Trajectory Prediction Failure Detection for Motion Planner Reuse
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face safety risks due to failures in trajectory prediction modules, leading to unsafe decision-making and excessive computational resource consumption.
Innovation Solution
Implementing systems and methods for task-relevant failure detection by determining cost probability distributions and evaluating them using anomaly detection, with threshold values to ensure false-positive and false-negative rate guarantees, allowing for adaptive control operations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If trajectory prediction modules are used to plan motion in autonomous vehicles, then the vehicle can navigate through the environment, but prediction failures can lead to unsafe decisions and excessive computational resource consumption
Solution Approach 1:
The system performs preliminary anomaly detection on predicted trajectories before they are executed by the motion planner. By evaluating whether predicted trajectories satisfy anomaly conditions in advance, the system prevents unsafe trajectories from reaching the execution stage, thereby improving safety without requiring continuous full-replanning computations.
Solution Approach 2:
The patent extracts and separates the anomaly detection function from the main trajectory planning and execution pipeline. This modular approach allows the system to quickly evaluate predicted trajectories against anomaly conditions without interfering with the normal planning operations, reducing computational overhead while maintaining safety monitoring.
2Reliability
If the autonomous vehicle performs frequent replanning operations to avoid collision, then safety is improved, but computational resources are consumed
Solution Approach 1:
The system performs preliminary anomaly detection on predicted trajectories before they are executed by the motion planner. By evaluating whether predicted trajectories satisfy anomaly conditions in advance, the system prevents unsafe trajectories from reaching the execution stage, thereby improving safety without requiring continuous full-replanning computations.
3Measurement precision
If anomaly detection with strict threshold values is implemented, then false-positive and false-negative rates are controlled, but detection accuracy requirements increase
Solution Approach 1:
The system uses the predicted trajectory data itself to establish the anomaly detection thresholds. By sampling predicted costs from the trajectory predictions and using these samples to determine threshold values (such as p-quantiles), the system adapts the detection criteria to the specific prediction model being used, eliminating the need for external calibration or complex statistical modeling.
Data Source
AI summary
In various examples, cost probability distributions corresponding to predicted locations of an object in an environment and potential locations for a machine in the environment and may be evaluated using corresponding observed costs corresponding to the machine and the object. The cost probability distributions may be evaluated based on comparing the observed costs to threshold values, which may be determined based on sampling a predicted cost function. A threshold value may be selected to provide false-positive rate and/or false-negative rate guarantees for anomaly detection. Control operations may be performed based on results of the evaluation of the cost probability distributions. For example, based on the results, a motion planner may reuse a planned trajectory for a future planning cycle (e.g., thereby avoiding re-planning computations) or generate and/or select a new planned trajectory (e.g., based at least on one or more anomalies being detected).