ITS Berkeley

In this article, we demonstrate a zero-shot transfer of an autonomous driving policy from simulation to University of Delaware's scaled smart city with adversarial multi-agent reinforcement learning, in which an adversary attempts to decrease the net reward by perturbing both the inputs and outputs of the autonomous vehicles during training. We train the autonomous vehicles to coordinate with each other while crossing a roundabout in the presence of an adversary in simulation. The adversarial policy successfully reproduces the simulated behavior and incidentally outperforms, in terms of...

A wide range of reinforcement learning (RL) problems --- including robustness, transfer learning, unsupervised RL, and emergent complexity --- require specifying a distribution of tasks or environments in which a policy will be trained. However, creating a useful distribution of environments is error prone, and takes a significant amount of developer time and effort. We propose Unsupervised Environment Design (UED) as an alternative paradigm, where developers provide environments with unknown parameters, and these parameters are used to automatically produce a distribution over valid,...

Curb space has been traditionally designed for private vehicle parking, public transit, and passenger and commercial loading. However, in recent years, a growing number of newservices and activities have increased the demand for limited curb space, including passenger pick-up and drop-off; last-mile delivery (e.g., courier network services, personal delivery devices); electric vehicle (EV) charging; micromobility parking and use (e.g., personally owned and shared bikes and scooters); and carsharing services. The curb serves a variety of functions such as vehicle and device storage (...

Level of Traffic Stress (LTS) metrics are widely used to examine how bicyclists may perceive stress along urban streets and identify opportunities for infrastructure improvements. The intuitiveness of the original method, which condensed 18 input variables into four levels, has made LTS very popular among practitioners. Nonetheless, it can be challenging to collect all required inputs. In response, numerous alternative methods have been developed with fewer or different inputs drawn from more general sources, such as OpenStreetMap (OSM) or GIS datasets from local agencies. These...