Autonomous Vehicle Task Scheduling for Dependency-Driven Low Latency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current operating system schedulers in autonomous vehicles fail to effectively manage task dependencies and resource allocation, leading to poor performance in executing concurrent real-time tasks, which is critical for low latency in autonomous vehicle systems.
Innovation Solution
A method and system for scheduling tasks in autonomous vehicles that identify fixed and variable parameters prior to runtime, develop a schedule with an event loop to minimize execution time, and implement the plan using data from sensors, allowing for dynamic adjustments based on task dependencies and maximum execution times.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If current operating system schedulers are used, then the system is simple to implement, but task execution efficiency and latency performance deteriorate
Solution Approach 1:
The patent applies preliminary action by pre-compiling task schedules and execution plans before runtime. The offline compiler analyzes task dependencies, resource requirements, and timing constraints to generate optimized execution schedules in advance. During runtime, the system only needs to execute the pre-complied plan, significantly improving real-time performance without adding complex runtime scheduling logic.
Solution Approach 2:
The scheduling system is segmented into distinct phases: offline task analysis, schedule compilation, and runtime execution. The offline phase handles complex dependency analysis and optimization, while the runtime phase executes the pre-complied schedule. This segmentation separates the complexity of schedule generation from the simplicity of schedule execution, improving overall efficiency.
2Speed
If task schedules are optimized for minimum execution time, then processing speed improves, but the ability to handle variable task parameters deteriorates
Solution Approach 1:
The patent implements dynamics by distinguishing between fixed parameters (task dependencies, resource requirements) and variable parameters (sensor data, environmental conditions). The offline compiler optimizes schedules based on fixed parameters, while the runtime system adapts to variable parameters by selecting appropriate pre-complied schedules or adjusting execution based on actual sensor data and changing conditions.
Solution Approach 2:
The system changes parameters by separating static task characteristics (optimized offline) from dynamic runtime conditions (handled during execution). The offline compiler generates multiple schedules for different parameter configurations, and the runtime system selects the appropriate schedule based on current variable parameters, maintaining both optimization and adaptability.
3Reliability
If concurrent real-time tasks are executed with strict dependencies, then task correctness is ensured, but overall system latency increases
Solution Approach 1:
The patent resolves this contradiction by pre-analyzing task dependencies and execution paths offline. The compiler identifies critical paths and optimizes the execution order to minimize latency while ensuring all dependency constraints are satisfied. This preliminary optimization eliminates the need for conservative sequential execution at runtime, reducing latency while maintaining correctness.
Solution Approach 2:
The system ensures continuity of useful action by overlapping independent task executions and eliminating idle waiting time. The offline compiler identifies tasks that can execute in parallel and schedules them accordingly, ensuring that the processor is continuously productive. Dependency constraints are respected, but unnecessary delays are eliminated, reducing overall latency while maintaining task correctness.
Data Source
AI summary
A method of scheduling a plurality of tasks in an autonomous vehicle system (AVS) includes, by a processor, prior to runtime of an autonomous vehicle, identifying a plurality of tasks to be implemented by the AVS of the autonomous vehicle, for each of the tasks, identifying at least one fixed parameter and at least one variable, and developing a schedule for each of the tasks. The schedule includes an event loop that minimizes an overall time for execution of the tasks. The method includes compiling the schedule into an execution plan, and saving the execution plan to a memory of the autonomous vehicle. During runtime of the autonomous vehicle, the processor receives data corresponding to the variables of the tasks, and uses the variables to implement the execution plan on the autonomous vehicle.