AN ULTRA-LOW POWER AND DISTRIBUTED ACCESS PROTOCOL FOR BROADBAND WIRELESS SENSOR NETWORKS Lizhi Charlie Zhong, Rahul Shah, Chunlong Guo, Jan Rabaey Berkeley Wireless Research Center Department of EECS University of California at Berkeley Abstract - An ultra-low power access control scheme has been proposed for broadband wireless sensor networks. This is a distributed method, which does not require the existence of a central base station, eliminating the possibility of a single point failure. This access protocol combines the best of simple carrier sense multiple access (CSMA) and spread spectrum techniques. It trades the bandwidth in broadband applications for higher power efficiency and throughput. Compared with other access schemes, it consumes far less power. This access protocol does not require a dedicated control channel, or synchronization, whether global or local. Additionally, it has very low delay and does not have the problem of coordinating broadcast and scheduled unicasts. For sensor network s, the inherent redundancy can be exploited to further improve power efficiency. It achieves much higher power efficiency than much more complicated algorithms seeking to reduce the redundancy. I. SENSOR NETWORKS Sensor network has attracted quite a n attention recently in both academia and industry. The idea is simple enough. Tens or hundreds of sensors are spread out. Each of them can measure the temperature, light intensity, humidity, noise level, airflow speed, etc. in its locality. But the beauty of the sensor network is that all these nodes can be self -organized into a network, over which they can do cooperative processing to accomplish tasks that they cannot do individually. The idea of sensor network applies directly to office building environm ent, where sensors can work together to create a microclimate based on personal settings. Similar things can be done inside a car, airplane, train, bus, or in any other public transportation and in any public places. But the applications for sensor network s go far beyond these. They can be used in smart home, in the battlefield, in scientific research, in home nurse care and interactive museums. Sensor networks can also be used together with Internet to create a net surfing experience impossible before. Loc al sensor networks can recreate the environment where a web page tries to bring one to, be it smells, strong wind or cold weather. Similarly, a game designer can put in all the details in a compute game, which allows sensors to recreate the scene in a player’s local environment. Most importantly, the “sensor” in a sensor network does not have to be limited to a mechanic device that measures certain parameters. It can be anything that has certain knowledge of its local environment, for example, an individual person can be a “sensor”. Therefore, almost everything can be thought of as a form of sensor networks, so design used in a traditional sensor network can be ported to it easily. Despite the diversity of the applications in sensor networks, the following requirements are true for all of them: low power, low cost, wireless and ad -hoc. Among these requirements, low power is the most important. It would be a maintenance nightmare having to replace the batteries of the sensors every day or even every month. In some applications, the sensors are buried inside a wall so it is impossible to take them out to replace their batteries. In yet some other applications, the sensors are used just once and live only as long as their batteries can support. So it is very imp ortant to make the sensors self powered or have battery life of several years. In either case, the design of the entire system has to be optimized for low power consumption. II. MEDIA ACCESS CONTROL PROTOCOLS There are many different Media Access Contro l (MAC) protocols in the literature. They can be classified into different categories based on different principles. Some are centralized, with the base station or group leader doing the access control; some are distributed. Some use a single channel; some use multiple channels. Some use various versions of random access; some use reservation and scheduling. They are also optimized for different things: power, delay, throughput, fairness, Quality of Service (Qos) or support for multiple services. The MAC for sensor networks has to be distributed and optimized for power. A distributed method using multiple channels and random access is the way to go. Distributed methods do not require a central control point, avoiding the problem of a single point of failure . Global synchronization is typically not required in distributed methods. Using multiple channels reduces collisions and retransmissions. It also reduces delay and increases throughput. Random access does not require a node to have any knowledge of the ne twork, therefore synchronization is not needed and there is no overhead in exchanging schedules and reservation information. Protocols using reservation and scheduling have better channel utilization, but they only make sense for periodic events, therefore they are not suited for mobile or any dynamic environment.