An official website of the United States government
ObjectiveOur objectives were to assess the efficacy of active inference models for capturing driver takeovers from automated vehicles and to evaluate the links between model parameters and self-reported cognitive fatigue, trust, and situation awareness. BackgroundControl transitions between human drivers and automation pose a substantial safety and performance risk. Models of driver behavior that predict these transitions from data are a critical tool for designing safer, human-centered, systems but current models do not sufficiently account for human factors. Active inference theory is a promising approach to integrate human factors because of its grounding in cognition and translation to a quantitative modeling framework. MethodWe used data from a driving simulation to develop an active inference model of takeover performance. After validating the model’s predictions, we used Bayesian regression with a spike and slab prior to assess substantial correlations between model parameters and self-reported trust, situation awareness, fatigue, and demographic factors. ResultsThe model accurately captured driving takeover times. The regression results showed that increases in cognitive fatigue were associated with increased uncertainty about the need to takeover, attributable to mapping observations to environmental states. Higher situation awareness was correlated with a more precise understanding of the environment and state transitions. Higher trust was associated with increased variance in environmental conditions associated with environmental states. ConclusionThe results align with prior theory on trust and active inference and provide a critical connection between complex driver states and interpretable model parameters. ApplicationThe active inference framework can be used in the testing and validation of automated vehicle technology to calibrate design parameters to ensure safety.
more »
« less
World Model Learning from Demonstrations with Active Inference: Application to Driving Behavior
Active inference proposes a unifying principle for perception and action as jointly minimizing the free energy of an agent’s internal world model. In the active inference literature, world models are typically pre-specified or learned through interacting with an environment. This paper explores the possibility of learning world models of active inference agents from recorded demonstrations, with an application to human driving behavior modeling. The results show that the presented method can create models that generate human-like driving behavior but the approach is sensitive to input features.
more »
« less
- Award ID(s):
- 2048395
- PAR ID:
- 10440154
- Date Published:
- Journal Name:
- International Workshop on Active Inference IWAI 2022
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Explaining the results of Machine learning algorithms is crucial given the rapid growth and potential applicability of these methods in critical domains including healthcare, defense, autonomous driving, etc. In this paper, we address this problem in the context of Markov Logic Networks (MLNs) which are highly expressive statistical relational models that combine first-order logic with probabilistic graphical models. MLNs in general are known to be interpretable models, i.e., MLNs can be understood more easily by humans as compared to models learned by approaches such as deep learning. However, at the same time, it is not straightforward to obtain human-understandable explanations specific to an observed inference result (e.g. marginal probability estimate). This is because, the MLN provides a lifted interpretation, one that generalizes to all possible worlds/instantiations, which are not query/evidence specific. In this paper, we extract grounded-explanations, i.e., explanations defined w.r.t specific inference queries and observed evidence. We extract these explanations from importance weights defined over the MLN formulas that encode the contribution of formulas towards the final inference results. We validate our approach in real world problems related to analyzing reviews from Yelp, and show through user-studies that our explanations are richer than state-of-the-art non-relational explainers such as LIME .more » « less
-
Many papers have addressed the problem of learning the behavior (i.e., the local interaction function at each node) of a networked system through active queries, assuming that the network topology is known. We address the problem of inferring both the network topology and the behavior of such a system through active queries. Our results are for systems where the state of each node is from {0, 1} and the local functions are Boolean. We present inference algorithms under both batch and adaptive query models for dynamical systems with symmetric local functions. These algorithms show that the structure and behavior of such dynamical systems can be learnt using only a polynomial number of queries. Further, we establish a lower bound on the number of queries needed to learn such dynamical systems. We also present experimental results obtained by running our algorithms on synthetic and real-world networks.more » « less
-
When faced with accomplishing a task, human experts exhibit intentional behavior. Their unique intents shape their plans and decisions, resulting in experts demonstrating diverse behaviors to accomplish the same task. Due to the uncertainties encountered in the real world and their bounded rationality, experts sometimes adjust their intents, which in turn influences their behaviors during task execution. This paper introduces IDIL, a novel imitation learning algorithm to mimic these diverse intent-driven behaviors of experts. Iteratively, our approach estimates expert intent from heterogeneous demonstrations and then uses it to learn an intent-aware model of their behavior. Unlike contemporary approaches, IDIL is capable of addressing sequential tasks with high-dimensional state representations, while sidestepping the complexities and drawbacks associated with adversarial training (a mainstay of related techniques). Our empirical results suggest that the models generated by IDIL either match or surpass those produced by recent imitation learning benchmarks in metrics of task performance. Moreover, as it creates a generative model, IDIL demonstrates superior performance in intent inference metrics, crucial for human-agent interactions, and aptly captures a broad spectrum of expert behaviors.more » « less
-
Abstract This paper investigates temporal correlations in human driving behavior using real-world driving to improve speed forecasting accuracy. These correlations can point to a measurement weighting function with two parameters: a forgetting factor for past speed measurements that the vehicle itself drove with, and a discount factor for the speeds of vehicles ahead based on information from vehicle-to-vehicle communication. The developed weighting approach is applied to a vehicle speed predictor using polynomial regression, a prediction method well-known in the literature. The performance of the developed approach is then assessed in both real-world and simulated traffic scenarios for accuracy and robustness. The new weighting method is applied to an ecological adaptive cruise control system, and its influence is analyzed on the prediction accuracy and the performance of the ecological adaptive cruise control in an electric vehicle powertrain model. The results show that the new prediction method improves energy saving from the eco-driving by up to 4.7% compared to a baseline least-square-based polynomial regression. This is a 10% improvement over the constant speed/acceleration model, a conventional speed predictor.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- The DOI is not currently available.
