A Lightweight MAC Protocol with Adaptive Transmission
Power Control in WSNs
Chen Caijun
Department of Information Engineering Wuhan University of Technology
Wuhan , PRC
Abstract—This paper proposes an enhanced lightweight
medium access control protocol. This protocol combines schedule-based MAC mechanism and adaptive transmission power control (ADTPC-LMAC). This paper introduces a slotting communication mechanism, through which, ADTPC-LMAC solves the problems that normally occur in WSN such as idle listening, over hearing and hidden terminal problem, etc. Nodes only transmit data in their own timeslots and sleep in other timeslots that they don’t occupy. In addition, ADTPC-LMAC combines this protocol with an adaptive transmission power level without introducing packet delivery ratio. We compare our protocol with traditional carrier sense multiple access (CSMA) mechanism and the result shows our protocol gives a more reliable packet transmission and prolongs the network lifetime. Keywords-ADTPC-LMAC； Adaptive Transmission Power Control； CSMA
I．INTRODUCTION
WSN consists of mass small, low data rate and cheap nodes which are used to perceive or control the real physical environment. Most WSN applications depend on designed nodes to minimize complexity and energy consumption[1]. MAC is the second layer of the OSI model[2]. MAC layer has crucial influence on energy consumption of node, and in turn, energy consumption decides the survival time of node in WSN. The traditional MAC protocols such as ALOHA, CSMA and MACA[3],etc. are all designed on the basis of channel competition methods. During a certain time interval, CSMA monitors channel. If no data sending has been investigated, node regards it as idle channel and will start sending messages. However, CSMA cannot solve the hidden terminal problem. MACA agreement introduced handshake mechanism three times to circumvent hidden terminal problem[4].
Facing with SCMA, excessive energy consumption is caused for MACA must continue monitoring channel; SMAC agreement introduced active\\sleep switching mechanism on the basis of MACA. Nodes schedule their communication process by maintaining NAV with adjacent node information [5]. According to the TDMA mechanism, frame is divided into
Yin Haiming
Department of Information Engineering Wuhan University of Technology
Wuhan , PRC
many timeslots to avoid channel competition problems of high concurrent access. When nodes neither send nor receive data, they move to sleep[6]. TDMA must synchronize by the veracity of the center base station. It may not be acceptable in mobile equivalence WSN. This paper blends adaptive power control mechanism on the basis of LMAC and comes up with ADTPC - LMAC algorithm which is based on TDMA timeslot schedule.
II. THE DESCRIPTION AND REALIZATION OF ADTPC-LMAC A. Timeslot mechanism description
ADTPC-LMAC inherited LMAC frame structure; LMAC adopts the schedule-based or the time division multiple access mechanism. As is shown in fig.1, frame is divided into fixed length of time slot. Time slot is divided into two parts beacon domain and data fields. Each node in network chooses an available slot by gathering adjacent nodes information. In its starting place, node sends beacon packets to broadcast its own state towards adjacent node.
Figure 1. Frame structure for ADTPC-LMAC protocol
The included state information of beacon message is shown in table 1. Timeslot domain identifies current time slot sequence, used for time synchronization, then level domain. Occupied timeslot domain is used to inform the occupation of the neighbor node time slot and available slots information. Occupied timeslot domain has totally 32-bit binary. Each binary 0\\1 shows the state of corresponding time slot information representation, 0 marks the slots available, 1 marks the corresponding slots occupied. The last domain is used to tell adjacent nodes destination node of the next data
978-1-61284-181-6/11/$26.00 ©2011 IEEE 2208
stream.
Table 1. BEACON MASSAGE STRUCTURE FOR ADTPC-LMAC
Field Size(byte) Timeslot 1 Level 1 Occupied timeslot 4 Destination Node 2
following two circumstances: first, when no data need to be
sent after sending a beacon message; second, no data will be sent to this node according to the beacon message received from other adjacent nodes.
The MAC protocol which is based on TDMA requires that nodes synchronize with each other. In ADTCP - LMAC agreement, we adopt hierarchical timeslots to realize synchronization. Node will contrast its own slots timer with the beacon message of adjacent nodes through the number of minimum level.
