Multilevel Cloud Computing for Vehicle Surveillance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vehicle surveillance systems lack effective methods to identify and respond to erratic vehicle behavior in real-time across a fleet, potentially leading to collisions, as they rely on centralized data processing that can be overwhelmed by the volume of data from multiple vehicles and infrastructure sensors.
Innovation Solution
A multilevel cloud computing system that distributes data collection and processing across central, local servers, and vehicles, using infrastructure-mounted sensors to identify erratically moving vehicles and instruct them to move to the side of the road, with machine learning algorithms to set thresholds for normal behavior, allowing for decentralized computation and fleet-wide management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If centralized data processing is used to monitor vehicle behavior, then comprehensive fleet management is achieved, but the system becomes overwhelmed by the volume of data from multiple vehicles and infrastructure sensors
Solution Approach 1:
The patent divides the centralized monitoring system into multiple levels: local servers at intersections handle initial data collection and processing, regional servers aggregate data from multiple local servers, and a central server performs high-level analysis. This segmentation distributes the computational burden and prevents any single point from being overwhelmed by raw data volume.
Solution Approach 2:
The patent introduces a hierarchical dimension to the data processing architecture, organizing servers across multiple levels (local, regional, central) rather than using a single flat centralized structure. This dimensional organization allows data to be processed at appropriate levels of granularity, improving overall system efficiency.
2Reliability
If real-time identification of erratic vehicle behavior is implemented, then collision prevention is improved, but the computational requirements increase significantly
Solution Approach 1:
The system performs preliminary data collection and basic processing at local servers before data reaches higher levels. Machine learning models pre-process sensor data to identify potential erratic behaviors, so that when data reaches regional and central servers, the computational workload is already reduced and focused on critical cases.
Solution Approach 2:
Different levels of the hierarchy perform different types of processing appropriate to their capabilities and data volume. Local servers handle immediate sensor data and basic anomaly detection, regional servers perform aggregated analysis, and the central server focuses on fleet-wide patterns. This distributes computational power requirements across the hierarchy.
3Measurement precision
If machine learning algorithms are used to set thresholds for normal behavior, then detection accuracy improves, but system complexity increases
Solution Approach 1:
Machine learning algorithms automatically learn and adapt the thresholds for normal versus erratic vehicle behavior from historical data, eliminating the need for manual configuration and expert tuning. The system self-adjusts to changing traffic patterns and vehicle types, maintaining high detection accuracy without increasing operational complexity.
Data Source
AI summary
A computer includes a processor and a memory, the memory storing instructions executable by the processor to collect steering, speed, and position data about a plurality of vehicles from one or more infrastructure sensors, identify a vehicle that varies from a specified position in a roadway lane relative to a roadway lane marker or exceeds a threshold speed based on the collected data, instruct the identified vehicle to move to a side of a roadway, and send a message to a central server including an identification of the vehicle.