INTERACT: Interactive Planning for Autonomous Driving via Anchor-Conditioned Prediction and Trust-Region Refinement

Aron Distelzweig; Andreas Look; Faris Janjoš; Steffen Hagedorn; Luigi Palmieri; Joschka Boedecker

cs.ROSep 25, 2026

INTERACT: Interactive Planning for Autonomous Driving via Anchor-Conditioned Prediction and Trust-Region Refinement

Authors: Aron Distelzweig, Andreas Look, Faris Janjoš, Steffen Hagedorn, Luigi Palmieri, Joschka Boedecker

Organizations: Robert Bosch GmbH, Germany · University of Freiburg, Germany · Coburg University, Germany

Abstract

Driving in dense urban traffic is interactive: whether a merge or an unprotected turn succeeds depends on how surrounding agents respond to the ego vehicle. Conventional planners predict first and plan second and, therefore, cannot account for this dependency. Methods that integrate prediction and planning either train both jointly, which introduces task interference, or keep them separate and are restricted to a predefined set of proposals. We present INTERACT: Interactive Planning for Autonomous Driving via Anchor-Conditioned Prediction and Trust-Region Refinement. Our key insight is that surrounding agents react to the intent a trajectory expresses rather than to its exact realization, so a single reactive prediction stays valid across an entire family of plans. INTERACT therefore decomposes interactive planning into prediction across driving intents and optimization within each intent. We derive a small set of diverse intents, which we call anchors, from map geometry, query a dedicated ego-conditioned prediction model once per anchor, and refine every anchor with the Cross-Entropy Method under a trust-region penalty that keeps the refined plan close enough to its anchor for the conditioned reaction to still apply. Prediction thus remains a separate model, avoiding task interference, while conditioning on anchors preserves the dependency. Because each anchor is refined continuously, the final plan is not restricted to the anchor set, yet INTERACT requires only one predictor query per anchor rather than one per candidate plan, with all anchors processed in parallel. On the nuPlan and interPlan closed-loop benchmarks, INTERACT sets a new state of the art, with the largest gains precisely in the interactive scenarios that motivate the method. The code will be released upon acceptance.

Figures & tables

Fig. 1: Interactive planning beyond a fixed proposal set. Proposal-based planners (top) score a fixed set of candidate trajectories and select the best one; the executed plan can therefore never be better than the best candidate in that set. Interact (bottom) treats each candidate as an anchor that expresses a driving intent, queries an ego-conditioned predictor once per anchor, and refines the anchor continuously under a trust-region penalty that keeps the reaction valid. Scoring and selection then operate on refined plans, so the final trajectory is no longer confined to the initial set, at the cost of one predictor query per anchor rather than one per candidate plan.

Fig. 2: Overview of INTERACT. 1 Anchor Generation: We generate a diverse set of anchors, representing different ego intents. 2 Anchor-Conditioned Prediction: For each anchor, a conditional prediction model predicts how surrounding agents react to the corresponding ego intent. 3 Trajectory Refinement: Each anchor is independently refined with Cross-Entropy Maximization (CEM) within a local region, preserving its maneuver intent and thereby the validity of the corresponding anchor-conditioned prediction.

Fig. 3: Comparison of our continuous score components (left) with their discrete PDM counterparts (right) on a nuPlan scenario. For every map location, the progress, TTC, and collision terms are evaluated. Color encodes the score value. The discrete terms are piecewise constant: gradients exist only at edges, leaving the search without guidance on the flat plateaus, backing up scores identically to standing still, and every colliding pose is equally bad. The continuous terms provide an informative gradient throughout: progress degrades smoothly behind the ego and pulls laterally toward the goal lane, TTC decays with the closing distance to slower vehicles, and the collision term grades severity by footprint overlap, pointing out of the conflict.

Predictions	Method	Val14 (R) $\uparrow$	Test14-hard (R) $\uparrow$	Test14-random (R) $\uparrow$	Dynamic-10 (R) $\uparrow$	interPlan (R) $\uparrow$	interPlan-LC (R) $\uparrow$	interPlan-full (R) $\uparrow$
None	Urban Driver [ 24 ]	64.11	49.95	67.15	65.98	4.00	9.65	—
	GC-PGP [ 25 ]	63.82	39.36	51.39	—	14.55	34.99	—
	PlanCNN [ 26 ]	72.00	52.20	67.50	—	—	58.35	—
	PlanTF [ 27 ]	76.95	61.61	79.58	—	30.53	55.18	—
	PDM-Open [ 4 ]	54.24	35.83	57.23	—	—	23.35	—
	DiffusionPlanner [ 28 ]	82.80	69.22	82.93	—	25.76	25.96	—

TABLE I: Comparison of planning performance on the nuPlan benchmarks, grouped by the type of agent prediction the planner consumes. We report closed-loop reactive (R) scores for the Val14, Test14-hard, Test14-random, and Interactive splits, as well as for interPlan, interPlan-LC, and interPlan-full. Higher values indicate better performance. Best and second-best values per column are marked in bold and underlined , respectively. Our method is highlighted . *BIBeR-CV filters candidate plans based on constant-velocity predictions but uses reactive learned predictions afterwards.

Fig. 4: Sensitivity to the trust-region weight

\lambda

on Test14-hard (left) and interPlan-LC (right): closed-loop score (solid), TTC compliance (dashed) and progress along the route (dash-dotted). A small

\lambda

lets the CEM trade safety for progress, a large

\lambda

keeps it close to the anchors and sacrifices progress without further safety gains.

