Smart parking model based on Internet of Things (IoT) and TensorFlow

Abstract

Our paper proposes a smart parking model to reduce a wastage of time by allocating the free parking spaces, ensures security and proposed a recommender system. We have targeted main three segments of our smart parking model. First, we have proposed a cloud based architecture. Through the architecture we shows how to allocate spaces of parking area among the users. We have used Google Cloud IoT core service to generate our cloud architecture. Secondly, we introduced the security portion of a parking system. License plate is an important aspect for those parking places where people reserve parking places for their comfortable journey. We used Deep Learning (DL) based architecture to detect and recognize the license plates of users. TensorFlow framework is used to detect and recognize the license plates for user authentication. Finally, we focused on building a recommender system to recommend a parking place where multiple parking areas are exist in one areas. The Reinforcement Learning (RL) algorithm is used to show how a recommender system can work. We have got 98% accuracy in terms of detecting and recognizing the license plates. Also, our recommender system performs 50% better than the existing algorithms. Overall, our proposed model is able to reduce the wastage of time by 50% which is a very good accuracy and ideal for practical deployment in our real life.