Autonomous Vehicle Microgrid Scheduling for Idle Compute Utilization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles have high computing potential but face challenges in efficiently utilizing this resource due to geographical distance, isolation, and unpredictable downtime, making it difficult to provide user access and orchestrate computing power usage effectively.
Innovation Solution
A system and method for clustering autonomous vehicles into microgrids based on location and utilization rates, scheduling tasks using a cloud-based control system that includes a management unit and scheduler, and incentivizing vehicles to stay in the grid for task completion.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If autonomous vehicles are clustered into microgrids based on location and time, then computing resource utilization is improved, but system complexity increases
Solution Approach 1:
The system segments the fleet of autonomous vehicles into multiple microgrids based on geographic location and time of day. Each microgrid operates as a semi-independent computing cluster, allowing localized resource management while maintaining overall system coordination through the cloud-based control system.
Solution Approach 2:
The microgrid clustering is dynamically adjusted based on real-time vehicle locations, charging status, and computing workload demands. The system continuously reconfigures which vehicles belong to which microgrids, optimizing resource allocation as vehicles move between idle and active states throughout the day.
2Productivity
If tasks are scheduled on microgrids based on estimated runtime, then task completion efficiency is improved, but scheduling complexity increases
Solution Approach 1:
The system requires task submitters to provide estimated runtime information in advance. This preliminary data allows the cloud-based control system to pre-calculate optimal scheduling decisions, matching tasks to appropriate microgrids before execution begins, thereby reducing runtime coordination overhead.
Solution Approach 2:
The scheduling system uses feedback from actual task execution times and microgrid availability status to continuously refine future scheduling decisions. The cloud-based control system learns from past scheduling outcomes to improve task-to-microgrid matching accuracy over time.
3Reliability
If autonomous vehicles are held back to complete allocated tasks, then computing task reliability is improved, but transportation service quality deteriorates
Solution Approach 1:
The system holds back only the minimum necessary number of vehicles from transportation duties to ensure computing task completion. Not all idle vehicles are restricted, and the holding back duration is precisely calibrated to match task requirements, minimizing impact on transportation availability while ensuring sufficient computing resources.
Solution Approach 2:
The system dynamically adjusts the parameter of vehicle availability by temporarily changing the operational state of selected vehicles from 'transportation-ready' to 'computing-task-assigned'. This parameter change is time-bound and reversible, allowing vehicles to return to transportation service after task completion.
4Productivity
If users are provided access to computing resources of autonomous vehicles, then resource utilization is improved, but security and isolation challenges increase
Solution Approach 1:
The system segments user computing workloads into isolated containers that run on the autonomous vehicle's computing hardware. Each user's task is confined to its own computational environment, preventing interference or security breaches from affecting other users or the vehicle's core systems.
Solution Approach 2:
The cloud-based control system acts as an intermediary layer between users and the autonomous vehicle computing resources. It manages authentication, task allocation, and resource access control, providing a secure interface that protects both the vehicle systems and user data while enabling resource sharing.
Data Source
AI summary
One or more autonomous vehicles are clustered into one or more microgrids and at least one computing task is scheduled on at least one microgrid. Activation signals are received from a client of one or more autonomous vehicles when the vehicles are plugged into charging stations. Utilization rates of the autonomous vehicles are determined based on a set of parameters, which includes at least one of location of the autonomous vehicle and time. The autonomous vehicles are clustered into one or more microgrids of autonomous vehicles based on the utilization rates. At least one request for performing at least one computing task is received from a user device. The at least one request includes an estimated runtime of the at least one computing task. The at least one computing task is scheduled on at least one microgrid of autonomous vehicles based on the estimated runtime.