B. ADTPC-LMAC adaptive power control
Nodes save the neighbor node information by maintaining a nearby table. As is shown in table 2.
Table 2. FIELDS IN NEIGHBORHOOD TABLE APPLIED IN ADTPC-LMAC
Field Size(byte) Node address 2 Node Level 1
Node timeslot
1 Data Sequence Number
1 Node Transmit Level
1 Different from TDMA which uses tuner to allot time slot for cluster nodes, our ADTPC-LMAC agreement inherited
distributed time slot allocation mechanism in LMAC agreement, and made proper simplification. In ADTPC-LMAC, nodes take three steps to gain time slots:
initial, wait and discover, active, As is shown in fig. 2.
Figure 2. State diagram for distributed timeslot assignment in ADTPC-LMAC
When a new node joins the network, it firstly starts initialization routines protected reliably against detective channel to search beacon signal from adjacent nodes. After
receiving beacon messages and completing frame synchronization, nodes change into wait and discover state. Before choosing a time slot, nodes randomly wait a delay time, meanwhile, nodes collect state information of nearby nodes through receiving beacon news from other nearby nodes. The
ending of the state is that nodes gain a free timeslot by (1).
ZOR(x1,x2,...,xN)=x1∨x2∨...∨xN (1) In (1), Xi is time slot field that the ninth node takes. V
stands for operation.
After gaining a time slot successfully, node comes into active state. When a node is activated, it constantly sends beacon message at each timeslot starting place. Node must also monitor beacon messages from other adjacent nodes at other timeslot starting places. Nodes come into sleep in the
In order to minimize the energy consumption, Adaptive
Power Control(APC) mechanism is adopted in the protocol. Scholars have carried out many experiments in the real scenario[7,8,9]. They are all about the sensing device data communication. The result suggests that, with the given sending power and communication distance, the power of signal receipt is changing all the time and the link quality is
unstable. Shan lin, etc. have put forward the ATPC[10] which can
search for the optimal rank of sending power in the current
situation. The experimental data indicate that the power of signal receipt and the condition of temporal-spatial domain are related to each other. In order to find the optimal sending power in the given circumstance, the node will send datagram to its adjacent nodes with different ranks of power. According to the feedback information from the adjacent nodes, APTC algorithm can create a prospective model with the least squares technology. However, this algorithm needs nodes to periodically update their prospective models in order to adapt to the environment topology change.
In the ATPC-LMAC, we simplify ATPC algorithm. This still effectively decreases the energy consumption. In ATPC-LMAC algorithm, each packet sent by node will receive the confirm message from the receiving node. When receiving node is sending confirm message, it binding sends its RSSI. In this way, through (2), the sender can work out the packet’s energy attenuation in this link.
Ploss=Ptr−Pre (2)
ploss stands for the number of energy attenuation; ptr stands for the strength of sending power; prs stands for the strength of accepting power. If prs is bigger than pmaxth or prs smaller than pminth, the sender needs to adjust the sendingpower according to (3).
ppmaxth+pminth
newtr=
2
+Ploss (3)
The maximum and minimum thresholds of received power are crucial to the accuracy of both energy saving and power adjustment. If the range between pmaxth and pminth is too small, the debility of radio signals will likely result in the power jittering; on the contrary, if the range is too large, it will
2209
result in a low accuracy of adjustment of the sending power, and the goal of sending power of optimal rank will be missed in the end.
In the beginning of sending data, the maximal sending power is used by node, then according to the received confirm message node adjusts the sending power. If sending node cannot receive the confirm message, it will resend the packet with the power of 10 times rank. If this power rank has reached the maximum, then it is believed that this sending of this packet is failed.
Receiving node adjusts the power strength of sending confirm message in the same optimization way. If the timeslot at singular time. In this way, the conflict between sending data packets will be avoided. If the duty cycle is 20%, the length of timeslot grows, the waiting time of next timeslot will in turn grow longer. When the data bulks exceed the data capacity at current MAC layer, the data packet throughput rate decreases by 40%.
B.ADTPC-LMAC energy efficiency
receiving node receives a packet with the same serial number as the former one, it will increase the power strength of sending confirm message.
