Vehicle Compute Platooning for Ad Hoc Task Sharing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Automated vehicles face inefficiencies in handling complex computational tasks due to the impracticality of relying solely on their on-board processors, especially when tasks require significant computational resources and stable network connections, which may not always be available.
Innovation Solution
The system enables platooning of computational resources by dynamically forming ad hoc networks between automated vehicles, allowing them to share computational tasks that are predicted to consume a threshold amount of resources, by identifying nearby vehicles with available processors and scheduling task execution across multiple processors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If automated vehicles handle all computational tasks on their own-board processors, then task execution is independent and reliable, but computational resource demands become overwhelming and processing efficiency decreases
Solution Approach 1:
The patent combines computational resources from multiple automated vehicles into a shared pool, allowing tasks to be distributed across multiple processors. This merging of resources enables complex computational tasks to be handled more efficiently while maintaining reliability through distributed processing and task replication across the vehicle network.
2Productivity
If automated vehicles use ad hoc networks for computational sharing, then processing efficiency and resource utilization improve, but network stability and connection reliability may deteriorate
Solution Approach 1:
The patent implements dynamic task scheduling that adapts to changing network conditions. Tasks are scheduled and reassigned based on real-time vehicle proximity, network availability, and computational resource status. This dynamic approach allows the system to maintain high resource utilization while compensating for network instability through flexible task reallocation.
Solution Approach 2:
The system performs preliminary assessments of vehicle trajectories and predicted proximity to identify potential computational partners in advance. By pre-evaluating network conditions and task compatibility before actual execution, the system prepares contingency plans that cushion against potential network disruptions and ensure task completion reliability.
3Power
If computational tasks are distributed across multiple vehicles, then processing capacity increases, but system complexity and coordination overhead increase
Solution Approach 1:
The patent assumes homogeneous computational capabilities across vehicles in the network, using standardized task description formats and uniform scheduling protocols. This homogeneity simplifies the coordination complexity by allowing the system to treat all vehicles similarly, reducing the need for complex vehicle-specific coordination logic while still distributing tasks effectively across the network.
Data Source
AI summary
Novel techniques are described for platooning of computational resources in automated vehicle networks. An on-board computational processor of an automated vehicle typically performs a large number of computational tasks, and some of those computational tasks can be computationally intensive. Some such tasks, referred to as platoonable tasks herein, are well-suited for parallel processing by multiple processors. Embodiments can detect one or more on-board computational processors in one or more automated vehicles that are likely, during the time window in which the platoonable task will be executed, to have available computational resources and to be traveling along respective paths that are close enough to each other to allow for ad hoc network communications to be established between the processors. In response to detecting such cases, embodiments can schedule and instruct shared execution of the platoonable tasks by the multiple processors via the ad hoc network.