Continuous Trajectory Prediction for Real-Time Vehicle Control
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining the future trajectory of objects from sensor data in automotive applications provide coarse predictions due to the discrete nature of data samples, which are insufficient for precise vehicle control, and increasing prediction frequency conflicts with real-time processing limitations.
Innovation Solution
A computer-implemented method using a machine-learning model, such as an artificial neural network, to determine continuous information on an object's trajectory by calculating parameter values for a continuous function, allowing precise and safe vehicle control at any desired time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of predictions per time span is increased to improve trajectory accuracy, then measurement precision improves, but processing resources are exceeded and real-time capability is lost
Solution Approach 1:
The patent applies continuity by transitioning from discrete pointwise predictions to continuous trajectory predictions. The machine learning model outputs continuous trajectory information that provides prediction values for any point in time within a time span, eliminating the need for multiple discrete predictions and enabling real-time processing with high accuracy.
Solution Approach 2:
The patent changes the parameter representation from discrete time samples to continuous function parameters. By representing trajectories as continuous functions with parameters (e.g., polynomial coefficients), the system achieves high measurement precision while reducing the computational burden of generating multiple discrete predictions.
2Device complexity
If discrete data samples are used for trajectory prediction, then processing resources are conserved, but manufacturing precision of control functions deteriorates
Solution Approach 1:
The patent replaces discrete data sampling with continuous trajectory prediction. The machine learning model generates continuous trajectory information that can be evaluated at any point in time, providing the precision needed for control functions without requiring numerous discrete predictions that would consume excessive processing resources.
Solution Approach 2:
The patent substitutes the mechanical approach of generating multiple discrete prediction points with a mathematical/algorithmic approach using continuous functions. The machine learning model outputs parametric representations (e.g., polynomial coefficients) that can be efficiently evaluated anywhere in the time domain, reducing computational complexity while maintaining high precision.
Data Source
AI summary
Computer-implemented method for determining continuous information on an expected trajectory of an object, the method comprising at least the following steps carried out by computer hardware components: determining data related to an expected trajectory of an object; and determining at least one parameter value for a continuous function on the basis of the data, wherein the continuous function and the at least one parameter value represent continuous information on the expected trajectory of the object.