III. ADTPC-LMAC ALGORITHM SIMULATION
The simulation operation system scenario in this paper is Redhat9, using NS-2.34. Pminth value is set as -85dBm, Pmaxth value is -75dBm. Taking CSMA and CSMA-ACK with ACK function as the targets of evaluation we evaluate ADTPC-LMAC algorithm in this paper. In this simulation experiment, we set 9 nodes in the 3*3 square, the peripheral 8 nodes periodically sending 25 byte packets to the central node, as is shown in fig.3.
Figure 3. Nodes’ topology in simulation environment
A. Increase of throughput rate of data packet by ADTPC-LMAC
Figure 4. Delivery ratio for different packet rate
The result of fig. 4 indicates that ADTPC-LMAC can obtain better packet data throughput rate. ADPTC-LMAC protocol with 50% of duty cycle can realize optimal data packet delivery ratio. This is because only one node can have
Figure 5. Average energy consume in a node and average energy per bit for
different packet rate
It is clear to see from fig. 5 that, in ADTPC-LMAC, the average energy consume both in a node and per bit packet have greatly reduced and showed a better performance. The APC algorithm of ADTPC-LMAC can greatly reduce the energy consumption and increase the energy efficiency by using the optimal rank of sending power to send data.
IV. CONCLUSION AND FUTURE WORK
The ADTPC-LMAC Protocol proposed in this article has introduced time slots mechanism and APT, both can avoid the competition in Channel access, and raise the network’s data throughput rate while reduce the energy consumption by choosing the optimal sending power to send packet in accordance to environmental factor adaptively. The simulation indicates that, different duty cycles have different superiorities in network traffic. In heavy network traffic, ADTPC-LMAC with a low duty cycle does not perform well in data throughput rate, although it has lower energy consumption. What’s more, in light network traffic, ADTPC-LMAC with high duty cycle will lead to unnecessary energy consumption. Moreover, the nodes in this paper are arranged based on WPAN network without the consideration of node’s mobility which can bring the change of topology structure. As a result, our next work will focus on the studying of node’s topology transformation and adaptively adjusting the duty cycle depending on the detection of network traffic.
2210
REFERENCES
[1]Sun Limin,Wireless sensor networks.Beijing:TsingHua University Press,2005.(in Chinese)
[2]Wang Jinlong, Ad hoc mobile wireless network. Beijing: National Defence Industry Press, 2004. (in Chinese)
[3] Zhang Dazong,Yang Tao, Wei Dongmei. Overview of low-power design in wireless sensor networks. Transducer and Microsystem Technologies,2006, (5):141-144(in Chinese)
[4] L V. Hoesel, P. Havinga, “Collision-free Time Slot Reuse in Multi-hop Wireless Sensor Networks”, Proc. Int. Conf. on Intelligent Sensor, Sensor Networks and Information Processing , Dec. 2005, pp. 101-107.
[5] L. F. W. van Hoesel and P. J. M. Havinga, “Design Aspects of An Energy-Efficient Lightweight Medium Access Control Protocol for Wireless Sensor Networks”, July 17, 2006.
[6] J. Zhao and R. Govindan, “Understanding Packet Delivery Performance in Dense Wireless Sensor Networks”, ACM SenSys., Nov. 2003.
[7] A. Woo, T. Tong and D. Culler, “Taming the Underlying Challenges of Reliable Multihop Routing in Sensor Networks”, ACM SenSys., Nov. 2003. [8] G. Zhou, T. He, S. Krishnamurthy and J. A. Stankovic, “Impact of Radio Irregularity on Wireless Sensor Networks”, ACM MobiSys., June 2004, pp. 125-138.
[9] A.Cerpa, J.L. Wong, L. Kuang, M. Potkonjak and D. Estrin, “Statistical Model of Lossy Link in Wireless Sensor Networks”, ACM/IEEE IPSN, Apr. 2005.
[10] S. Lin, J. Zhang, G. Zhou, L. Gu, T. He and J. A. Stankovic, “ATPC: Adaptive Transmission Power Control for Wireless Sensor Networks”, SenSys., Nov. 2006.
2211
因篇幅问题不能全部显示，请点此查看更多更全内容