Distributed ECU Processing via CAN Bus Scheduling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The limitations of the Controller Area Network (CAN) protocol, such as small message size and dynamic availability of processing power in automotive systems, hinder the effective implementation of distributed computing systems, which are crucial for optimizing computing resources and reducing costs in the automotive industry.
Innovation Solution
A distributed computing system where ECUs on a CAN network dynamically report and utilize available processing power, with a scheduler ECU designating tasks to worker ECUs based on availability, allowing for efficient resource allocation and data collection across the network.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If CAN protocol is used for distributed computing, then network reliability and security are improved, but message size is limited to 8 bytes which restricts data transmission capability
Solution Approach 1:
The patent segments large data into multiple 8-byte CAN messages by dividing the data field into separate messages with sequential numbering. Each message contains a portion of the original data along with sequence identifiers, allowing reconstruction of the complete data set at the receiving end. This resolves the contradiction by maintaining CAN's reliable 8-byte message structure while enabling transmission of larger data volumes through multiple segmented messages.
2Productivity
If ECUs dynamically report available processing power, then computing resource allocation is optimized, but network traffic and coordination complexity increase
Solution Approach 1:
The patent implements self-service through autonomous ECU resource reporting where each ECU independently monitors and broadcasts its own processing power availability to the network. ECUs automatically detect their computational state and publish resource status messages without requiring centralized management or manual configuration. This reduces coordination complexity while maintaining optimized resource allocation, as the system self-regulates through decentralized ECUs serving their own resource management needs.
3Ease of manufacture
If multiple ECUs are connected to a single CAN bus, then system cost is reduced, but network traffic congestion increases with 20-30 nodes per bus
Solution Approach 1:
The patent transitions from a single-dimensional CAN bus architecture to a multi-dimensional network topology by organizing ECUs into logical groups or clusters that can communicate through hierarchical message routing. This dimensional expansion allows maintaining low-cost single-bus connections while implementing multi-level communication paths that reduce congestion by distributing traffic across multiple logical channels and priority levels.
Data Source
AI summary
A system includes a first electronic control unit (ECU) processor configured to function as a scheduler, connected to a bus on a controller area network. The processor is also configured to receive an availability notification, indicating ECU available processing power, from a second ECU connected to the bus. The processor is further configured to instruct the second ECU to become the scheduler, responsive to a determination that the ECU available processing power is greater than processing power available locally on the first ECU.


