Probabilistic Object Tracking for Autonomous Vehicles
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 the trajectories of static and dynamic objects, which can lead to inefficient and unsafe operation.
Innovation Solution
A method and system for world objects tracking and prediction, where an autonomous vehicle's world model module receives sensor data, associates it with objects, determines hypotheses about their intentions, calculates likelihoods, and predicts their positions, enabling the vehicle to plan a safe trajectory.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If autonomous vehicles use basic object detection methods, then the system complexity is low, but the tracking and prediction accuracy is insufficient leading to unsafe operation
Solution Approach 1:
The system segments the tracking and prediction task into multiple independent modules: object detection module, hypothesis generation module, likelihood calculation module, and trajectory prediction module. Each module handles a specific aspect of the problem, allowing complex functionality to be achieved through coordinated simple components rather than a single complex system.
Solution Approach 2:
The system dynamically generates and evaluates multiple hypotheses about object intentions and trajectories. Instead of using a static detection approach, the system continuously updates hypothesis likelihoods based on new sensor data and object behavior patterns, allowing the tracking system to adapt to changing conditions while maintaining manageable complexity through probabilistic reasoning.
2Reliability
If autonomous vehicles implement comprehensive object tracking and prediction, then navigation safety is improved, but the computational processing time increases
Solution Approach 1:
The system performs preliminary actions by pre-generating multiple possible trajectory hypotheses for each detected object based on its current state and behavior patterns. This allows the system to have prediction candidates ready before they are needed for collision avoidance decisions, reducing real-time computational burden while maintaining safety through pre-computed likelihood assessments.
Solution Approach 2:
The system changes parameters by dynamically adjusting hypothesis likelihoods based on observed object behavior and contextual factors. Instead of computing all possible trajectories with equal weight, the system modifies probability parameters to focus computational resources on the most likely scenarios, achieving safe navigation with reduced processing time by eliminating low-probability hypotheses early in the decision pipeline.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
World objects tracking and prediction by an autonomous vehicle is disclosed. A method includes receiving, from sensors of the AV, a first observation data; associating the first observation data with a first world object; determining hypotheses for the first world object, wherein a hypothesis corresponds to an intention of the first world object; determining a respective hypothesis likelihood of each of the hypotheses indicating a likelihood that the first world object follows the intention; determining, for at least one hypothesis of the hypotheses, a respective state, wherein the respective state comprises predicted positions of the first world object; and in response to a query, providing a hypothesis of the hypotheses based on the respective hypothesis likelihood of each of the hypotheses.