College of Engineering, Pune

About: College of Engineering, Pune is a based out in . It is known for research contribution in the topics: Computer science & Sliding mode control. The organization has 4264 authors who have published 3492 publications receiving 19371 citations. The organization is also known as: COEP.

Hybrid Cluster Head Election for WSN Based on Firefly and Harmony Search Algorithms

Abstract: Design of energy efficient routing protocols for Wireless Sensor Network (WSN) is a great challenge for researchers. Recently, WSNs have gained lot of popularity and many energy efficient routing solutions are proposed. Most of the existing routing protocols focus on cluster head election and ignoring other important aspects of routing such cluster formation, data aggregation, etc. This research article presents a hybrid cluster head election for WSN based on firefly and harmony search algorithms. The contributions of the proposed protocols are (1) two level cluster head election strategy. In the first stage harmony search algorithm is used to determine initial set of energy efficient cluster head nodes that are sufficiently separated from on another by certain optimal distance. Then tentatively elected cluster head nodes are refined by firefly algorithm by considering the parameters such as node density, cluster compactness and energy to be consumed. Sometimes nature inspired optimization techniques may end up in early convergence and to avoid such problems, cluster head election scheme is divided at two levels. (2) a refined cluster formation strategy is designed where a normal node has privilege of joining to cluster head node either based on distance based metric or based on residual energy of cluster heads. This process of cluster formation helps in reduced energy consumption. The presented protocol is compared with some of the well-known clustering protocols such as LEACH, LEACH-C, EOICHD, and simple firefly based routing protocol based on the evaluation metrics such as number of alive nodes, energy consumption of network, number of packets received by Base Station, First Node Dead, Half Node Dead and Last Node Dead. Implementation is carried out using Network Simulator (NS 2.34) and results show that proposed hybrid cluster head election scheme outperforms the mentioned routing protocols.

System identification black box approach for modeling performance of PEM fuel cell

Abstract: A polymer electrolyte membrane (PEM) fuel cell is very useful for distributed generation and for portable users like electric vehicles because it is very efficient, emission free and operated at low temperature. However, it is not so easy to find direct experimental estimates of the actual performance of PEM fuel cell through various phenomena and operating conditions like chemical reactions, fuel pressure, working temperature and fuel humidity. Mathematical modeling, therefore, plays an important role in understanding operational performance of PEM fuel cell. The actual operating performance of PEM fuel cell depends on a number of parameters, therefore developing an accurate model that includes its dynamic behavior is most important. In this paper, first time, the system identification black box approach is used to develop a number of simple but more realistic mathematical model structures for a PEM fuel cell. The performance of each model structure is compared with the data from a 25 cm2 active area practical PEM fuel cell for result validation. The presented models can be used to predict polarization behavior of the PEM fuel cell under different loading conditions.

Design of a discrete-time sliding mode controller for a magnetic levitation system using multirate output feedback

Abstract: The advent of highly efficient and superior digital computers and microprocessors in controller implementation has led to a significant interest in the field of discrete time controller design Design of a discrete time sliding mode controller for a magnetic levitation system based on a multirate output feedback strategy is presented The control objective is to precisely stabilize the position of a ferromagnetic ball above the ground by levitating it against the force of gravity using an electromagnet For this application, we designed a discrete time sliding mode controller using the equivalent discrete time reaching law Furthermore, we developed a new formulation on discrete time sliding mode control using multirate output feedback for this application, guaranteeing a system response close to that obtained through implementation of a continuous time sliding mode control Simulation results show that for a magnetic levitation system the later, multirate output feedback based discrete time sliding mode controller is more robust than the former, discrete time sliding mode controller Also, the phenomenon of chattering present in practical continuous time sliding mode controllers is significantly reduced in the discrete time counterparts

Performance enhancement of cooperative learning algorithms by improved decision making for context based application

Abstract: Cooperation in learning (CL) can be understood in a multiagent system. In this the agents are capable of learning from both their own experiments and other agents' knowledge and expertise. Implementation of CL is a complicated task in the real world. In distributed systems several agents cooperate to achieve a common goal or accomplish a shared task. In particular, if there are different people or organizations with different goals and information, then a multiagent system (MAS) is needed to handle their interactions. In this paper, various issues related with cooperative machine learning are studied and implemented. A new set of improved cooperative learning algorithms is proposed in the paper. Expertness measuring criteria which were used in earlier work is further enhanced in proposed method. Six methods for measuring the agents' expertness are used i.e. Normal (Nrm), Absolute (Abs), Positive (P), Negative (N), Certainty (Cer) and Entropy (Ent). The novelty of this approach lies in the implementation of Weighted Strategy Sharing with expertness measuring criteria by means of Q-learning, Sarsa learning, Q(λ) and Sarsa(λ) learning algorithms. The paper shows implementation results and performance comparison of all these algorithms.

Data Mining Technology with Fuzzy Logic, Neural Networks and Machine Learning for Agriculture

Abstract: Farmers countenance failure as the crop cultivation decisions by farmers always depend on current market price as the production sustainability processes are not taken into consideration So there should be some platform which guides the farmer for taking correct decision depending on their need, environment, and changing seasons The system proposes Marathi calendar using nakshatras which guide farmer for crop cultivation decision It aims to create methodologies to strengthen the farmers’ economic conditions by providing informed decisions The methodology used for the system specially uses data mining to generate expert decision along with the fuzzy logic, machine learning to give decisions appropriately to farmer for cultivation of expected crops