Vehicle Track Prediction Using Potential Positions

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing autonomous vehicle tracking prediction methods struggle to accurately forecast the trajectory of a target vehicle, especially during emergency maneuvers or when the driver's intention changes, leading to inaccurate predictions based solely on current state and lane information.

Innovation Solution

A method and device that calculate potential positions and predict traveling tracks for a target vehicle by combining historical traveling data, road information, and environment data, using a prediction model and algorithms to select the most probable tracks based on probabilities and environment factors.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If track prediction is performed based solely on current state and lane information, then the prediction method is simple, but the prediction accuracy deteriorates during emergency maneuvers or driver intention changes

Engineering Contradiction:
Improveprediction method complexityVSAvoidtrack prediction accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The system pre-calculates multiple potential positions where the target vehicle may reach within a preset time based on historical traveling data and road information, before the actual maneuver occurs. This preliminary preparation of multiple candidate positions enables accurate prediction even when emergency maneuvers or driver intention changes happen, resolving the contradiction between simple methodology and high accuracy.

Inventive Principle:
Principle #10Preliminary action

2Duration of action of moving object

If track prediction is performed within a relatively long time period, then the prediction covers more scenarios, but the prediction fails to reflect real-time driver intention changes

Engineering Contradiction:
Improveprediction time periodVSAvoidreflection of real driver intention
Core Design Contradiction:
Duration of action of moving objectVSReliability

Solution Approach 1:

The system dynamically adjusts the prediction approach by selecting multiple potential positions within a preset time that is optimized to balance coverage and responsiveness. By using historical traveling data to identify patterns and combining this with real-time environment information, the system maintains reliability in reflecting driver intention while covering sufficient time horizon for scenario planning.

Inventive Principle:
Principle #15Dynamics

3Measurement precision

If multiple potential positions and tracks are calculated, then the prediction accuracy is improved, but the calculation complexity increases

Engineering Contradiction:
Improvetrack prediction accuracyVSAvoidcalculation complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The prediction process is segmented into distinct modules: historical data analysis module that identifies traveling patterns, potential position calculation module that generates candidate positions, and track prediction module that selects most probable tracks. This segmentation allows the system to handle multiple potential positions and tracks systematically, improving accuracy while managing calculation complexity through structured processing.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3699048B1Travelling track prediction method and device for vehicle
Publication Date: 2022.11.30 APOLLO INTELLIGENT DRIVING (BEIJING) TECHNOLOGY CO LTD
  • EP3699048B1 patent drawingFigure 1
  • EP3699048B1 patent drawingFigure 2~3
  • EP3699048B1 patent drawingFigure 4~6

AI summary

A travelling track prediction method and device for a vehicle are provided according to the present invention. The method includes: calculating (S100) a plurality of potential positions to which is to be reached by a target vehicle (2) within a preset time within a sensible range of a main vehicle (1); selecting (S200) at least two positions (a, b, c) from the plurality of potential positions as target positions; predicting (S300) a plurality of travelling tracks (a1, a2, a3, a4, b1, b2, b3) to each target position for the target vehicle (2) based on travelling state information of the target vehicle (2); and selecting (S400) at least one travelling track from the plurality of travelling tracks as a prediction result of the target vehicle (2) based on environment information around the target vehicle. The travelling track of the target vehicle (2) around the main vehicle (1) may thus be accurately predicted based on travelling state information and the environment information.