Autonomous Vehicle Trajectory Prediction via Hypothesis Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in safely navigating through transportation networks due to the inability to accurately detect and predict static and dynamic objects, which can lead to collisions and inefficient route planning.
Innovation Solution
The implementation of a method and system for world objects tracking and prediction by an autonomous vehicle, using sensors to identify oncoming vehicles and road objects, generate hypotheses about their intentions, and compute reference drivelines to plan either lane-following or go-around trajectories, ensuring safe navigation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If autonomous vehicles use basic sensor detection and simple trajectory planning, then the system complexity is low, but the ability to accurately detect and predict static and dynamic objects is insufficient, leading to collisions and unsafe navigation
Solution Approach 1:
The system segments the detection and prediction process into distinct modules: sensor data acquisition module, object detection module, trajectory prediction module (generating multiple hypotheses), and trajectory planning module. This segmentation allows each module to specialize in specific tasks, improving overall reliability while managing complexity through modular design.
Solution Approach 2:
The system performs preliminary actions by generating multiple trajectory hypotheses for dynamic objects before final trajectory planning. The prediction module pre-computes possible future positions and paths of other vehicles and objects, allowing the planning module to make safer decisions based on anticipated scenarios rather than reacting to current states only.
2Measurement precision
If autonomous vehicles implement multiple hypothesis generation and complex trajectory prediction for oncoming vehicles, then the accuracy of predicting road user trajectories improves, but the computational complexity and processing time increase
Solution Approach 1:
The system applies dynamics by generating multiple trajectory hypotheses that represent different possible future behaviors of oncoming vehicles (e.g., continuing straight, turning left, turning right). Each hypothesis dynamically adapts based on detected road objects and current vehicle states, allowing the system to maintain high prediction accuracy while managing computational complexity through selective hypothesis generation.
Solution Approach 2:
The system creates simplified copies of potential trajectory scenarios rather than computing all possible physical interactions. Each hypothesis is a representative model of a possible future state, allowing the system to evaluate multiple outcomes efficiently without requiring exhaustive simulation of every variable interaction.
3Manufacturing precision
If autonomous vehicles generate multiple trajectory hypotheses and compute reference drivelines for each, then the route planning accuracy improves, but the processing time and computational resources increase
Solution Approach 1:
The system performs preliminary computation of reference drivelines for multiple trajectory hypotheses before selecting the final trajectory. By pre-computing these reference paths based on current sensor data and predicted object positions, the system reduces the time required for final trajectory selection and execution, maintaining high precision while managing processing time.
Solution Approach 2:
The system applies local quality by computing reference drivelines with high precision only for the most promising trajectory hypotheses rather than all possible trajectories. The hypothesis evaluation process identifies which trajectories are most likely based on detected road objects and vehicle behaviors, allowing concentrated computational resources to be applied where they provide the greatest precision benefit.
Data Source
AI summary
World objects tracking and prediction by an autonomous vehicle (AV) is disclosed. A method includes identifying, based on first observation data received from sensors of the AV, an oncoming vehicle; identifying, based on second observation data received from the sensors of the AV, a road object; generating a lane-following hypothesis for the oncoming vehicle, the lane-following hypothesis indicating an intention that the oncoming vehicle remain in a current road lane; computing a lane-following reference driveline for the lane-following hypothesis of the oncoming vehicle; in response to determining that the lane-following reference driveline is blocked by the road object, generating a go-around hypothesis for the oncoming vehicle, and computing a go-around reference driveline for the go-around hypothesis; and providing at least one of a go-around trajectory corresponding to the go-around hypothesis or a lane-following trajectory corresponding to the lane-following hypothesis.