An Optimized Deep Learning-Based Smart Parking Mechanism for Smart City Environment

In the recent past, identifying a parking spot at the appropriate time and place has become increasingly crucial around the globe due to the ever-increasing parking demands introduced by vehicles. At this juncture, the process involved in the interested drivers’ search for identifying an unoccupied parking lot is an NP-hard problem with imposed constraints of optimization. Thus, the swarm-intelligent optimized deep learning model would be a better solution for achieving ideal smart parking, such that profit-cum-services management could be addressed more predominantly. In this paper, a deep learning model using Long Short-Term Memory (LSTM) and Least Absolute Shrinkage and Selection Operator (LASSO) (DLLLIDOA) is proposed, incorporating an enhanced Dingo optimization algorithm that focuses on minimizing cruising time and alleviating congestion in parking spaces. This LLIDOA-based deep learning model helped in determining timely and accurate information related to availability and parking occupancy. It specifically adopted the Absolute Shrinkage and Selection Operator (LASSO) into a long short-term memory neural network (LSTM) model to improve the degree of predicting parking lots. It further utilized an improved Dingo Optimization Algorithm (IDOA) to optimize the factors that need to be minimized during the parking process. This IDOA approach balanced local and global search during the optimization process, achieving rapid convergence in optimal time. A dataset is utilized to test the model’s performance, and metrics like RMSE, MSE, R2, MDAE, MSLE, and MAE are used to gauge accuracy. The results of this LLIDOA mechanism confirmed better pricing, minimized delay, and maximized profit during the allocation of a bright parking space.