Adaptive Trajectory Planning Using Imitation Learning

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing trajectory planning systems for autonomous vehicles rely on fixed, non-adaptive parameters that fail to accommodate diverse customer preferences and environmental variations, leading to suboptimal driving experiences.

Innovation Solution

Implement imitation learning to dynamically adapt parameters of the trajectory planner using a machine-learning model trained on driving data, allowing the selection of parameters based on current state and customer preferences.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If fixed, non-adaptive parameters are used in trajectory planning, then system complexity is reduced and reliability is improved, but adaptability to diverse customer preferences and environmental variations deteriorates

Engineering Contradiction:

Improveadaptability to customer preferences and environmental conditionsVSAvoidcomplexity of trajectory planning system

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent applies dynamics by transitioning from fixed, static parameters to dynamic, adaptive parameters in the trajectory planning system. The machine-learning model continuously learns from driving data and environmental conditions, allowing the trajectory planner to adapt its parameters in real-time based on current state, customer preferences, and environmental variations, thereby resolving the contradiction between adaptability and complexity

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system implements self-service through the machine-learning model that automatically learns and adjusts trajectory planning parameters without requiring manual intervention. The model trains on driving data and autonomously optimizes parameters based on observed patterns, enabling the system to serve itself in adapting to diverse conditions while maintaining manageable complexity through automated learning

Inventive Principle:

Principle #25Self-service

2Manufacturing precision

If fixed parameters are used in trajectory planning, then ease of operation is improved, but manufacturing precision of driving experience deteriorates

Engineering Contradiction:

Improveprecision of driving experienceVSAvoidoperational simplicity of trajectory planner

Core Design Contradiction:

Manufacturing precisionVSEase of operation

Solution Approach 1:

The patent replaces the mechanical system of fixed, hard-coded parameters with a machine-learning-based adaptive parameter system. The machine-learning model processes driving data and environmental information to dynamically determine optimal parameters, substituting rigid mechanical configuration with flexible intelligent processing, thereby achieving precise driving experiences while maintaining ease of operation through automated learning

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If imitation learning with machine-learning model is implemented, then adaptability and driving experience are improved, but device complexity and computational requirements increase

Engineering Contradiction:

Improveefficiency of trajectory planningVSAvoidcomplexity of machine-learning based system

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent applies preliminary action by pre-training the machine-learning model on extensive driving data before deployment. The model learns optimal parameter selections in advance from historical driving scenarios, customer preferences, and environmental conditions. This pre-learning phase enables the system to make rapid, efficient decisions during actual trajectory planning without requiring complex real-time computations, thereby improving productivity while managing device complexity

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS12448007B2Imitation learning for trajectory planning

Publication Date: 2025.10.21 NISSAN MOTOR CO LTD

AI summary

Trajectory planning include identifying state data based on sensor data from sensors of a vehicle. The state data are input to a machine-learning model to obtain parameters of a trajectory planner. The parameters are input to the trajectory planner to obtain a short term speed plan. The vehicle is autonomously controlled according to at least a portion of the short term speed plan.