Vehicle Perception Compute Resource Allocation via Intent-Based Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face a trade-off between accurate perception of their environment and quick reaction time due to the high latency caused by large amounts of image data from high-resolution sensors like computer vision devices and LIDAR, which decreases reaction time.
Innovation Solution
A perception system that dynamically adjusts the resolution, region of interest (ROI), and compute resources based on the vehicle's intent and current state, prioritizing resources for the most critical areas, such as the planned route and velocity, to optimize image data processing and reduce overall system latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If high-resolution imagers are used to provide large amounts of image data for accurate perception, then measurement precision is improved, but loss of time increases due to increased latency
Solution Approach 1:
The patent segments the imaging data into different regions of interest (ROIs) based on spatial priority. Instead of processing the entire high-resolution image uniformly, the system divides the image into multiple ROIs with different resolution requirements. Critical areas receive high-resolution processing while less critical areas use lower resolution, thereby reducing overall processing time while maintaining perception accuracy where needed.
Solution Approach 2:
The patent applies local quality by assigning different processing qualities to different spatial regions. Each ROI is assigned a quality level based on its importance for the current driving task. This allows the system to maintain high measurement precision in critical regions while reducing processing time in less critical regions, resolving the contradiction between overall accuracy and overall latency.
2Measurement precision
If high-resolution imagers are used to provide large amounts of image data for accurate perception, then measurement precision is improved, but device complexity increases
Solution Approach 1:
The patent implements dynamics by making the compute resource allocation adaptive rather than static. The system dynamically adjusts the resolution and processing resources allocated to different ROIs based on real-time driving context, vehicle state, and task priorities. This dynamic adaptation allows the system to maintain perception accuracy when needed while simplifying resource allocation complexity through automated decision-making.
Solution Approach 2:
The patent changes key parameters including resolution, ROI boundaries, and compute resource allocation based on driving context. By adjusting these parameters dynamically, the system can maintain high measurement precision for critical perception tasks while managing compute resource complexity through parameter-based control rather than complex hardware configurations.
3Device complexity
If compute resources are allocated uniformly to all imaging devices, then device complexity is simplified, but productivity decreases due to inefficient resource utilization
Solution Approach 1:
The patent applies local quality to resource allocation by assigning different compute resource levels to different ROIs based on their importance. Critical ROIs receive more compute resources while less critical ROIs receive fewer resources, thereby improving processing efficiency without requiring complex manual allocation mechanisms. The system automatically determines resource distribution based on spatial priority and driving context.
Solution Approach 2:
The system implements self-service by automatically determining optimal compute resource allocation based on the driving context and ROI priorities. Rather than requiring complex external control mechanisms, the perception system autonomously adjusts resource distribution to maximize processing efficiency, thereby improving productivity while keeping the allocation mechanism relatively simple.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The present disclosure provides perception system for a vehicle that includes a plurality of imaging devices for producing images of an environment of the vehicle; a perception filter for receiving the images produced by the imaging devices, wherein the perception filter determines compute resource priority instructions based on an intent of the vehicle and a current state of the vehicle; and a compute module for receiving the compute resource priority instructions from the perception filter and allocating compute resources among the imaging devices in accordance with the compute resource priority instructions.