Autonomous Vehicle Trajectory Planning for Low-Latency Evasive Maneuvers

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Autonomous vehicles face challenges in reacting quickly to dynamic road situations due to latency in existing trajectory planning systems, which can lead to potential collisions.

Innovation Solution

Implementing a secondary planning system with reduced latency to generate trajectories, leveraging high compute power and prioritized processing to quickly adapt to changing road user behaviors, while the primary planning system generates multiple trajectories for evaluation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a primary planning system generates multiple trajectories for evaluation, then trajectory selection accuracy is improved, but reaction latency increases

Engineering Contradiction:

Improvetrajectory selection accuracyVSAvoidreaction latency

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The planning system is divided into two independent components: a primary planning system that generates multiple trajectories for accuracy, and a secondary planning system that generates a single trajectory for speed. This segmentation allows each subsystem to optimize for its specific function without compromising the other.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The secondary planning system performs a simplified version of trajectory planning by generating only one trajectory instead of multiple trajectories. This partial action reduces computational overhead and latency while still providing a viable trajectory option for emergency situations.

Inventive Principle:

Principle #16Partial or excessive action

2Productivity

If a streamlined behavior prediction system is used, then processing speed is improved, but prediction accuracy may be reduced

Engineering Contradiction:

Improveprocessing speedVSAvoidprediction accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

Different behavior prediction systems are applied to different situations: the comprehensive first behavior prediction system is used when time permits and high accuracy is needed, while the streamlined second behavior prediction system is used when rapid response is critical. This local quality approach optimizes the balance between speed and accuracy based on situational requirements.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The second behavior prediction system performs a simplified prediction with reduced computational steps compared to the first system. This partial action enables faster processing by focusing only on the most critical prediction elements, sacrificing some accuracy for speed in time-sensitive scenarios.

Inventive Principle:

Principle #16Partial or excessive action

3Reliability

If a secondary planning system is added to reduce latency, then system complexity increases, but collision avoidance capability is improved

Engineering Contradiction:

Improvecollision avoidance capabilityVSAvoidsystem complexity

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The secondary planning system acts as an intermediary emergency response mechanism that activates when the primary system's latency is unacceptable. This intermediary layer provides a simplified fast-path trajectory generation that complements rather than replaces the primary system, improving reliability without requiring complete system redesign.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS20260077786A1Trajectory planning systems for autonomous vehicles

Publication Date: 2026.03.19 WAYMO LLC

AI summary

Aspects of the disclosure may enable an autonomous vehicle to perform evasive maneuvers in order to avoid potential collisions. For instance, a vehicle may be controlled in an autonomous driving mode using a first trajectory generated by a first planning system. Information identifying a characteristic of a road user in an environment of the vehicle may be received and used to determine whether to generate a new behavior prediction for the road user or to use a prior determined behavior prediction. While using the first trajectory, a second planning system may be used to generate a second trajectory based on the determination. The second trajectory may be compared to the first trajectory. The vehicle may be controlled in the autonomous driving mode based on a result of the comparing.