Autonomous Vehicle Object Tracking via Historical Data Association

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Autonomous vehicles face challenges in accurately detecting and tracking objects in their environment, particularly when objects are temporarily concealed or moving at high speeds, leading to inaccuracies in navigation and prediction of object behavior.

Innovation Solution

A computer system configured to receive sensor data from autonomous vehicles, detect objects, associate detection data with tracking data, and navigate based on predicted object positions, using a combination of real-time and historical data to refine object location and behavior prediction, thereby improving tracking accuracy and reducing transient errors.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If sensors are used to detect objects in real-time, then navigation capability is improved, but tracking accuracy deteriorates when objects are temporarily concealed

Engineering Contradiction:

Improvereal-time detection speedVSAvoidobject tracking accuracy

Core Design Contradiction:

SpeedVSMeasurement precision

Solution Approach 1:

The system performs preliminary tracking by maintaining historical tracking data for objects before they are concealed. When an object is temporarily obscured from sensor detection, the pre-established tracking data allows the system to continue predicting and maintaining object position and behavior, thereby preserving tracking accuracy during periods when real-time sensor detection is unavailable.

Inventive Principle:

Principle #10Preliminary action

2Device complexity

If real-time sensor data only is used, then system complexity is reduced, but tracking reliability deteriorates during object concealment

Engineering Contradiction:

Improvedata processing complexityVSAvoidtracking reliability

Core Design Contradiction:

Device complexityVSReliability

Solution Approach 1:

The system merges real-time sensor data with historical tracking data to create a comprehensive object tracking solution. By combining current sensor detections with previously accumulated tracking information, the system achieves reliable object tracking during concealment events while maintaining manageable system complexity through integrated data processing architecture.

Inventive Principle:

Principle #5Merging (Combining)

3Measurement precision

If historical tracking data is integrated, then prediction accuracy is improved, but data processing time increases

Engineering Contradiction:

Improvebehavior prediction accuracyVSAvoiddata processing time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system extracts only the essential and relevant features from historical tracking data that are necessary for predicting object behavior, rather than processing complete historical datasets. This selective extraction of critical tracking attributes reduces data processing time while preserving the accuracy benefits of historical information for behavior prediction.

Inventive Principle:

Principle #2Taking out (Extraction)

Data Source

PatentUS11675366B2Long-term object tracking supporting autonomous vehicle navigation

Publication Date: 2023.06.13 MOTIONAL AD LLC

AI summary

This disclosure relates in general to systems and methods for tracking objects proximate an autonomous vehicle. In particular, an object tracking system capable of re-identifying objects it has temporarily lost line of sight to is described. Re-identification of the objects allows earlier object detections to be used more effectively to predict motion likely to be taken by the objects.