Received: 24 May 2019 Revised: 27 July 2019 Accepted: 19 August 2019 DOI: 10.1002/ett.3798 SPECIAL ISSUE ARTICLE Efficient scheduling of video camera sensor networks for IoT systems in smart cities Muhammad Naeem 1 Waleed Ejaz 2 Muhammad Iqbal 1 Farkhund Iqbal 3 Alagan Anpalagan 4 Joel J. P. C. Rodrigues 5,6 1 Department of Electrical Engineering, COMSATS University Islamabad, Wah Campus, Wah Cantonment, Pakistan 2 Department of Applied Science & Engineering, Thompson Rivers University, Kamloops, Kamloops, British Columbia, Canada 3 College of Technological Innovation, Zayed University, Abu Dhabi, UAE 4 Department of Electrical and Computer Engineering, Ryerson University, Toronto, Ontario, Canada 5 Federal University of Piauí (UFPI), Teresina - Pi, Brazil 6 Instituto de Telecomunicacções, Portugal Correspondence Waleed Ejaz, Department of Applied Science & Engineering, Thompson Rivers University, Kamloops, V2C 0C8, Canada Email: waleed.ejaz@ieee.org Funding information Cluster Research Project, Grant/Award Number: R16083 and R18055; Zayed University; National Funding from the FCT - Fundação para a Ciência e a Tecnologia, Grant/Award Number: UID/EEA/50008/2019 Project; National Council for Scientific and Technological Development (CNPq), Grant/Award Number: 309335/2017-5 Abstract Video camera sensor networks (VCSN) has numerous applications in smart cities, including vehicular networks, environmental monitoring, and smart houses. Scheduling of video camera sensor networks (VCSN) can reduce the computational complexity, increase energy efficiency, and enhance throughput for the Internet of things (IoT) systems. In this paper, we apply the iterative low-complexity probabilistic evolutionary method for scheduling video cam- eras to maximize throughput in VCSNs for IoT systems. Scheduling of video cameras in VCSNs to maximize throughput is a combinatorial optimization problem whose computational complexity increases exponentially with the increase in the number of video cameras. We propose an iterative probabilistic method named as cross-entropy optimization (CEO), which is an evolution- ary algorithm. The combinatorial optimization problems can be solved using the CEO which is a generalized Monte Carlo technique. The proposed method updates its selected population (video cameras) at each iteration based on the Kullback Leibler (KL) distance/divergence. The KL distance/divergence is min- imized using the probability distribution obtained from the learned from the group of selected samples of better solutions found in the previous iterations. The effectiveness of the CEO is verified in terms of optimality and simplicity through simulations. In addition, the results of the CEO are better than the suboptimal algorithms (ie, best norm-based algorithm, genetic algorithm, and capacity upper-bound–based greedy algorithm) and maximum of 2%-3% devi- ation from the exhaustive search (optimal) with less complexity. The trade-off between CEO and optimal is the computational complexity. 1 INTRODUCTION Video camera sensor networks (VCSNs) are a type of wireless sensor networks (WSNs) in which each sensor is equipped with a video camera. 1 Similar to WSNs, each sensor is capable of sensing, processing, and transmitting video data to the sink node using wireless technologies. VCSNs have been emerging as the key component for Internet-of-Things (IoT) systems in smart cities. 2 For example, Internet of vehicles (IoV) is an environment that consists of IoT enabled vehicles to provide safety, traffic management, service efficiency, etc. Particularly, VCSNs can be used for vehicle searching and tracking in automobiles as well as providing useful signals to the drivers for safety. VCSNs are also important for obstacle detection in IoT era. 3 The efficiency of traffic can be enhanced using VCSNs on the roadside in vehicular networks. Trans Emerging Tel Tech. 2019;e3798. wileyonlinelibrary.com/journal/ett © 2019 John Wiley & Sons, Ltd. 1 of 13 https://doi.org/10.1002/ett.3798