Collision Avoidance Perception System Using Secondary Sensor Corridor Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in accurately distinguishing objects from the ground in environments with changes in grade, leading to potential false positive or false negative classifications, which can result in unsafe trajectories and increased computational bandwidth, memory, and power consumption.
Innovation Solution
A collision avoidance system with a secondary perception component that classifies sensor data as either ground or object by fitting a spline or curve to model the roadway surface, using techniques like least squares regression and weighting sensor data to improve accuracy, and determining threshold distances to validate or reject trajectories.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the autonomous vehicle uses multiple sensors and systems to detect and track all objects in a 360-degree view, then the safety and collision avoidance capability is improved, but the computational bandwidth, memory, and power consumption increase
Solution Approach 1:
The patent segments the sensor data processing by dividing it into a primary perception component that processes all sensor data comprehensively, and a secondary perception component that focuses specifically on collision avoidance. This segmentation allows the system to maintain high safety standards while reducing the computational burden on any single processing component, as the secondary component only needs to analyze data relevant to immediate collision risks rather than all environmental objects.
Solution Approach 2:
The patent applies partial action by implementing a secondary perception component that performs a focused, simplified analysis of sensor data specifically for collision avoidance purposes. Rather than requiring all sensors to process all data at maximum detail, the system performs a targeted partial analysis on a subset of data (those potentially relevant to collision risks), which reduces overall computational resources while maintaining adequate safety monitoring.
2Difficulty of detecting and measuring
If the system classifies all sensor data as objects or ground in complex environments with changes in grade, then the detection coverage is improved, but the classification accuracy deteriorates due to false positives and false negatives
Solution Approach 1:
The patent segments the classification task into two distinct processing stages: a primary perception component that performs comprehensive classification of all sensor data, and a secondary perception component that performs a focused classification specifically on data within the corridor of interest. This segmentation allows each component to optimize its classification approach for its specific task, improving overall accuracy by reducing the complexity of individual classification decisions.
Solution Approach 2:
The patent introduces an intermediary mechanism in the form of a secondary perception component that acts as a mediator between the raw sensor data and the final collision avoidance decisions. This intermediary performs an additional layer of verification and classification specifically for collision-relevant objects, refining the initial classification results and reducing false positives and false negatives by applying specialized analysis to critical data.
3Loss of information
If the autonomous vehicle monitors the entire 360-degree environment, then the situational awareness is improved, but the time and computational resources required for processing increase
Solution Approach 1:
The patent segments the situational awareness function into a primary perception component that maintains comprehensive 360-degree environmental awareness, and a secondary perception component that focuses specifically on the corridor for collision avoidance. This segmentation allows the system to preserve complete situational awareness while reducing processing time for collision detection by directing focused attention only to relevant areas when needed.
Solution Approach 2:
The patent applies partial action by implementing a secondary perception component that performs focused analysis only on the corridor region rather than the entire 360-degree environment. This partial monitoring approach reduces processing time significantly while maintaining adequate collision avoidance capability, as the system only needs detailed analysis in the forward corridor where collision risks are most imminent.
Data Source
AI summary
A collision avoidance system may validate, reject, or replace a trajectory generated to control a vehicle. The collision avoidance system may comprise a secondary perception component that may receive sensor data, receive and/or determine a corridor associated with operation of a vehicle, classify a portion of the sensor data associated with the corridor as either ground or an object, determine a position and/or velocity of at least the nearest object, determine a threshold distance associated with the vehicle, and control the vehicle based at least in part on the position and/or velocity of the nearest object and the threshold distance.