An Intelligent and Automated Approach for Smart Minimarkets

Abstract

This paper presents the design and implementation of a smart and safe minimarket prototype for deployment in busy smart cities to mitigate the overhead of shopping experience. The prototype allows customers to remotely access and browse the available products at the minimarket using a special smart-phone application. The system can intelligently detect the nearby location of customers and subsequently provide location-dependent services such as allowing orders to be placed using the application, predicting weekly customer expenditures based on artificial-neural-network machine-learning approach, and automatically delivering purchased products using a robotic shopping cart. This proposal is believed to support safe shopping which became a critical issue after COVID-19 pandemic. From a service provider view point, the application allows the provider to remotely manage the minimarket by adding/removing product items, keeping track of shortage in products, and getting revenue information. Empirical results show that the average service time of the minimarket is ≈ 60 seconds per customer. However, an analytical model based on queueing theory was used to analyze the performance of the system when customers arrive according to a Poisson random process and get served according to a general-service-time distribution (M/G/1). The case of batch customer arrivals (M[H] /G/1) was also analyzed, where batch size is also assumed to be random. Various traffic intensities and the effect of variable service times were studied and cross-validated with simulation results. Worst-case scenario shows that under heavy load of 95%, when customers arrive at the minimarket every 63 seconds on average, the average response time for each customer is ≈ 10 minutes.