Autonomous Vehicle Path Prediction Using Adaptive Error Bounds

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Current path prediction technologies in autonomous vehicles are inaccurate, leading to potential failures in decision-making processes, particularly in predicting the paths of target vehicles, which can impact the safe operation of autonomous vehicles.

Innovation Solution

A controller in the host vehicle calculates an adaptive error bound for confidence intervals of the predicted position of a target vehicle, using a quadratic polynomial method to generate a range of likely positions, thereby enhancing the robustness of target path prediction and autonomous operation by continuously updating as the vehicle travels.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If current path prediction technologies are used, then the autonomous vehicle can operate with basic path prediction capability, but the prediction accuracy is insufficient leading to potential decision-making failures

Engineering Contradiction:

Improvepath prediction accuracyVSAvoiddecision-making reliability

Core Design Contradiction:

Measurement precisionVSReliability

Solution Approach 1:

The system performs preliminary path prediction for multiple candidate trajectories before the vehicle reaches the decision point. By pre-calculating predicted paths for multiple possible trajectories using a receding horizon approach, the system prepares advance information about potential future positions of other vehicles, enabling more reliable real-time decision-making.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The path prediction system dynamically adapts to changing conditions by continuously updating predictions as the autonomous vehicle moves along its trajectory. The receding horizon approach allows the prediction model to be re-evaluated at each time step with updated sensor data, making the prediction accuracy adapt to dynamic environmental changes and improving overall reliability.

Inventive Principle:

Principle #15Dynamics

2Adaptability or versatility

If a single predicted path is used, then the computation is simple, but the system cannot account for multiple possible trajectories and uncertainties

Engineering Contradiction:

Improvemulti-trajectory prediction capabilityVSAvoidprediction system complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The prediction problem is segmented into multiple independent candidate trajectories rather than attempting to predict a single complex path. Each candidate trajectory is evaluated separately using the receding horizon approach, allowing the system to handle multiple possibilities while keeping the computational complexity of each individual prediction manageable.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system generates predictions for multiple candidate trajectories (excessive action) rather than just the single most likely path. This approach provides a more comprehensive view of possible future scenarios, enabling the decision-making system to select the safest trajectory by comparing multiple predicted outcomes.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS10118610B2Autonomous vehicle using path prediction

Publication Date: 2018.11.06 FORD GLOBAL TECH LLC

AI summary

A controller in a host vehicle includes a processor and a memory storing processor-executable instructions. The processor is programmed to calculate an error bound for confidence intervals of a predicted position at a future time of a target vehicle relative laterally to a current position and orientation of the host vehicle at a current time based on a current position and velocity of the target vehicle and the future time.