Complexity-Scored Agent Prediction Scheduling for AV Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in efficiently allocating computing resources for generating prediction data for multiple agents in their vicinity, leading to sub-optimal schedules that delay or neglect timely and accurate prediction data, which is crucial for safe and comfortable vehicle trajectory planning.
Innovation Solution
A system that uses complexity and importance scoring models to dynamically determine a task assignment schedule for generating prediction data, prioritizing agents with the greatest impact on planning decisions and optimizing the use of limited computing resources, allowing for timely and adequate data generation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If computing resources are allocated to generate prediction data for all agents, then the quantity of prediction data increases, but the time to generate data increases and critical agents may be delayed
Solution Approach 1:
The patent segments the set of all agents into high-priority and low-priority subsets based on their importance to planning decisions. This segmentation allows the system to process agents in groups, ensuring critical agents are handled first while still maintaining the ability to process additional agents if resources permit, thus resolving the contradiction between processing all agents and meeting time constraints.
Solution Approach 2:
The system performs preliminary identification of high-priority agents before full prediction data generation. By pre-classifying agents based on their importance to planning decisions, the system ensures that critical agents are immediately processed while other agents can be processed in parallel or sequentially based on available resources, preventing delays for critical agents.
2Measurement precision
If computing resources are allocated to generate prediction data for multiple agents, then the accuracy of planning decisions improves, but the computing power required increases
Solution Approach 1:
The patent divides computing resources into dedicated pools for high-priority and low-priority agents. This segmentation allows the system to allocate sufficient computing power to ensure accurate prediction data for critical agents while using remaining resources for other agents, thus maintaining accuracy for planning-critical agents without requiring excessive total computing power.
Solution Approach 2:
The system applies different levels of processing quality to different agent groups. High-priority agents receive full computational attention and resources to ensure maximum accuracy, while low-priority agents receive processing based on available resources. This local quality differentiation ensures that planning decisions are based on accurate data for critical agents without uniformly increasing computing power requirements for all agents.
3Device complexity
If a fixed schedule is used for generating prediction data, then the system complexity is reduced, but the productivity decreases due to inability to adapt to varying agent priorities
Solution Approach 1:
The system performs preliminary classification of agents into priority groups before scheduling. This pre-processing step enables the use of relatively simple scheduling rules within each priority group while still achieving adaptive behavior. The complexity is localized to the classification stage, allowing straightforward scheduling execution that maintains high productivity.
Solution Approach 2:
The patent changes the scheduling parameter from treating all agents uniformly to processing agents in priority-based groups. This parameter change enables the system to maintain simple scheduling logic within groups while adapting to varying priorities through the grouping parameter, thus resolving the contradiction between simplicity and adaptability.
Data Source
AI summary
Methods, systems, and apparatus, including computer programs encoded on a computer storage medium, for determining a task schedule for generating prediction data for different agents. In one aspect, a method comprises receiving data that characterizes an environment in a vicinity of a vehicle at a current time step, the environment comprising a plurality of agents; receiving data that identifies high-priority agents for which respective data characterizing the agents must be generated at the current time step; identifying available computing resources at the current time step; processing the data that characterizes the environment using a complexity scoring model to determine a respective complexity score for each of the high-priority agents; and determining a schedule for the current time step that allocates the generation of the data characterizing the high-priority agents across the available computing resources based on the complexity scores.