Target Motion State Estimation Without Fixed Driving Models

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing motion state estimation methods for target objects in intelligent driving systems are limited by fixed motion models, requiring cumbersome and time-consuming adjustments of noise matrices to adapt to changing driving scenarios, leading to inaccurate estimations and reduced applicability.

Innovation Solution

Estimate motion states of target objects using historical distances and a motion state estimation model, eliminating the need for fixed motion models and Kalman gain corrections, thereby expanding applicability and reducing estimation time.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If an explicitly constructed motion model with fixed motion mode is used for motion state estimation, then the estimation process can be performed with a defined mathematical model, but the application scenarios are limited and the noise matrix adjustment process is cumbersome and time-consuming

Engineering Contradiction:
Improvemotion state estimation accuracyVSAvoidapplication scenario adaptability
Core Design Contradiction:
ReliabilityVSAdaptability or versatility

Solution Approach 1:

The patent transforms the fixed motion model into a dynamic learning-based model that can adapt to different driving scenarios. The motion state estimation model is trained on diverse scenario data and dynamically adjusts its parameters through learning, enabling it to handle various motion patterns without requiring manual noise matrix adjustments for each scenario.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes the approach from manually adjusting noise matrix parameters to automatically learning optimal parameters from data. The model learns motion patterns, noise characteristics, and correction factors during training, replacing the need for expert-driven parameter tuning with data-driven parameter optimization that adapts to different scenarios automatically.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If noise matrix adjustment is performed based on actual driving environment to determine Kalman gain, then the motion state estimation can be corrected for specific scenarios, but the adjustment process is rather cumbersome and time-consuming

Engineering Contradiction:
Improvemotion state estimation precisionVSAvoidnoise matrix adjustment time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent performs preliminary learning during the training phase, where the model learns optimal motion patterns and correction factors from labeled data across multiple scenarios. This preliminary action embeds scenario-specific knowledge into the model parameters, eliminating the need for time-consuming noise matrix adjustments during actual operation. The model is pre-adapted to handle different driving environments through offline training.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent replaces the mechanical parameter adjustment process (manually tuning noise matrices and Kalman gains) with an intelligent learning system. The model automatically learns and applies appropriate corrections based on the input data, substituting the manual control theory-based adjustment with an automated machine learning-based estimation that achieves similar or better precision without the time cost.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Ease of manufacture

If a fixed motion model is used for motion state estimation, then the mathematical framework is well-defined, but the model cannot adapt to changing driving scenarios and motion patterns

Engineering Contradiction:
Improvemodel implementation simplicityVSAvoidscenario adaptability
Core Design Contradiction:
Ease of manufactureVSAdaptability or versatility

Solution Approach 1:

The patent creates a universal motion state estimation model that can handle multiple driving scenarios and motion patterns through learning. The model is designed to be scenario-agnostic during deployment, automatically adapting to different situations based on the input data it processes. This single model replaces the need for multiple scenario-specific fixed models, achieving multi-functionality while maintaining implementation simplicity.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP4636690A1Method and apparatus for motion state estimation, method and apparatus for training motion state estimation model, storage medium and electronic device
Publication Date: 2025.10.22 HORIZON JOURNEY (SHANGHAI) TECHNOLOGY CO LTD
  • EP4636690A1 patent drawingFigure 1
  • EP4636690A1 patent drawingFigure 2
  • EP4636690A1 patent drawingFigure 3

AI summary

Disclosed are a method and apparatus for motion state estimation, a method and apparatus for training a motion state estimation model, a storage medium and an electronic device, which relate to technical field of intelligent driving. The method includes: determining first distances between an ego vehicle and a target object at a plurality of historical moments and a second distance between the ego vehicle and the target object at a current moment; performing fitting processing on the first distances and the second distance; and processing, based on a motion state estimation model, motion states of the target object at the plurality of historical moments obtained by the performing fitting processing on , to obtain a first estimation motion state of the target object at the current moment and a second estimation motion state of the target object at a future moment.