Blocking Vehicle Detection for Autonomous Trajectory Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in accurately detecting stationary vehicles that may be blocking their path, due to limitations in sensor visibility and environmental cues, leading to potential unsafe operations and increased power consumption.
Innovation Solution
A machine-learning model is used to determine the probability of a stationary vehicle being a blocking vehicle by analyzing sensor data, including features such as distance, speed, and traffic flow, rather than relying on hard-coded conditional rules, allowing for more accurate trajectory planning and decision-making.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensor visibility is used to detect blocking vehicles, then detection capability is limited, but system complexity is reduced
Solution Approach 1:
The patent introduces an intermediary machine learning model that processes sensor data to infer the presence of blocking vehicles. Instead of directly detecting blocking vehicles through sensors, the system uses the stationary vehicle detector and machine learning model as intermediaries to analyze sensor residuals and determine whether a stationary vehicle is actually blocking the path, thereby improving detection accuracy without proportionally increasing sensor complexity
Solution Approach 2:
The patent replaces traditional mechanical/rule-based detection systems with a machine learning-based detection system. Instead of using hard-coded rules to determine if a vehicle is blocking, the system uses trained machine learning models that analyze sensor data patterns, improving detection accuracy while maintaining manageable system complexity through software-based solutions
2Device complexity
If hard-coded conditional rules are used to detect blocking vehicles, then system complexity is reduced, but detection accuracy deteriorates
Solution Approach 1:
The patent changes the detection approach from fixed conditional parameters to dynamic learned parameters. Instead of using static hard-coded rules, the system employs machine learning models that learn optimal detection parameters from training data, allowing the system to adapt to various driving scenarios and improve detection accuracy while managing complexity through parameter optimization
Solution Approach 2:
The patent introduces dynamics into the detection system by using machine learning models that can adapt and learn from data. The detection system transitions from static hard-coded rules to dynamic models that continuously improve their detection capabilities based on training data and environmental variations, improving accuracy while maintaining reasonable system complexity
3Reliability
If all stationary vehicles are treated as blocking vehicles, then safety is improved, but power consumption increases
Solution Approach 1:
The patent applies partial action by selectively detecting and responding to blocking vehicles rather than treating all stationary vehicles uniformly. The machine learning model determines the probability that a stationary vehicle is blocking, allowing the system to take action only when necessary, thereby improving safety while reducing unnecessary power consumption from continuous full-alert operation
Solution Approach 2:
The patent implements feedback mechanisms where the machine learning model continuously evaluates sensor data and updates its determination of blocking vehicle probability. This feedback loop allows the system to adjust its response based on current conditions, improving safety through accurate real-time detection while reducing power consumption by avoiding unnecessary actions when vehicles are not actually blocking
4Measurement precision
If machine-learning models are used to detect blocking vehicles, then detection accuracy is improved, but device complexity increases
Solution Approach 1:
The patent segments the detection system into distinct functional modules: a stationary vehicle detector that identifies stationary vehicles, a machine learning model that determines blocking probability, and a trajectory planner that responds to detections. This segmentation allows each component to be optimized independently, improving overall detection accuracy while managing system complexity through modular design
Solution Approach 2:
The patent creates a multi-functional machine learning model that performs multiple detection and classification tasks. The model not only detects blocking vehicles but also determines the probability of blocking, enabling the system to handle various driving scenarios with a single integrated component, improving detection accuracy while reducing the need for multiple separate systems
Data Source
AI summary
A method and system of determining whether a stationary vehicle is a blocking vehicle to improve control of an autonomous vehicle. A perception engine may detect a stationary vehicle in an environment of the autonomous vehicle from sensor data received by the autonomous vehicle. Responsive to this detection, the perception engine may determine feature values of the environment of the vehicle from sensor data (e.g., features of the stationary vehicle, other object(s), the environment itself). The autonomous vehicle may input these feature values into a machine-learning model to determine a probability that the stationary vehicle is a blocking vehicle and use the probability to generate a trajectory to control motion of the autonomous vehicle.