Risk-Sensitive Rendezvous Algorithm for Heterogeneous Agents in Urban Environments*

Demand for fast and inexpensive parcel deliveries in urban environments has risen considerably in recent years. A framework is envisioned to enforce efficient last-mile delivery in urban environments by leveraging a network of ride-sharing vehicles, where Unmanned Aerial Systems (UASs) drop packages on said vehicles, which then cover the majority of the distance before final aerial delivery. This approach presents many engineering challenges, including the safe rendezvous of both agents: the UAS and the human-operated ground vehicle. This paper introduces a framework to minimize the risk of failure while allowing for the controlled agent's optimal usage. We discuss the downfalls of traditional approaches and formulate a fast, compact planner to drive a UAS to a passive ground vehicle with inexact behavior. To account for uncertainty, we learn driver behavior while leveraging historical data, and a Model Predictive Controller minimizes a risk-enabled cost function. The resulting algorithm is shown to be fast and implementable in real-time in qualitative scenarios.