Maximum transmit power
Normal transmit power
Figure 8b presents the retransmission attempts versus the network size. We find the retransmission attempts in our proposed scheme are lower than that of the legacy one. It can be achieved by the fact that our proposed scheme enables RTS/CTS frames to be transmitted on a maximum power and it prevents hidden and exposed terminals occurring in a wide range. In addition, unlike the legacy scheme, the neighbor AP can adjust its power to a proper level such that it can respond with immediate ACKs.
In this paper, we have presented a dynamic transmit power control scheme, namely DTPC, for enhancing the performance of OBSSs. Stations can dynamically adjust their transmit powers using the proposed DTPC, which enables the overlapping BSSs to transmit simultaneously and to enhance the spatial reuse. We have developed a Markov chain model in order to analyze the performance of DTPC, and the simulation results prove that the analytical model is properly built. Both analytic and simulation results show that the proposed DTPC significantly improves the performance of OBSSs. As the future work, we plan to investigate the performance of DTPC operating in multiple OBSSs rather than two OBSSs. Also the hidden and/or exposed terminals existed in a BSS and various network topologies should be studied as a future work.
This work was supported by Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education, Science and Technology (2013008855), and in part by the Research Grant of Kwangwoon University in 2013.
- W Li, Y Cui, X Cheng, MA Al-Rodhaan, A Al-Dhelaan, Achieving proportional fairness via AP power control in multi-rate WLANs. IEEE Trans. Wireless Commun. 10(11), 3784–3792 (2011).View ArticleGoogle Scholar
- A Jow, C Schurgers, Borrowed channel relaying: a novel method to improve infrastructure network throughput. EURASIP J. Wireless Commun. Netw. 2009174730 (2010).View ArticleGoogle Scholar
- B Han, L Ji, S Lee, RR Miller, B Bhattacharjee, in IEEE International Conference on Communications. Channel access throttling for overlapping BSS management (Dresden, 2009), pp. 1–6.Google Scholar
- O Oteri, P Xia, F LaSita, R Olesen, in 16th International Symposium on Wireless Personal Multimedia Communications, WPMC. Advanced power control techniques for interference mitigation in dense 802.11 networks (Atlantic, 2013), pp. 1–7.Google Scholar
- X Wang, H Lou, M Ghosh, G Zhang, P Xia, O Oteri, F La Sita, R Olesen, N Shah, in Systems, Applications and Technology Conference, LISAT, 2014 IEEE Long Island. Carrier grade Wi-Fi: air interface requirements and technologies (Farmingdale, 2014), pp. 1–6.Google Scholar
- X Liu, S Seshan, P Steenkiste, in proceedings of the annual conference of ITA. Interference-aware transmission power control for dense wireless networks, (2007), pp. 1–7.Google Scholar
- H Ma, J Zhu, S Roy, SY Shin, Joint transmit power and physical carrier sensing adaptation based on loss differentiation for high density IEEE 802.11 WLAN. Comput. Netw. 52, 1703–1720 (2008).View ArticleMATHGoogle Scholar
- M Cesana, D Maniezzo, P Bergamo, M Gerla, in Vehicular Technology Conference 2003. Interference aware (IA) MAC: an enhancement to IEEE802.11b DCF (Orlando, Florida, USA, 2003), pp. 2799–2803.Google Scholar
- D Qiao, S Choi, A Jain, KG Shin, in Proceedings of the 9th annual international conference on mobile computing and networking. MiSer: an optimal low energy transmission strategy for IEEE 802.11a/h (San Diego, California, USA, 2003), pp. 161–175.Google Scholar
- X Wang, GB Giannakis, CSMA/CCA: a modied CSMA/CA protocol mitigating the fairness problem for IEEE 802.11 DCF. EURASIP J. Wireless Commun. Netw. 2006, 039604 (2006).Google Scholar
- Y Li, X Wang, SA Mujtaba, in Vehicular Technology Conference, 2003. Co-channel interference avoidance algorithm in 802.11 wireless LANs (Orlando, Florida, USA, 2003), pp. 2610–2614.Google Scholar
- Y Fang, D Gu, AB McDonald, J Zhang, in The 14th IEEE Workshop on Local and Metropolitan Area Networks. Two-level carrier sensing in overlapping basic service sets (BSSs) (Island of Crete, Greece, 2005), p. 6.Google Scholar
- T Tandai, K Toshimitsu, T Sakamoto, in Personal, Indoor and Mobile Radio Communications, 2006. Interferential packet detection scheme for a solution to overlapping BSS issues in IEEE 802.11 WLANs (Finland, 2006), pp. 1–5.Google Scholar
- IEEE 802.11-2012. Part 11, Wireless LAN medium access control (MAC) and physical layer (PHY) specifications. IEEE (2012).Google Scholar
- J Mo, H-S So, J Walrand, Comparison of multichannel MAC protocols. IEEE Trans. Mobile Comput. 7(1), 60–65 (2008).Google Scholar
This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited.