To make the massive deployment of automated vehicles possible in complex urban environments, it is essential to provide them with the ability of making safe and useful decisions. To that end, it is necessary to improve their capability to infer the intentions of the surrounding vehicles and their associated collision risk for the ego-vehicle in complex driving scenes. This work shows the implementation and validation in simulation of a probabilistic approach to estimate the risk of driving under uncertain conditions, combining (i) intention estimations and (ii) the expected behaviour of vehicles according to the topology and the subsequent traffic rules of the considered driving scenario. Promising results in terms of success rate and prediction horizon have been obtained testing the proposed approach in driving situations where lateral intention estimation is relevant, namely in multi-lane roundabouts and highways.

Efficient testing and validation of software components for highly automate vehicles is one of the key challenges to be solved for their massive deployment. The number of driving situation and environment variables makes validation almost intractable with real vehicles in open roads, and the testing reproducibility can only be achieved via simulation. This manuscript presents a framework and preliminary results for motion prediction of vehicles in a simulation environment that is being currently developed by the AUTOPIA Program.

No abstract

Autonomous vehicles (AVs) promise to bring many benefits to society, such as safety, an increase of accessibility and life quality, among others. Unlike humans, they do not get tired and, supposedly, do not fail. However, there might be cases where, due to limit visibility, occlusions or even a sensor failure, the system might not be capable to detect one or more obstacles along the vehicle's path early enough to avoid a crash. Although these situations might be rare if one considers a single vehicle, if predictions are correct, these AVs are to be adopted in large quantities in the near future, making even rare situations more commonplace. AVs will have to deal with these forced-choice situations in the best possible way. This paper presents a review of the ethical discussion regarding the matter of AVs and an analysis of a questionnaire implemented by the authors. Our results show evidence for several types of contradictory choices made by the subjects, which suggest the moral choices do not necessarily follow strict logical reasoning.

The behavior of traffic participants is full of uncertainties in the real world. It depends on their intentions, the road layout, and the interaction between them. Probabilistic intention and motion predictions are unavoidable to safely navigate in complex scenarios. In this work, we propose a framework to compute the motion prediction of the surrounding vehicles taking into account all possible routes obtained from a given map. To that end, a Dynamic Bayesian Network is used to model the problem and a particle filter is applied to infer the probability of being on a specific route and the intention to change lanes. Our framework, based on Markov chains, is generic and can handle various road layouts and any number of vehicles. The framework is evaluated in two scenarios: a two-lane highway and a three-lane merging highway. Finally, the influence of a set of lane-changing methods is evaluated on the predictions of the vehicles present on the scene.