Identifying Driver Interactions via Conditional Behavior Prediction

Tolstaya, Ekaterina; Mahjourian, Reza; Downey, Carlton; Vadarajan, Balakrishnan; Sapp, Benjamin; Anguelov, Dragomir

doi:10.1109/icra48506.2021.9561967

Cited by 61 publications

(23 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Nearly all work assumes an independent, per-agent output space, in which agent interactions cannot be explicitly captured. A few works are notable in describing joint interactions as output, either in an asymmetric [28,47] or symmetric way [18,35,41].…”

Section: Related Workmentioning

confidence: 99%

MultiPath++: Efficient Information Fusion and Trajectory Aggregation for Behavior Prediction

Varadarajan¹,

Hefny²,

Srivastava³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Predicting the future behavior of road users is one of the most challenging and important problems in autonomous driving. Applying deep learning to this problem requires fusing heterogeneous world state in the form of rich perception signals and map information, and inferring highly multimodal distributions over possible futures. In this paper, we present MultiPath++, a future prediction model that achieves state-of-the-art performance on popular benchmarks. MultiPath++ improves the MultiPath architecture [45] by revisiting many design choices. The first key design difference is a departure from dense image-based encoding of the input world state in favor of a sparse encoding of heterogeneous scene elements: MultiPath++ consumes compact and efficient polylines to describe road features, and raw agent state information directly (e.g., position, velocity, acceleration). We propose a context-aware fusion of these elements and develop a reusable multi-context gating fusion component. Second, we reconsider the choice of pre-defined, static anchors, and develop a way to learn latent anchor embeddings end-to-end in the model. Lastly, we explore ensembling and output aggregation techniques-common in other ML domains-and find effective variants for our probabilistic multimodal output representation. We perform an extensive ablation on these design choices, and show that our proposed model achieves state-of-the-art performance on the Argoverse Motion Forecasting Competition [12] and the Waymo Open Dataset Motion Prediction Challenge [18].

show abstract

Section: Related Workmentioning

confidence: 99%

MultiPath++: Efficient Information Fusion and Trajectory Aggregation for Behavior Prediction

Varadarajan¹,

Hefny²,

Srivastava³

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

show abstract

“…Our framework is complementary to alternative approaches for characterizing interaction, such as the interactivity score [43] (mutual information) and distribution-based KL-divergence. The Interactivity score may miss crucial interaction events: scores can be large when there is high correlation between two trajectories (e.g., one car following another), but small when trajectories are dissimilar (e.g., cars crossing an intersection).…”

Section: Discussionmentioning

confidence: 99%

“…DeCastro et al [11] construct a representation of multi-vehicle interaction outcomes based on latent parameters using a generative model. Tolstaya et al [43] propose an Interactivity score that enables the identification of interesting interactive scenarios for training generative models. Our work is similar in spirit and complementary to this latter line of work.…”

Section: Related Workmentioning

confidence: 99%

Analyzing Multiagent Interactions in Traffic Scenes via Topological Braids

Mavrogiannis¹,

DeCastro²,

Srinivasa³

2021

Preprint

View full text Add to dashboard Cite

We focus on the problem of analyzing multiagent interactions in traffic domains. Understanding the space of behavior of real-world traffic may offer significant advantages for algorithmic design, data-driven methodologies, and benchmarking. However, the high dimensionality of the space and the stochasticity of human behavior may hinder the identification of important interaction patterns. Our key insight is that traffic environments feature significant geometric and temporal structure, leading to highly organized collective behaviors, often drawn from a small set of dominant modes. In this work, we propose a representation based on the formalism of topological braids that can summarize arbitrarily complex multiagent behavior into a compact object of dual geometric and symbolic nature, capturing critical events of interaction. This representation allows us to formally enumerate the space of outcomes in a traffic scene and characterize their complexity. We illustrate the value of the proposed representation in summarizing critical aspects of real-world traffic behavior through a case study on recent driving datasets. We show that despite the density of real-world traffic, observed behavior tends to follow highly organized patterns of low interaction. Our framework may be a valuable tool for evaluating the richness of driving datasets, but also for synthetically designing balanced training datasets or benchmarks.

show abstract

“…It is because passive prediction models ignore the fact that the autonomous agent's future actions can influence other agents' behavior. To this end, researchers started to investigate a more coherent interactive prediction and planning framework which relies on predicting the surrounding agents' future trajectories conditioned on the ego agent's future actions [4], [5], [6], [7], [8], [9], [10], [11]. Under such frameworks, the autonomous agents can reason over potential actions while considering their influence on surrounding agents.…”

Section: Introductionmentioning

confidence: 99%

“…More interestingly, some existing works formulated an alternative prediction task to evaluate the prediction module in a self-contained way [9], [8], [11]. We follow [11] to refer to this task as conditional behavior prediction (CBP).…”

Section: Introductionmentioning

confidence: 99%

Interventional Behavior Prediction: Avoiding Overly Confident Anticipation in Interactive Prediction

Tang¹,

Zhan²,

Tomizuka³

2022

Preprint

View full text Add to dashboard Cite

Conditional behavior prediction (CBP) builds up the foundation for a coherent interactive prediction and planning framework that can enable more efficient and less conservative maneuvers in interactive scenarios. In CBP task, we train a prediction model approximating the posterior distribution of target agents' future trajectories conditioned on the future trajectory of an assigned ego agent. However, we argue that CBP may provide overly confident anticipation on how the autonomous agent may influence the target agents' behavior. Consequently, it is risky for the planner to query a CBP model. Instead, we should treat the planned trajectory as an intervention and let the model learn the trajectory distribution under intervention. We refer to it as the interventional behavior prediction (IBP) task. Moreover, to properly evaluate an IBP model with offline datasets, we propose a Shapley-value-based metric to testify if the prediction model satisfies the inherent temporal independence of an interventional distribution. We show that the proposed metric can effectively identify a CBP model violating the temporal independence, which plays an important role when establishing IBP benchmarks.

show abstract

Identifying Driver Interactions via Conditional Behavior Prediction

Cited by 61 publications

References 19 publications

MultiPath++: Efficient Information Fusion and Trajectory Aggregation for Behavior Prediction

MultiPath++: Efficient Information Fusion and Trajectory Aggregation for Behavior Prediction

Analyzing Multiagent Interactions in Traffic Scenes via Topological Braids

Interventional Behavior Prediction: Avoiding Overly Confident Anticipation in Interactive Prediction

Contact Info

Product

Resources

About