Application of Mm-Wave Radar Detection Technology and Artificial Intelligence Learning to Implement the Real-Time and Predictive Design of Parking Spaces

Abstract

In modern cities, the number of vehicles is increasing day by day, which makes it difficult to find parking spaces during travel and takes a lot of time to drive at a slow speed. This not only causes local traffic chaos but also interferes with the driving efficiency of the entire roads. This phenomenon is a common problem faced by modern metropolitan areas. To reduce the time spent in finding parking spaces, fuel consumption, and carbon dioxide emissions, and to effectively improve overall driving efficiency, the systemization of urban roadside or off-street parking has become a key issue that needs attention and demonstration in so-called “smart” cities. In this paper, we have designed and applied an mm-Wave Radar to detect the presence or absence of available parking spaces, and the parking space data can be quickly uploaded to the cloud. Therefore, the status of the parking space can be updated in real-time. Based on this parking space status data, we build a Long Short-Term Memory (LSTM) model to conduct deep learning. The input data of this model includes parking space location, parking space status, and parking time factors. Using mm-Wave radar experiment with the trained LSTM output provides parking space status prediction which enables drivers to obtain parking space status before they arrive at the destination. This can save half of the drivers' time wasted in finding parking spaces and head to the smart cities.