Test14-hard
Configuration	At- Fault	TTC	DAC	Making Progress	Progress along Route	Comfort	Score
Ours	97.43	89.71	98.90	96.32	85.10	81.25	85.35
w/o anchors	$-$ 4.25	$-$ 10.00	$-$ 4.43	$-$ 5.92	$-$ 4.17	+12.48	$-$ 13.35
w/o refinement	+0.74	+2.94	$-$ 0.37	$-$ 1.47	$-$ 7.40	$-$ 2.57	$-$ 2.92
w/o continuous score	$-$ 1.47	$-$ 3.31	0.00	+0.37	$-$ 1.16	+0.74	$-$ 2.19
interPlan-LC

TABLE II: Component ablation on Test14-hard and the lane-change scenarios of interPlan (interPlan-LC). Each row removes one component of the full approach and reports the difference to the full approach (blue: better, red: worse).

Test14-hard
Prediction	At- Fault	TTC	DAC	Making Progress	Progress along Route	Comfort	Score
Conditioned (Ours)	97.43	89.71	98.90	96.32	85.10	81.25	85.35
Unconditioned	+0.37	+0.74	0.00	$-$ 0.37	0.00	+2.21	+0.14
Constant velocity	$-$ 0.37	$-$ 1.47	0.00	0.00	$-$ 0.19	$-$ 1.47	$-$ 1.24
interPlan-LC
Conditioned (Ours)	100.00	100.00	100.00	100.00	80.34	76.67	89.07

TABLE III: Effect of prediction used on Test14-hard and lane-change scenarios of interPlan (interPlan-LC). Rows below ours report the difference to our anchor-conditioned SMART prediction (blue: better, red: worse).

Fig. 5: Sensitivity to the number of anchors

N

on Test14-hard (left) and interPlan-LC (right): closed-loop score (solid), at-fault collisions (dashed) and progress along the route (dash-dotted). Going from one to five anchors unlocks progress additional anchors only add risk (at-fault collisions on Test14-hard) without improving the score.

Stage	Runtime (ms)
Anchor generation	32.6
SMART inference	17.2
Refinement	31.8

TABLE IV: Runtime (ms) per anchor averaged over ten interactive scenarios.

Autonomous Driving Planning

Autonomous driving planning selects safe, feasible vehicle actions and trajectories amid traffic, uncertainty, road rules, and changing goals.

See all 385 papers

Driving is a Game: Combining Planning and Prediction with Bayesian Iterative Best Response
Dec 3, 2025Aron Distelzweig, Yiwei Wang, Faris Janjoš +5Autonomous Driving Planning Multi-Agent Planning
University of Freiburg, Germany · Bosch Center for Artificial Intelligence, Germany · Technical University of Munich, Germany +2
A Unified Framework for Trajectory Prediction with Explicit Planning and Reaction Decomposition
Aug 6, 2026Jiaheng Chen, Jiaxing Li, Tinghe Zhang +1Pedestrian Trajectory Prediction Multi-Agent Trajectory Prediction
Software College, Northeastern University Shenyang, China
Bridging Predictive Uncertainty and Safe Action: Sample-Conditioned Differentiable Planning for Autonomous Driving
Jun 2, 2026Chengzhen Meng, Pei Liu, Zhiyu Huang +2Diffusion-Based Planning Risk-Sensitive Planning
Robotics and Autonomous Systems Thrust, The Hong Kong University of Science and Technology, Guangzhou, China, and also the Department of Electronic and Computer Engineering, The Hong Kong University of Science and Technology, Hong Kong SAR, China · Department of Civil and Environmental Engineering, University of California, Los Angeles, CA 90095 USA · School of Mechanical and Aerospace Engineering, Nanyang Technological University, Singapore 639798

Dec 3, 2025J/K move · Enter open · S save

Driving is a Game: Combining Planning and Prediction with Bayesian Iterative Best Response

Aron Distelzweig, Yiwei Wang, Faris Janjoš, Marcel Hallgarten+4

University of Freiburg, Germany · Bosch Center for Artificial Intelligence, Germany · Technical University of Munich, Germany+2

Autonomous Driving Planning Multi-Agent Planning

Open paper AskPDF arXiv

Autonomous driving planning systems perform nearly perfectly in routine scenarios using lightweight, rule-based methods but still struggle in dense urban traffic, where lane changes and merges require anticipating and influencing other agents. Modern motion predictors offer highly accurate forecasts, yet their integration into planning is mostly rudimental: discarding unsafe plans. Similarly, end-to-end models offer a one-way integration that avoids the challenges of joint prediction and planning modeling under uncertainty. In contrast, game-theoretic formulations offer a principled alternative but have seen limited adoption in autonomous driving. We present Bayesian Iterative Best Response (BIBeR), a framework that unifies motion prediction and game-theoretic planning into a single interaction-aware process. BIBeR is the first to integrate a state-of-the-art predictor into an Iterative Best Response (IBR) loop, repeatedly refining the strategies of the ego vehicle and surrounding agents. This repeated best-response process approximates a Nash equilibrium, enabling bidirectional adaptation where the ego both reacts to and shapes the behavior of others. In addition, our proposed Bayesian confidence estimation quantifies prediction reliability and modulates update strength, more conservative under low confidence and more decisive under high confidence. BIBeR is compatible with modern predictors and planners, combining the transparency of structured planning with the flexibility of learned models. Experiments show that BIBeR achieves an 11% improvement over state-of-the-art planners on highly interactive interPlan lane-change scenarios, while also outperforming existing approaches on standard nuPlan benchmarks.

Abstract

Autonomous Driving Planning

Explore similar work

Driving is a Game: Combining Planning and Prediction with Bayesian Iterative Best Response

A Unified Framework for Trajectory Prediction with Explicit Planning and Reaction Decomposition

Bridging Predictive Uncertainty and Safe Action: Sample-Conditioned Differentiable Planning for Autonomous Driving