Our research provides a novel hardware/software autonomous car model that can be effectively deployed to carry out intelligent pickup and delivery missions. We develop a model of a Cooperative Autonomous Reactive Taxi System (CARTS) that contributes to solving the unsustainable urban traffic gridlock in large cities. Our model formulation is inspired by general the pickup/delivery problem (GPDP). Given dynamic stochastic variables that include a set of cars, a set of customers, a set of stations (e.g., bus stop), and a set of transport missions, devise autonomous intelligent carts capable of effectively carrying out these missions to satisfy the quality of service requirements and any associated constraints. Our model integrates novel capabilities, such as decentralized architecture, cooperative decision-making, autonomy, and intelligent navigation, as well as a transport requests logistics model. Our deployment prototype is a laboratory experimental set-up using autonomous robots and relevant infrastructure that demonstrate the feasibility of our pickup and delivery model. Analysis of collected data from the simulation and the physical implementation demonstrates the effectiveness of our model.