A Survey on Centralised and Distributed Clustering Routing Algorithms for WSNs Morteza M. Zanjireh and Hadi Larijani School of Engineering and Built Environment Glasgow Caledonian University Glasgow, UK {Morteza.Zanjireh, H.Larijani}@gcu.ac.uk Abstract—Wireless sensor networks have a wide range of applications because they can be adapted for various environments. They can operate independently in harsh places where a human presence is risky or even impossible. Since their life time is dependent on their batteries and replacing or recharging their batteries is impossible in rough places, it is necessary to find energy efficient routing protocols for them. In this paper, a number of well-know energy efficient routing algorithms for WSNs have been classified and presented based on their attributes. Index Terms—Wireless Sensor Networks, Clustering, Energy Efficiency, Centralised algorithms, Distributed algorithms, Uni- form distribution. I. I NTRODUCTION Wireless Sensor Networks (WSNs) have experienced no- ticeable growth over the past decade due to their adaptability and rapid advances in technology. In order to monitor an environment, a large number of sensors, numbering hundreds or even thousands of nodes, can be deployed and cooperate with one another in data collection. Data sensed by each sensor can be collected by the network and transmitted to the sink. Due to their adjustable and in-expensive attributes, they have a wide range of applications such as environmental monitoring, chemical detection, health-care services, emergency response, surveillance missions, vehicular movements, volcanic earth- quake timing, and weather forecasting [1]–[3]. Distributed sensors in WSNs can be organised in an ad-hoc manner and can be used in harsh places where a human presence is hopeless or risky [4]. Energy efficiency is essential for this sensor network’s effectiveness because recharging or replacing their batteries is hard in harsh places. Over the past decade, a large number of studies have been conducted in order to propose energy efficient algorithms for WSNs [5]–[8]. The main factor of energy consumption for WSNs is their communications [9]. Clustering sensors into smaller groups has an important role in conserving the energy of network sensors, and therefore in increasing network lifetimes, especially when we consider that neighbouring sensors might sense similar events. In clustering, a high number of sensors are grouped into small clusters, each cluster has a Cluster Head (CH) and other sensors are Cluster Members (CMs). Sensors need to transmit their sensed data to their CH using low-power short-distance transmitting. CHs aggregate and send the collected data to a sink using high- power long-distance transmitting. Thus CHs’ energy might run out earlier than others [10]. The proper number and scale of clusters is essential for clustering effectiveness, otherwise the network can not benefit from clustering advantages. A large number of clusters leads to a large number of CHs in the network. Consequently, a large number of CHs have to communicate with the sink using long-distance transmission. On the other hand, a small number of clusters leads to clusters with big diameters, and in each cluster a large amount of energy is consumed to send data from CMs to CHs. Consequently, a trade-off should be made between these two opposite circumstances [11]. Distributing CHs uniformly throughout the network is an- other influencing factor for clustering effectiveness. Arranging CHs too close or too far from each other might lead once more to clusters that are too small or big. In both cases, the result is inefficient energy protocols [6], [12]. In this paper, we describe a number of well-known cluster- ing energy efficient routing algorithms for WSNs which have been proposed by researchers so far. II. FEATURES OF WSNS WSNs have different features which can have an impact on designing efficient routing algorithms. These features can be used to compare different protocols and algorithms. A number of WSN features are as follows [1]: • WSNs vs Ad-hoc Networks: WSNs are a kind of ad- hoc network. Ad-hoc networks are infrastructure-less and cooperation-based networks which means that the net- work topologies must be decided by the network sensors themselves. Nevertheless, sensor nodes are more limited in capabilities and are deployed more densely than other ad-hoc networks [13]. • Sensor Positions are not Engineered or Predefined: This feature allows random separation of sensors in the environment. Thus, all sensors, their algorithms, and their protocols need to have self-organising capabilities. • Data Gathering: The sensor nodes have processing ability. They therefore can carry out simple data gathering procedures and transmit only the required and processed data to the next sensor.