A new integrated collision risk assessment methodology for autonomous vehicles

Abstract

Real-time risk assessment of autonomous driving at tactical and operational levels is extremely challenging since both contextual and circumferential factors should concurrently be considered. Recent methods have started to simultaneously treat the context of the traffic environment along with vehicle dynamics. In particular, interaction-aware motion models that take inter-vehicle dependencies into account by utilizing the Bayesian interference are employed to mutually control multiple factors. However, communications between vehicles are often assumed and the developed models are required many parameters to be tuned. Consequently, they are computationally very demanding. Even in the cases where these desiderata are fulfilled, current approaches cannot cope with a large volume of sequential data from organically changing traffic scenarios, especially in highly complex operational environments such as dense urban areas with heterogeneous road users. To overcome these limitations, this paper develops a new risk assessment methodology that integrates a network-level collision estimate with a vehicle-based risk estimate in real-time under the joint framework of interaction-aware motion models and Dynamic Bayesian Networks (DBN). Following the formulation and explanation of the required functions, machine learning classifiers were utilized for the real-time network-level collision prediction and the results were then incorporated into the integrated DBN model for predicting collision probabilities in real-time. Results indicated an enhancement of the interaction-aware model by up to 9%, when traffic conditions are deemed as collision-prone. Hence, it was concluded that a wellcalibrated collision prediction classifier provides a crucial hint for better risk perception by autonomous vehicles.

Simulating a transition to autonomous mobility

“…The results of their study showed some disbenefit of adopting CAVs in low penetrations (like 20%), whereas a high penetration rate (like 90%) has significantly reduced the conflicts at junctions, which substantially help increase safety. Katrakazas et al (2019) further explore these issues using a SSAM approach and find that this approach could be used in AVs to find dangerous road users and produce manoeuvres for other road users to avoid possible accidents. Table 1 reviews studies with a focus on the simulation of AVs and CAVs and the acquired methods.…”

A review on key challenges in intelligent vehicles: Safety and driver‐oriented features

“…Katrakazas et al. [38] constructed a joint risk assessment framework based on interactive perception motion model and dynamic Bayesian network (DBN).…”

Predicting real-time traffic conflicts using deep learning

“…However, at present, the safety aspect of these automated driving has primarily been tackled with data and information gathered from in-vehicle sensors and equipment. Given the fact that a sensor failure is inevitable even with redundant measurements in a chain of operations, a holistic approach is required to guarantee a collision-free path planning of automated driving (Katrakazas et al, 2017). For example, a conflict detection technique may provide a critical hint of any collision-risk arising from the contextual and circumstantial factors such as collision hot-spots identified through historical collision data, unsafe traffic dynamics in real-time, infrastructure deficiencies, complex road layout, and other imminent threats from surrounding objects.…”